diff --git a/llvm/lib/Target/ARM/ARMISelLowering.cpp b/llvm/lib/Target/ARM/ARMISelLowering.cpp index 84b37ae6833ae..6b290135c5bcb 100644 --- a/llvm/lib/Target/ARM/ARMISelLowering.cpp +++ b/llvm/lib/Target/ARM/ARMISelLowering.cpp @@ -4971,14 +4971,14 @@ SDValue ARMTargetLowering::getVFPCmp(SDValue LHS, SDValue RHS, SelectionDAG &DAG, const SDLoc &dl, bool Signaling) const { assert(Subtarget->hasFP64() || RHS.getValueType() != MVT::f64); - SDValue Cmp; + SDValue Flags; if (!isFloatingPointZero(RHS)) - Cmp = DAG.getNode(Signaling ? ARMISD::CMPFPE : ARMISD::CMPFP, - dl, MVT::Glue, LHS, RHS); + Flags = DAG.getNode(Signaling ? ARMISD::CMPFPE : ARMISD::CMPFP, dl, FlagsVT, + LHS, RHS); else - Cmp = DAG.getNode(Signaling ? ARMISD::CMPFPEw0 : ARMISD::CMPFPw0, - dl, MVT::Glue, LHS); - return DAG.getNode(ARMISD::FMSTAT, dl, MVT::Glue, Cmp); + Flags = DAG.getNode(Signaling ? ARMISD::CMPFPEw0 : ARMISD::CMPFPw0, dl, + FlagsVT, LHS); + return DAG.getNode(ARMISD::FMSTAT, dl, MVT::Glue, Flags); } /// duplicateCmp - Glue values can have only one use, so this function @@ -4991,15 +4991,11 @@ ARMTargetLowering::duplicateCmp(SDValue Cmp, SelectionDAG &DAG) const { return DAG.getNode(Opc, DL, MVT::Glue, Cmp.getOperand(0),Cmp.getOperand(1)); assert(Opc == ARMISD::FMSTAT && "unexpected comparison operation"); - Cmp = Cmp.getOperand(0); - Opc = Cmp.getOpcode(); - if (Opc == ARMISD::CMPFP) - Cmp = DAG.getNode(Opc, DL, MVT::Glue, Cmp.getOperand(0),Cmp.getOperand(1)); - else { - assert(Opc == ARMISD::CMPFPw0 && "unexpected operand of FMSTAT"); - Cmp = DAG.getNode(Opc, DL, MVT::Glue, Cmp.getOperand(0)); - } - return DAG.getNode(ARMISD::FMSTAT, DL, MVT::Glue, Cmp); + SDValue Flags = Cmp.getOperand(0); + assert((Flags.getOpcode() == ARMISD::CMPFP || + Flags.getOpcode() == ARMISD::CMPFPw0) && + "unexpected operand of FMSTAT"); + return DAG.getNode(ARMISD::FMSTAT, DL, MVT::Glue, Flags); } // This function returns three things: the arithmetic computation itself diff --git a/llvm/lib/Target/ARM/ARMInstrVFP.td b/llvm/lib/Target/ARM/ARMInstrVFP.td index 5b49f728ebb8d..a29753909ea99 100644 --- a/llvm/lib/Target/ARM/ARMInstrVFP.td +++ b/llvm/lib/Target/ARM/ARMInstrVFP.td @@ -10,7 +10,17 @@ // //===----------------------------------------------------------------------===// -def SDT_CMPFP0 : SDTypeProfile<0, 1, [SDTCisFP<0>]>; +def SDT_CMPFP : SDTypeProfile<1, 2, [ + SDTCisVT<0, FlagsVT>, // out flags + SDTCisFP<1>, // lhs + SDTCisSameAs<2, 1> // rhs +]>; + +def SDT_CMPFP0 : SDTypeProfile<1, 1, [ + SDTCisVT<0, FlagsVT>, // out flags + SDTCisFP<1> // operand +]>; + def SDT_VMOVDRR : SDTypeProfile<1, 2, [SDTCisVT<0, f64>, SDTCisVT<1, i32>, SDTCisSameAs<1, 2>]>; def SDT_VMOVRRD : SDTypeProfile<2, 1, [SDTCisVT<0, i32>, SDTCisSameAs<0, 1>, @@ -18,11 +28,18 @@ def SDT_VMOVRRD : SDTypeProfile<2, 1, [SDTCisVT<0, i32>, SDTCisSameAs<0, 1>, def SDT_VMOVSR : SDTypeProfile<1, 1, [SDTCisVT<0, f32>, SDTCisVT<1, i32>]>; -def arm_fmstat : SDNode<"ARMISD::FMSTAT", SDTNone, [SDNPInGlue, SDNPOutGlue]>; -def arm_cmpfp : SDNode<"ARMISD::CMPFP", SDT_ARMCmp, [SDNPOutGlue]>; -def arm_cmpfp0 : SDNode<"ARMISD::CMPFPw0", SDT_CMPFP0, [SDNPOutGlue]>; -def arm_cmpfpe : SDNode<"ARMISD::CMPFPE", SDT_ARMCmp, [SDNPOutGlue]>; -def arm_cmpfpe0: SDNode<"ARMISD::CMPFPEw0",SDT_CMPFP0, [SDNPOutGlue]>; +def arm_cmpfp : SDNode<"ARMISD::CMPFP", SDT_CMPFP>; +def arm_cmpfp0 : SDNode<"ARMISD::CMPFPw0", SDT_CMPFP0>; +def arm_cmpfpe : SDNode<"ARMISD::CMPFPE", SDT_CMPFP>; +def arm_cmpfpe0 : SDNode<"ARMISD::CMPFPEw0", SDT_CMPFP0>; + +def arm_fmstat : SDNode<"ARMISD::FMSTAT", + SDTypeProfile<0, 1, [ + SDTCisVT<0, FlagsVT> // in flags + ]>, + [SDNPOutGlue] // TODO: Change Glue to a normal result. +>; + def arm_fmdrr : SDNode<"ARMISD::VMOVDRR", SDT_VMOVDRR>; def arm_fmrrd : SDNode<"ARMISD::VMOVRRD", SDT_VMOVRRD>; def arm_vmovsr : SDNode<"ARMISD::VMOVSR", SDT_VMOVSR>; @@ -606,12 +623,12 @@ let Defs = [FPSCR_NZCV] in { def VCMPED : ADuI<0b11101, 0b11, 0b0100, 0b11, 0, (outs), (ins DPR:$Dd, DPR:$Dm), IIC_fpCMP64, "vcmpe", ".f64\t$Dd, $Dm", "", - [(arm_cmpfpe DPR:$Dd, (f64 DPR:$Dm))]>; + [(set FPSCR_NZCV, (arm_cmpfpe DPR:$Dd, (f64 DPR:$Dm)))]>; def VCMPES : ASuI<0b11101, 0b11, 0b0100, 0b11, 0, (outs), (ins SPR:$Sd, SPR:$Sm), IIC_fpCMP32, "vcmpe", ".f32\t$Sd, $Sm", "", - [(arm_cmpfpe SPR:$Sd, SPR:$Sm)]> { + [(set FPSCR_NZCV, (arm_cmpfpe SPR:$Sd, SPR:$Sm))]> { // Some single precision VFP instructions may be executed on both NEON and // VFP pipelines on A8. let D = VFPNeonA8Domain; @@ -620,17 +637,17 @@ def VCMPES : ASuI<0b11101, 0b11, 0b0100, 0b11, 0, def VCMPEH : AHuI<0b11101, 0b11, 0b0100, 0b11, 0, (outs), (ins HPR:$Sd, HPR:$Sm), IIC_fpCMP16, "vcmpe", ".f16\t$Sd, $Sm", - [(arm_cmpfpe (f16 HPR:$Sd), (f16 HPR:$Sm))]>; + [(set FPSCR_NZCV, (arm_cmpfpe (f16 HPR:$Sd), (f16 HPR:$Sm)))]>; def VCMPD : ADuI<0b11101, 0b11, 0b0100, 0b01, 0, (outs), (ins DPR:$Dd, DPR:$Dm), IIC_fpCMP64, "vcmp", ".f64\t$Dd, $Dm", "", - [(arm_cmpfp DPR:$Dd, (f64 DPR:$Dm))]>; + [(set FPSCR_NZCV, (arm_cmpfp DPR:$Dd, (f64 DPR:$Dm)))]>; def VCMPS : ASuI<0b11101, 0b11, 0b0100, 0b01, 0, (outs), (ins SPR:$Sd, SPR:$Sm), IIC_fpCMP32, "vcmp", ".f32\t$Sd, $Sm", "", - [(arm_cmpfp SPR:$Sd, SPR:$Sm)]> { + [(set FPSCR_NZCV, (arm_cmpfp SPR:$Sd, SPR:$Sm))]> { // Some single precision VFP instructions may be executed on both NEON and // VFP pipelines on A8. let D = VFPNeonA8Domain; @@ -639,7 +656,7 @@ def VCMPS : ASuI<0b11101, 0b11, 0b0100, 0b01, 0, def VCMPH : AHuI<0b11101, 0b11, 0b0100, 0b01, 0, (outs), (ins HPR:$Sd, HPR:$Sm), IIC_fpCMP16, "vcmp", ".f16\t$Sd, $Sm", - [(arm_cmpfp (f16 HPR:$Sd), (f16 HPR:$Sm))]>; + [(set FPSCR_NZCV, (arm_cmpfp (f16 HPR:$Sd), (f16 HPR:$Sm)))]>; } // Defs = [FPSCR_NZCV] //===----------------------------------------------------------------------===// @@ -669,7 +686,7 @@ let Defs = [FPSCR_NZCV] in { def VCMPEZD : ADuI<0b11101, 0b11, 0b0101, 0b11, 0, (outs), (ins DPR:$Dd), IIC_fpCMP64, "vcmpe", ".f64\t$Dd, #0", "", - [(arm_cmpfpe0 (f64 DPR:$Dd))]> { + [(set FPSCR_NZCV, (arm_cmpfpe0 (f64 DPR:$Dd)))]> { let Inst{3-0} = 0b0000; let Inst{5} = 0; } @@ -677,7 +694,7 @@ def VCMPEZD : ADuI<0b11101, 0b11, 0b0101, 0b11, 0, def VCMPEZS : ASuI<0b11101, 0b11, 0b0101, 0b11, 0, (outs), (ins SPR:$Sd), IIC_fpCMP32, "vcmpe", ".f32\t$Sd, #0", "", - [(arm_cmpfpe0 SPR:$Sd)]> { + [(set FPSCR_NZCV, (arm_cmpfpe0 SPR:$Sd))]> { let Inst{3-0} = 0b0000; let Inst{5} = 0; @@ -689,7 +706,7 @@ def VCMPEZS : ASuI<0b11101, 0b11, 0b0101, 0b11, 0, def VCMPEZH : AHuI<0b11101, 0b11, 0b0101, 0b11, 0, (outs), (ins HPR:$Sd), IIC_fpCMP16, "vcmpe", ".f16\t$Sd, #0", - [(arm_cmpfpe0 (f16 HPR:$Sd))]> { + [(set FPSCR_NZCV, (arm_cmpfpe0 (f16 HPR:$Sd)))]> { let Inst{3-0} = 0b0000; let Inst{5} = 0; } @@ -697,7 +714,7 @@ def VCMPEZH : AHuI<0b11101, 0b11, 0b0101, 0b11, 0, def VCMPZD : ADuI<0b11101, 0b11, 0b0101, 0b01, 0, (outs), (ins DPR:$Dd), IIC_fpCMP64, "vcmp", ".f64\t$Dd, #0", "", - [(arm_cmpfp0 (f64 DPR:$Dd))]> { + [(set FPSCR_NZCV, (arm_cmpfp0 (f64 DPR:$Dd)))]> { let Inst{3-0} = 0b0000; let Inst{5} = 0; } @@ -705,7 +722,7 @@ def VCMPZD : ADuI<0b11101, 0b11, 0b0101, 0b01, 0, def VCMPZS : ASuI<0b11101, 0b11, 0b0101, 0b01, 0, (outs), (ins SPR:$Sd), IIC_fpCMP32, "vcmp", ".f32\t$Sd, #0", "", - [(arm_cmpfp0 SPR:$Sd)]> { + [(set FPSCR_NZCV, (arm_cmpfp0 SPR:$Sd))]> { let Inst{3-0} = 0b0000; let Inst{5} = 0; @@ -717,7 +734,7 @@ def VCMPZS : ASuI<0b11101, 0b11, 0b0101, 0b01, 0, def VCMPZH : AHuI<0b11101, 0b11, 0b0101, 0b01, 0, (outs), (ins HPR:$Sd), IIC_fpCMP16, "vcmp", ".f16\t$Sd, #0", - [(arm_cmpfp0 (f16 HPR:$Sd))]> { + [(set FPSCR_NZCV, (arm_cmpfp0 (f16 HPR:$Sd)))]> { let Inst{3-0} = 0b0000; let Inst{5} = 0; } @@ -2492,7 +2509,8 @@ let DecoderMethod = "DecodeForVMRSandVMSR" in { let Defs = [CPSR], Uses = [FPSCR_NZCV], Predicates = [HasFPRegs], Rt = 0b1111 /* apsr_nzcv */ in def FMSTAT : MovFromVFP<0b0001 /* fpscr */, (outs), (ins), - "vmrs", "\tAPSR_nzcv, fpscr", [(arm_fmstat)]>; + "vmrs", "\tAPSR_nzcv, fpscr", + [(arm_fmstat FPSCR_NZCV)]>; // Application level FPSCR -> GPR let hasSideEffects = 1, Uses = [FPSCR], Predicates = [HasFPRegs] in diff --git a/llvm/lib/Target/ARM/ARMRegisterInfo.td b/llvm/lib/Target/ARM/ARMRegisterInfo.td index f37d0fe542b4f..f5a675e2976bb 100644 --- a/llvm/lib/Target/ARM/ARMRegisterInfo.td +++ b/llvm/lib/Target/ARM/ARMRegisterInfo.td @@ -413,7 +413,9 @@ def VCCR : RegisterClass<"ARM", [i32, v16i1, v8i1, v4i1, v2i1], 32, (add VPR)> { // FPSCR, when the flags at the top of it are used as the input or // output to an instruction such as MVE VADC. -def cl_FPSCR_NZCV : RegisterClass<"ARM", [i32], 32, (add FPSCR_NZCV)>; +def cl_FPSCR_NZCV : RegisterClass<"ARM", [i32], 32, (add FPSCR_NZCV)> { + let CopyCost = -1; +} // Scalar single precision floating point register class.. // FIXME: Allocation order changed to s0, s2, ... or s0, s4, ... as a quick hack diff --git a/llvm/test/CodeGen/ARM/fcmp-xo.ll b/llvm/test/CodeGen/ARM/fcmp-xo.ll index 3d5972f065859..908dbd7a11a6b 100644 --- a/llvm/test/CodeGen/ARM/fcmp-xo.ll +++ b/llvm/test/CodeGen/ARM/fcmp-xo.ll @@ -54,12 +54,12 @@ define arm_aapcs_vfpcc float @float128(float %a0) local_unnamed_addr { ; NEON-LABEL: float128: ; NEON: @ %bb.0: ; NEON-NEXT: mov.w r0, #1124073472 -; NEON-NEXT: vmov.f32 s2, #5.000000e-01 -; NEON-NEXT: vmov d3, r0, r0 -; NEON-NEXT: vmov.f32 s4, #-5.000000e-01 -; NEON-NEXT: vcmp.f32 s6, s0 +; NEON-NEXT: vmov.f32 s4, #5.000000e-01 +; NEON-NEXT: vmov d1, r0, r0 +; NEON-NEXT: vmov.f32 s6, #-5.000000e-01 +; NEON-NEXT: vcmp.f32 s2, s0 ; NEON-NEXT: vmrs APSR_nzcv, fpscr -; NEON-NEXT: vselgt.f32 s0, s4, s2 +; NEON-NEXT: vselgt.f32 s0, s6, s4 ; NEON-NEXT: bx lr %1 = fcmp nsz olt float %a0, 128.000000e+00 %2 = select i1 %1, float -5.000000e-01, float 5.000000e-01 diff --git a/llvm/test/CodeGen/ARM/fp16-instructions.ll b/llvm/test/CodeGen/ARM/fp16-instructions.ll index 1988cb1d2f903..7a1d5ddfa301b 100644 --- a/llvm/test/CodeGen/ARM/fp16-instructions.ll +++ b/llvm/test/CodeGen/ARM/fp16-instructions.ll @@ -700,9 +700,9 @@ define half @select_cc1(ptr %a0) { ; CHECK-LABEL: select_cc1: -; CHECK-HARDFP-FULLFP16: vcmp.f16 -; CHECK-HARDFP-FULLFP16-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-HARDFP-FULLFP16-NEXT: vseleq.f16 s0, +; CHECK-HARDFP-FULLFP16: vcmp.f16 +; CHECK-HARDFP-FULLFP16: vmrs APSR_nzcv, fpscr +; CHECK-HARDFP-FULLFP16: vseleq.f16 s0, ; CHECK-SOFTFP-FP16-A32: vcmp.f32 ; CHECK-SOFTFP-FP16-A32-NEXT: vmrs APSR_nzcv, fpscr @@ -728,9 +728,9 @@ define half @select_cc_ge1(ptr %a0) { ; CHECK-LABEL: select_cc_ge1: -; CHECK-HARDFP-FULLFP16: vcmp.f16 -; CHECK-HARDFP-FULLFP16-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-HARDFP-FULLFP16-NEXT: vselge.f16 s0, +; CHECK-HARDFP-FULLFP16: vcmp.f16 +; CHECK-HARDFP-FULLFP16: vmrs APSR_nzcv, fpscr +; CHECK-HARDFP-FULLFP16: vselge.f16 s0, ; CHECK-SOFTFP-FP16-A32: vcmp.f32 ; CHECK-SOFTFP-FP16-A32-NEXT: vmrs APSR_nzcv, fpscr @@ -751,9 +751,9 @@ define half @select_cc_ge2(ptr %a0) { ; CHECK-LABEL: select_cc_ge2: -; CHECK-HARDFP-FULLFP16: vcmp.f16 -; CHECK-HARDFP-FULLFP16-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-HARDFP-FULLFP16-NEXT: vselge.f16 s0, +; CHECK-HARDFP-FULLFP16: vcmp.f16 +; CHECK-HARDFP-FULLFP16: vmrs APSR_nzcv, fpscr +; CHECK-HARDFP-FULLFP16: vselge.f16 s0, ; CHECK-SOFTFP-FP16-A32: vcmp.f32 ; CHECK-SOFTFP-FP16-A32-NEXT: vmrs APSR_nzcv, fpscr @@ -774,9 +774,9 @@ define half @select_cc_ge3(ptr %a0) { ; CHECK-LABEL: select_cc_ge3: -; CHECK-HARDFP-FULLFP16: vcmp.f16 -; CHECK-HARDFP-FULLFP16-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-HARDFP-FULLFP16-NEXT: vselge.f16 s0, +; CHECK-HARDFP-FULLFP16: vcmp.f16 +; CHECK-HARDFP-FULLFP16: vmrs APSR_nzcv, fpscr +; CHECK-HARDFP-FULLFP16: vselge.f16 s0, ; CHECK-SOFTFP-FP16-A32: vcmp.f32 ; CHECK-SOFTFP-FP16-A32-NEXT: vmrs APSR_nzcv, fpscr @@ -797,9 +797,9 @@ define half @select_cc_ge4(ptr %a0) { ; CHECK-LABEL: select_cc_ge4: -; CHECK-HARDFP-FULLFP16: vcmp.f16 -; CHECK-HARDFP-FULLFP16-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-HARDFP-FULLFP16-NEXT: vselge.f16 s0, s{{.}}, s{{.}} +; CHECK-HARDFP-FULLFP16: vcmp.f16 +; CHECK-HARDFP-FULLFP16: vmrs APSR_nzcv, fpscr +; CHECK-HARDFP-FULLFP16: vselge.f16 s0, s{{.}}, s{{.}} ; CHECK-SOFTFP-FP16-A32: vcmp.f32 ; CHECK-SOFTFP-FP16-A32-NEXT: vmrs APSR_nzcv, fpscr @@ -821,9 +821,9 @@ define half @select_cc_gt1(ptr %a0) { ; CHECK-LABEL: select_cc_gt1: -; CHECK-HARDFP-FULLFP16: vcmp.f16 -; CHECK-HARDFP-FULLFP16-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-HARDFP-FULLFP16-NEXT: vselgt.f16 s0, s{{.}}, s{{.}} +; CHECK-HARDFP-FULLFP16: vcmp.f16 +; CHECK-HARDFP-FULLFP16: vmrs APSR_nzcv, fpscr +; CHECK-HARDFP-FULLFP16: vselgt.f16 s0, s{{.}}, s{{.}} ; CHECK-SOFTFP-FP16-A32: vcmp.f32 ; CHECK-SOFTFP-FP16-A32-NEXT: vmrs APSR_nzcv, fpscr @@ -844,9 +844,9 @@ define half @select_cc_gt2(ptr %a0) { ; CHECK-LABEL: select_cc_gt2: -; CHECK-HARDFP-FULLFP16: vcmp.f16 -; CHECK-HARDFP-FULLFP16-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-HARDFP-FULLFP16-NEXT: vselgt.f16 s0, s{{.}}, s{{.}} +; CHECK-HARDFP-FULLFP16: vcmp.f16 +; CHECK-HARDFP-FULLFP16: vmrs APSR_nzcv, fpscr +; CHECK-HARDFP-FULLFP16: vselgt.f16 s0, s{{.}}, s{{.}} ; CHECK-SOFTFP-FP16-A32: vcmp.f32 ; CHECK-SOFTFP-FP16-A32-NEXT: vmrs APSR_nzcv, fpscr @@ -867,9 +867,9 @@ define half @select_cc_gt3(ptr %a0) { ; CHECK-LABEL: select_cc_gt3: -; CHECK-HARDFP-FULLFP16: vcmp.f16 -; CHECK-HARDFP-FULLFP16-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-HARDFP-FULLFP16-NEXT: vselgt.f16 s0, s{{.}}, s{{.}} +; CHECK-HARDFP-FULLFP16: vcmp.f16 +; CHECK-HARDFP-FULLFP16: vmrs APSR_nzcv, fpscr +; CHECK-HARDFP-FULLFP16: vselgt.f16 s0, s{{.}}, s{{.}} ; CHECK-SOFTFP-FP16-A32: vcmp.f32 ; CHECK-SOFTFP-FP16-A32-NEXT: vmrs APSR_nzcv, fpscr @@ -890,9 +890,9 @@ define half @select_cc_gt4(ptr %a0) { ; CHECK-LABEL: select_cc_gt4: -; CHECK-HARDFP-FULLFP16: vcmp.f16 -; CHECK-HARDFP-FULLFP16-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-HARDFP-FULLFP16-NEXT: vselgt.f16 s0, s{{.}}, s{{.}} +; CHECK-HARDFP-FULLFP16: vcmp.f16 +; CHECK-HARDFP-FULLFP16: vmrs APSR_nzcv, fpscr +; CHECK-HARDFP-FULLFP16: vselgt.f16 s0, s{{.}}, s{{.}} ; CHECK-SOFTFP-FP16-A32: vcmp.f32 ; CHECK-SOFTFP-FP16-A32-NEXT: vmrs APSR_nzcv, fpscr @@ -923,10 +923,10 @@ entry: ; CHECK-LABEL: select_cc4: ; CHECK-HARDFP-FULLFP16: vldr.16 [[S2:s[0-9]]], .LCPI{{.*}} +; CHECK-HARDFP-FULLFP16: vcmp.f16 s0, [[S2]] ; CHECK-HARDFP-FULLFP16: vldr.16 [[S4:s[0-9]]], .LCPI{{.*}} +; CHECK-HARDFP-FULLFP16: vmrs APSR_nzcv, fpscr ; CHECK-HARDFP-FULLFP16: vmov.f16 [[S6:s[0-9]]], #-2.000000e+00 -; CHECK-HARDFP-FULLFP16: vcmp.f16 s0, [[S2]] -; CHECK-HARDFP-FULLFP16-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-HARDFP-FULLFP16-NEXT: vseleq.f16 [[S0:s[0-9]]], [[S6]], [[S4]] ; CHECK-HARDFP-FULLFP16-NEXT: vselvs.f16 s0, [[S6]], [[S0]] diff --git a/llvm/test/CodeGen/ARM/fp16-vminmaxnm-safe.ll b/llvm/test/CodeGen/ARM/fp16-vminmaxnm-safe.ll index 56e734c440433..996b46c51ab36 100644 --- a/llvm/test/CodeGen/ARM/fp16-vminmaxnm-safe.ll +++ b/llvm/test/CodeGen/ARM/fp16-vminmaxnm-safe.ll @@ -5,11 +5,11 @@ define half @fp16_vminnm_o(half %a, half %b) { ; CHECK-LABEL: fp16_vminnm_o: ; CHECK: @ %bb.0: @ %entry -; CHECK-NEXT: vmov.f16 s0, r1 -; CHECK-NEXT: vmov.f16 s2, r0 -; CHECK-NEXT: vcmp.f16 s0, s2 +; CHECK-NEXT: vmov.f16 s0, r0 +; CHECK-NEXT: vmov.f16 s2, r1 +; CHECK-NEXT: vcmp.f16 s2, s0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vselgt.f16 s0, s2, s0 +; CHECK-NEXT: vselgt.f16 s0, s0, s2 ; CHECK-NEXT: vmov r0, s0 ; CHECK-NEXT: bx lr entry: @@ -37,11 +37,11 @@ entry: define half @fp16_vminnm_u(half %a, half %b) { ; CHECK-LABEL: fp16_vminnm_u: ; CHECK: @ %bb.0: @ %entry -; CHECK-NEXT: vmov.f16 s0, r0 -; CHECK-NEXT: vmov.f16 s2, r1 -; CHECK-NEXT: vcmp.f16 s0, s2 +; CHECK-NEXT: vmov.f16 s0, r1 +; CHECK-NEXT: vmov.f16 s2, r0 +; CHECK-NEXT: vcmp.f16 s2, s0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vselge.f16 s0, s2, s0 +; CHECK-NEXT: vselge.f16 s0, s0, s2 ; CHECK-NEXT: vmov r0, s0 ; CHECK-NEXT: bx lr entry: @@ -53,11 +53,11 @@ entry: define half @fp16_vminnm_ule(half %a, half %b) { ; CHECK-LABEL: fp16_vminnm_ule: ; CHECK: @ %bb.0: @ %entry -; CHECK-NEXT: vmov.f16 s0, r0 -; CHECK-NEXT: vmov.f16 s2, r1 -; CHECK-NEXT: vcmp.f16 s0, s2 +; CHECK-NEXT: vmov.f16 s0, r1 +; CHECK-NEXT: vmov.f16 s2, r0 +; CHECK-NEXT: vcmp.f16 s2, s0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vselgt.f16 s0, s2, s0 +; CHECK-NEXT: vselgt.f16 s0, s0, s2 ; CHECK-NEXT: vmov r0, s0 ; CHECK-NEXT: bx lr entry: @@ -69,11 +69,11 @@ entry: define half @fp16_vminnm_u_rev(half %a, half %b) { ; CHECK-LABEL: fp16_vminnm_u_rev: ; CHECK: @ %bb.0: @ %entry -; CHECK-NEXT: vmov.f16 s0, r1 -; CHECK-NEXT: vmov.f16 s2, r0 -; CHECK-NEXT: vcmp.f16 s0, s2 +; CHECK-NEXT: vmov.f16 s0, r0 +; CHECK-NEXT: vmov.f16 s2, r1 +; CHECK-NEXT: vcmp.f16 s2, s0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vselge.f16 s0, s2, s0 +; CHECK-NEXT: vselge.f16 s0, s0, s2 ; CHECK-NEXT: vmov r0, s0 ; CHECK-NEXT: bx lr entry: diff --git a/llvm/test/CodeGen/ARM/fptosi-sat-scalar.ll b/llvm/test/CodeGen/ARM/fptosi-sat-scalar.ll index 4b27e804e6df9..84f6ee276ba5f 100644 --- a/llvm/test/CodeGen/ARM/fptosi-sat-scalar.ll +++ b/llvm/test/CodeGen/ARM/fptosi-sat-scalar.ll @@ -258,11 +258,11 @@ define i13 @test_signed_i13_f32(float %f) nounwind { ; VFP2: @ %bb.0: ; VFP2-NEXT: vmov s0, r0 ; VFP2-NEXT: vldr s2, .LCPI2_0 +; VFP2-NEXT: vldr s6, .LCPI2_1 ; VFP2-NEXT: vcvt.s32.f32 s4, s0 ; VFP2-NEXT: vcmp.f32 s0, s2 -; VFP2-NEXT: vldr s2, .LCPI2_1 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s0, s2 +; VFP2-NEXT: vcmp.f32 s0, s6 ; VFP2-NEXT: vmov r0, s4 ; VFP2-NEXT: itt lt ; VFP2-NEXT: movwlt r0, #61440 @@ -358,11 +358,11 @@ define i16 @test_signed_i16_f32(float %f) nounwind { ; VFP2: @ %bb.0: ; VFP2-NEXT: vmov s0, r0 ; VFP2-NEXT: vldr s2, .LCPI3_0 +; VFP2-NEXT: vldr s6, .LCPI3_1 ; VFP2-NEXT: vcvt.s32.f32 s4, s0 ; VFP2-NEXT: vcmp.f32 s0, s2 -; VFP2-NEXT: vldr s2, .LCPI3_1 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s0, s2 +; VFP2-NEXT: vcmp.f32 s0, s6 ; VFP2-NEXT: vmov r0, s4 ; VFP2-NEXT: itt lt ; VFP2-NEXT: movwlt r0, #32768 @@ -458,11 +458,11 @@ define i19 @test_signed_i19_f32(float %f) nounwind { ; VFP2: @ %bb.0: ; VFP2-NEXT: vmov s0, r0 ; VFP2-NEXT: vldr s2, .LCPI4_0 +; VFP2-NEXT: vldr s6, .LCPI4_1 ; VFP2-NEXT: vcvt.s32.f32 s4, s0 ; VFP2-NEXT: vcmp.f32 s0, s2 -; VFP2-NEXT: vldr s2, .LCPI4_1 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s0, s2 +; VFP2-NEXT: vcmp.f32 s0, s6 ; VFP2-NEXT: vmov r0, s4 ; VFP2-NEXT: itt lt ; VFP2-NEXT: movlt r0, #0 @@ -639,39 +639,31 @@ define i50 @test_signed_i50_f32(float %f) nounwind { ; ; VFP-LABEL: test_signed_i50_f32: ; VFP: @ %bb.0: -; VFP-NEXT: .save {r7, lr} -; VFP-NEXT: push {r7, lr} -; VFP-NEXT: .vsave {d8} -; VFP-NEXT: vpush {d8} -; VFP-NEXT: vmov s16, r0 +; VFP-NEXT: .save {r4, lr} +; VFP-NEXT: push {r4, lr} +; VFP-NEXT: mov r4, r0 ; VFP-NEXT: bl __aeabi_f2lz ; VFP-NEXT: vldr s0, .LCPI6_0 -; VFP-NEXT: vldr s2, .LCPI6_1 -; VFP-NEXT: vcmp.f32 s16, s0 +; VFP-NEXT: vmov s2, r4 +; VFP-NEXT: vldr s4, .LCPI6_1 +; VFP-NEXT: vcmp.f32 s2, s0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, s2 -; VFP-NEXT: itt lt +; VFP-NEXT: ittt lt ; VFP-NEXT: movlt r1, #0 ; VFP-NEXT: movtlt r1, #65534 +; VFP-NEXT: movlt r0, #0 +; VFP-NEXT: vcmp.f32 s2, s4 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, s0 -; VFP-NEXT: itt gt +; VFP-NEXT: ittt gt ; VFP-NEXT: movwgt r1, #65535 ; VFP-NEXT: movtgt r1, #1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it lt -; VFP-NEXT: movlt r0, #0 -; VFP-NEXT: vcmp.f32 s16, s2 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it gt ; VFP-NEXT: movgt.w r0, #-1 -; VFP-NEXT: vcmp.f32 s16, s16 +; VFP-NEXT: vcmp.f32 s2, s2 ; VFP-NEXT: vmrs APSR_nzcv, fpscr ; VFP-NEXT: itt vs ; VFP-NEXT: movvs r0, #0 ; VFP-NEXT: movvs r1, #0 -; VFP-NEXT: vpop {d8} -; VFP-NEXT: pop {r7, pc} +; VFP-NEXT: pop {r4, pc} ; VFP-NEXT: .p2align 2 ; VFP-NEXT: @ %bb.1: ; VFP-NEXT: .LCPI6_0: @@ -765,27 +757,18 @@ define i64 @test_signed_i64_f32(float %f) nounwind { ; VFP-NEXT: vldr s4, .LCPI7_1 ; VFP-NEXT: vcmp.f32 s2, s0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s4 -; VFP-NEXT: it lt +; VFP-NEXT: itt lt ; VFP-NEXT: movlt r0, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s2 -; VFP-NEXT: it gt -; VFP-NEXT: movgt.w r0, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s0 -; VFP-NEXT: it vs -; VFP-NEXT: movvs r0, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it lt ; VFP-NEXT: movlt.w r1, #-2147483648 ; VFP-NEXT: vcmp.f32 s2, s4 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it gt +; VFP-NEXT: itt gt ; VFP-NEXT: mvngt r1, #-2147483648 +; VFP-NEXT: movgt.w r0, #-1 ; VFP-NEXT: vcmp.f32 s2, s2 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it vs +; VFP-NEXT: itt vs +; VFP-NEXT: movvs r0, #0 ; VFP-NEXT: movvs r1, #0 ; VFP-NEXT: pop {r4, pc} ; VFP-NEXT: .p2align 2 @@ -923,51 +906,24 @@ define i100 @test_signed_i100_f32(float %f) nounwind { ; VFP-NEXT: vldr s4, .LCPI8_1 ; VFP-NEXT: vcmp.f32 s2, s0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s4 -; VFP-NEXT: it lt +; VFP-NEXT: itttt lt ; VFP-NEXT: movlt r0, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s2 -; VFP-NEXT: it gt -; VFP-NEXT: movgt.w r0, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s0 -; VFP-NEXT: it vs -; VFP-NEXT: movvs r0, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s4 -; VFP-NEXT: it lt ; VFP-NEXT: movlt r1, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s2 -; VFP-NEXT: it gt -; VFP-NEXT: movgt.w r1, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s0 -; VFP-NEXT: it vs -; VFP-NEXT: movvs r1, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s4 -; VFP-NEXT: it lt ; VFP-NEXT: movlt r2, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s2 -; VFP-NEXT: it gt -; VFP-NEXT: movgt.w r2, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s0 -; VFP-NEXT: it vs -; VFP-NEXT: movvs r2, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it lt ; VFP-NEXT: mvnlt r3, #7 ; VFP-NEXT: vcmp.f32 s2, s4 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it gt +; VFP-NEXT: itttt gt ; VFP-NEXT: movgt r3, #7 +; VFP-NEXT: movgt.w r2, #-1 +; VFP-NEXT: movgt.w r1, #-1 +; VFP-NEXT: movgt.w r0, #-1 ; VFP-NEXT: vcmp.f32 s2, s2 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it vs +; VFP-NEXT: itttt vs +; VFP-NEXT: movvs r0, #0 +; VFP-NEXT: movvs r1, #0 +; VFP-NEXT: movvs r2, #0 ; VFP-NEXT: movvs r3, #0 ; VFP-NEXT: pop {r4, pc} ; VFP-NEXT: .p2align 2 @@ -1108,51 +1064,24 @@ define i128 @test_signed_i128_f32(float %f) nounwind { ; VFP-NEXT: vldr s4, .LCPI9_1 ; VFP-NEXT: vcmp.f32 s2, s0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s4 -; VFP-NEXT: it lt +; VFP-NEXT: itttt lt ; VFP-NEXT: movlt r0, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s2 -; VFP-NEXT: it gt -; VFP-NEXT: movgt.w r0, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s0 -; VFP-NEXT: it vs -; VFP-NEXT: movvs r0, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s4 -; VFP-NEXT: it lt ; VFP-NEXT: movlt r1, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s2 -; VFP-NEXT: it gt -; VFP-NEXT: movgt.w r1, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s0 -; VFP-NEXT: it vs -; VFP-NEXT: movvs r1, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s4 -; VFP-NEXT: it lt ; VFP-NEXT: movlt r2, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s2 -; VFP-NEXT: it gt -; VFP-NEXT: movgt.w r2, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s2, s0 -; VFP-NEXT: it vs -; VFP-NEXT: movvs r2, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it lt ; VFP-NEXT: movlt.w r3, #-2147483648 ; VFP-NEXT: vcmp.f32 s2, s4 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it gt +; VFP-NEXT: itttt gt ; VFP-NEXT: mvngt r3, #-2147483648 +; VFP-NEXT: movgt.w r2, #-1 +; VFP-NEXT: movgt.w r1, #-1 +; VFP-NEXT: movgt.w r0, #-1 ; VFP-NEXT: vcmp.f32 s2, s2 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it vs +; VFP-NEXT: itttt vs +; VFP-NEXT: movvs r0, #0 +; VFP-NEXT: movvs r1, #0 +; VFP-NEXT: movvs r2, #0 ; VFP-NEXT: movvs r3, #0 ; VFP-NEXT: pop {r4, pc} ; VFP-NEXT: .p2align 2 @@ -1451,15 +1380,15 @@ define i13 @test_signed_i13_f64(double %f) nounwind { ; VFP2: @ %bb.0: ; VFP2-NEXT: vmov d16, r0, r1 ; VFP2-NEXT: vldr d17, .LCPI12_0 +; VFP2-NEXT: vldr d18, .LCPI12_1 ; VFP2-NEXT: vcvt.s32.f64 s0, d16 ; VFP2-NEXT: vcmp.f64 d16, d17 -; VFP2-NEXT: vldr d17, .LCPI12_1 -; VFP2-NEXT: vmov r0, s0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr +; VFP2-NEXT: vmov r0, s0 +; VFP2-NEXT: vcmp.f64 d16, d18 ; VFP2-NEXT: itt lt ; VFP2-NEXT: movwlt r0, #61440 ; VFP2-NEXT: movtlt r0, #65535 -; VFP2-NEXT: vcmp.f64 d16, d17 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr ; VFP2-NEXT: it gt ; VFP2-NEXT: movwgt r0, #4095 @@ -1568,15 +1497,15 @@ define i16 @test_signed_i16_f64(double %f) nounwind { ; VFP2: @ %bb.0: ; VFP2-NEXT: vmov d16, r0, r1 ; VFP2-NEXT: vldr d17, .LCPI13_0 +; VFP2-NEXT: vldr d18, .LCPI13_1 ; VFP2-NEXT: vcvt.s32.f64 s0, d16 ; VFP2-NEXT: vcmp.f64 d16, d17 -; VFP2-NEXT: vldr d17, .LCPI13_1 -; VFP2-NEXT: vmov r0, s0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr +; VFP2-NEXT: vmov r0, s0 +; VFP2-NEXT: vcmp.f64 d16, d18 ; VFP2-NEXT: itt lt ; VFP2-NEXT: movwlt r0, #32768 ; VFP2-NEXT: movtlt r0, #65535 -; VFP2-NEXT: vcmp.f64 d16, d17 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr ; VFP2-NEXT: it gt ; VFP2-NEXT: movwgt r0, #32767 @@ -1685,15 +1614,15 @@ define i19 @test_signed_i19_f64(double %f) nounwind { ; VFP2: @ %bb.0: ; VFP2-NEXT: vmov d16, r0, r1 ; VFP2-NEXT: vldr d17, .LCPI14_0 +; VFP2-NEXT: vldr d18, .LCPI14_1 ; VFP2-NEXT: vcvt.s32.f64 s0, d16 ; VFP2-NEXT: vcmp.f64 d16, d17 -; VFP2-NEXT: vldr d17, .LCPI14_1 -; VFP2-NEXT: vmov r0, s0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr +; VFP2-NEXT: vmov r0, s0 +; VFP2-NEXT: vcmp.f64 d16, d18 ; VFP2-NEXT: itt lt ; VFP2-NEXT: movlt r0, #0 ; VFP2-NEXT: movtlt r0, #65532 -; VFP2-NEXT: vcmp.f64 d16, d17 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr ; VFP2-NEXT: itt gt ; VFP2-NEXT: movwgt r0, #65535 @@ -1905,42 +1834,32 @@ define i50 @test_signed_i50_f64(double %f) nounwind { ; ; VFP2-LABEL: test_signed_i50_f64: ; VFP2: @ %bb.0: -; VFP2-NEXT: .save {r7, lr} -; VFP2-NEXT: push {r7, lr} -; VFP2-NEXT: .vsave {d8} -; VFP2-NEXT: vpush {d8} -; VFP2-NEXT: vmov d8, r0, r1 +; VFP2-NEXT: .save {r4, r5, r7, lr} +; VFP2-NEXT: push {r4, r5, r7, lr} +; VFP2-NEXT: mov r4, r1 +; VFP2-NEXT: mov r5, r0 ; VFP2-NEXT: bl __aeabi_d2lz ; VFP2-NEXT: vldr d16, .LCPI16_0 -; VFP2-NEXT: vldr d17, .LCPI16_1 -; VFP2-NEXT: vcmp.f64 d8, d16 +; VFP2-NEXT: vmov d17, r5, r4 +; VFP2-NEXT: vldr d18, .LCPI16_1 +; VFP2-NEXT: vcmp.f64 d17, d16 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: itt lt +; VFP2-NEXT: ittt lt ; VFP2-NEXT: movlt r1, #0 ; VFP2-NEXT: movtlt r1, #65534 -; VFP2-NEXT: vcmp.f64 d8, d17 +; VFP2-NEXT: movlt r0, #0 +; VFP2-NEXT: vcmp.f64 d17, d18 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: itt gt +; VFP2-NEXT: ittt gt ; VFP2-NEXT: movwgt r1, #65535 ; VFP2-NEXT: movtgt r1, #1 -; VFP2-NEXT: vcmp.f64 d8, d8 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r1, #0 -; VFP2-NEXT: vcmp.f64 d8, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vcmp.f64 d8, d17 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vcmp.f64 d8, d8 +; VFP2-NEXT: vcmp.f64 d17, d17 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs +; VFP2-NEXT: itt vs ; VFP2-NEXT: movvs r0, #0 -; VFP2-NEXT: vpop {d8} -; VFP2-NEXT: pop {r7, pc} +; VFP2-NEXT: movvs r1, #0 +; VFP2-NEXT: pop {r4, r5, r7, pc} ; VFP2-NEXT: .p2align 3 ; VFP2-NEXT: @ %bb.1: ; VFP2-NEXT: .LCPI16_0: @@ -2074,27 +1993,18 @@ define i64 @test_signed_i64_f64(double %f) nounwind { ; VFP2-NEXT: vldr d18, .LCPI17_1 ; VFP2-NEXT: vcmp.f64 d17, d16 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt +; VFP2-NEXT: itt lt ; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vcmp.f64 d17, d18 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vcmp.f64 d17, d17 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r0, #0 -; VFP2-NEXT: vcmp.f64 d17, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt ; VFP2-NEXT: movlt.w r1, #-2147483648 ; VFP2-NEXT: vcmp.f64 d17, d18 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt +; VFP2-NEXT: itt gt ; VFP2-NEXT: mvngt r1, #-2147483648 +; VFP2-NEXT: movgt.w r0, #-1 ; VFP2-NEXT: vcmp.f64 d17, d17 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs +; VFP2-NEXT: itt vs +; VFP2-NEXT: movvs r0, #0 ; VFP2-NEXT: movvs r1, #0 ; VFP2-NEXT: pop {r4, r5, r7, pc} ; VFP2-NEXT: .p2align 3 @@ -2118,27 +2028,18 @@ define i64 @test_signed_i64_f64(double %f) nounwind { ; FP16-NEXT: vldr d2, .LCPI17_1 ; FP16-NEXT: vcmp.f64 d1, d0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt +; FP16-NEXT: itt lt ; FP16-NEXT: movlt r0, #0 -; FP16-NEXT: vcmp.f64 d1, d2 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vcmp.f64 d1, d1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs -; FP16-NEXT: movvs r0, #0 -; FP16-NEXT: vcmp.f64 d1, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: movlt.w r1, #-2147483648 ; FP16-NEXT: vcmp.f64 d1, d2 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt +; FP16-NEXT: itt gt ; FP16-NEXT: mvngt r1, #-2147483648 +; FP16-NEXT: movgt.w r0, #-1 ; FP16-NEXT: vcmp.f64 d1, d1 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs +; FP16-NEXT: itt vs +; FP16-NEXT: movvs r0, #0 ; FP16-NEXT: movvs r1, #0 ; FP16-NEXT: pop {r4, r5, r7, pc} ; FP16-NEXT: .p2align 3 @@ -2287,51 +2188,24 @@ define i100 @test_signed_i100_f64(double %f) nounwind { ; VFP2-NEXT: vldr d18, .LCPI18_1 ; VFP2-NEXT: vcmp.f64 d17, d16 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt +; VFP2-NEXT: itttt lt ; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vcmp.f64 d17, d18 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vcmp.f64 d17, d17 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r0, #0 -; VFP2-NEXT: vcmp.f64 d17, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt ; VFP2-NEXT: movlt r1, #0 -; VFP2-NEXT: vcmp.f64 d17, d18 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r1, #-1 -; VFP2-NEXT: vcmp.f64 d17, d17 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r1, #0 -; VFP2-NEXT: vcmp.f64 d17, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt ; VFP2-NEXT: movlt r2, #0 -; VFP2-NEXT: vcmp.f64 d17, d18 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r2, #-1 -; VFP2-NEXT: vcmp.f64 d17, d17 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r2, #0 -; VFP2-NEXT: vcmp.f64 d17, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt ; VFP2-NEXT: mvnlt r3, #7 ; VFP2-NEXT: vcmp.f64 d17, d18 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt +; VFP2-NEXT: itttt gt ; VFP2-NEXT: movgt r3, #7 +; VFP2-NEXT: movgt.w r2, #-1 +; VFP2-NEXT: movgt.w r1, #-1 +; VFP2-NEXT: movgt.w r0, #-1 ; VFP2-NEXT: vcmp.f64 d17, d17 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs +; VFP2-NEXT: itttt vs +; VFP2-NEXT: movvs r0, #0 +; VFP2-NEXT: movvs r1, #0 +; VFP2-NEXT: movvs r2, #0 ; VFP2-NEXT: movvs r3, #0 ; VFP2-NEXT: pop {r4, r5, r7, pc} ; VFP2-NEXT: .p2align 3 @@ -2350,56 +2224,29 @@ define i100 @test_signed_i100_f64(double %f) nounwind { ; FP16-NEXT: mov r4, r1 ; FP16-NEXT: mov r5, r0 ; FP16-NEXT: bl __fixdfti -; FP16-NEXT: vldr d2, .LCPI18_0 -; FP16-NEXT: vmov d0, r5, r4 -; FP16-NEXT: vldr d1, .LCPI18_1 -; FP16-NEXT: vcmp.f64 d0, d2 +; FP16-NEXT: vldr d0, .LCPI18_0 +; FP16-NEXT: vmov d1, r5, r4 +; FP16-NEXT: vldr d2, .LCPI18_1 +; FP16-NEXT: vcmp.f64 d1, d0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt +; FP16-NEXT: itttt lt ; FP16-NEXT: movlt r0, #0 -; FP16-NEXT: vcmp.f64 d0, d1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vcmp.f64 d0, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs -; FP16-NEXT: movvs r0, #0 -; FP16-NEXT: vcmp.f64 d0, d2 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: movlt r1, #0 -; FP16-NEXT: vcmp.f64 d0, d1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r1, #-1 -; FP16-NEXT: vcmp.f64 d0, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs -; FP16-NEXT: movvs r1, #0 -; FP16-NEXT: vcmp.f64 d0, d2 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: movlt r2, #0 -; FP16-NEXT: vcmp.f64 d0, d1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r2, #-1 -; FP16-NEXT: vcmp.f64 d0, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs -; FP16-NEXT: movvs r2, #0 -; FP16-NEXT: vcmp.f64 d0, d2 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: mvnlt r3, #7 -; FP16-NEXT: vcmp.f64 d0, d1 +; FP16-NEXT: vcmp.f64 d1, d2 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt +; FP16-NEXT: itttt gt ; FP16-NEXT: movgt r3, #7 -; FP16-NEXT: vcmp.f64 d0, d0 +; FP16-NEXT: movgt.w r2, #-1 +; FP16-NEXT: movgt.w r1, #-1 +; FP16-NEXT: movgt.w r0, #-1 +; FP16-NEXT: vcmp.f64 d1, d1 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs +; FP16-NEXT: itttt vs +; FP16-NEXT: movvs r0, #0 +; FP16-NEXT: movvs r1, #0 +; FP16-NEXT: movvs r2, #0 ; FP16-NEXT: movvs r3, #0 ; FP16-NEXT: pop {r4, r5, r7, pc} ; FP16-NEXT: .p2align 3 @@ -2550,51 +2397,24 @@ define i128 @test_signed_i128_f64(double %f) nounwind { ; VFP2-NEXT: vldr d18, .LCPI19_1 ; VFP2-NEXT: vcmp.f64 d17, d16 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt +; VFP2-NEXT: itttt lt ; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vcmp.f64 d17, d18 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vcmp.f64 d17, d17 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r0, #0 -; VFP2-NEXT: vcmp.f64 d17, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt ; VFP2-NEXT: movlt r1, #0 +; VFP2-NEXT: movlt r2, #0 +; VFP2-NEXT: movlt.w r3, #-2147483648 ; VFP2-NEXT: vcmp.f64 d17, d18 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt +; VFP2-NEXT: itttt gt +; VFP2-NEXT: mvngt r3, #-2147483648 +; VFP2-NEXT: movgt.w r2, #-1 ; VFP2-NEXT: movgt.w r1, #-1 +; VFP2-NEXT: movgt.w r0, #-1 ; VFP2-NEXT: vcmp.f64 d17, d17 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs +; VFP2-NEXT: itttt vs +; VFP2-NEXT: movvs r0, #0 ; VFP2-NEXT: movvs r1, #0 -; VFP2-NEXT: vcmp.f64 d17, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r2, #0 -; VFP2-NEXT: vcmp.f64 d17, d18 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r2, #-1 -; VFP2-NEXT: vcmp.f64 d17, d17 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs ; VFP2-NEXT: movvs r2, #0 -; VFP2-NEXT: vcmp.f64 d17, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt.w r3, #-2147483648 -; VFP2-NEXT: vcmp.f64 d17, d18 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt -; VFP2-NEXT: mvngt r3, #-2147483648 -; VFP2-NEXT: vcmp.f64 d17, d17 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs ; VFP2-NEXT: movvs r3, #0 ; VFP2-NEXT: pop {r4, r5, r7, pc} ; VFP2-NEXT: .p2align 3 @@ -2613,56 +2433,29 @@ define i128 @test_signed_i128_f64(double %f) nounwind { ; FP16-NEXT: mov r4, r1 ; FP16-NEXT: mov r5, r0 ; FP16-NEXT: bl __fixdfti -; FP16-NEXT: vldr d2, .LCPI19_0 -; FP16-NEXT: vmov d0, r5, r4 -; FP16-NEXT: vldr d1, .LCPI19_1 -; FP16-NEXT: vcmp.f64 d0, d2 +; FP16-NEXT: vldr d0, .LCPI19_0 +; FP16-NEXT: vmov d1, r5, r4 +; FP16-NEXT: vldr d2, .LCPI19_1 +; FP16-NEXT: vcmp.f64 d1, d0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt +; FP16-NEXT: itttt lt ; FP16-NEXT: movlt r0, #0 -; FP16-NEXT: vcmp.f64 d0, d1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vcmp.f64 d0, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs -; FP16-NEXT: movvs r0, #0 -; FP16-NEXT: vcmp.f64 d0, d2 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: movlt r1, #0 -; FP16-NEXT: vcmp.f64 d0, d1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r1, #-1 -; FP16-NEXT: vcmp.f64 d0, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs -; FP16-NEXT: movvs r1, #0 -; FP16-NEXT: vcmp.f64 d0, d2 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: movlt r2, #0 -; FP16-NEXT: vcmp.f64 d0, d1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r2, #-1 -; FP16-NEXT: vcmp.f64 d0, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs -; FP16-NEXT: movvs r2, #0 -; FP16-NEXT: vcmp.f64 d0, d2 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: movlt.w r3, #-2147483648 -; FP16-NEXT: vcmp.f64 d0, d1 +; FP16-NEXT: vcmp.f64 d1, d2 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt +; FP16-NEXT: itttt gt ; FP16-NEXT: mvngt r3, #-2147483648 -; FP16-NEXT: vcmp.f64 d0, d0 +; FP16-NEXT: movgt.w r2, #-1 +; FP16-NEXT: movgt.w r1, #-1 +; FP16-NEXT: movgt.w r0, #-1 +; FP16-NEXT: vcmp.f64 d1, d1 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs +; FP16-NEXT: itttt vs +; FP16-NEXT: movvs r0, #0 +; FP16-NEXT: movvs r1, #0 +; FP16-NEXT: movvs r2, #0 ; FP16-NEXT: movvs r3, #0 ; FP16-NEXT: pop {r4, r5, r7, pc} ; FP16-NEXT: .p2align 3 @@ -2949,11 +2742,11 @@ define i13 @test_signed_i13_f16(half %f) nounwind { ; VFP2-NEXT: bl __aeabi_h2f ; VFP2-NEXT: vmov s0, r0 ; VFP2-NEXT: vldr s2, .LCPI22_0 +; VFP2-NEXT: vldr s6, .LCPI22_1 ; VFP2-NEXT: vcvt.s32.f32 s4, s0 ; VFP2-NEXT: vcmp.f32 s0, s2 -; VFP2-NEXT: vldr s2, .LCPI22_1 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s0, s2 +; VFP2-NEXT: vcmp.f32 s0, s6 ; VFP2-NEXT: vmov r0, s4 ; VFP2-NEXT: itt lt ; VFP2-NEXT: movwlt r0, #61440 @@ -3055,11 +2848,11 @@ define i16 @test_signed_i16_f16(half %f) nounwind { ; VFP2-NEXT: bl __aeabi_h2f ; VFP2-NEXT: vmov s0, r0 ; VFP2-NEXT: vldr s2, .LCPI23_0 +; VFP2-NEXT: vldr s6, .LCPI23_1 ; VFP2-NEXT: vcvt.s32.f32 s4, s0 ; VFP2-NEXT: vcmp.f32 s0, s2 -; VFP2-NEXT: vldr s2, .LCPI23_1 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s0, s2 +; VFP2-NEXT: vcmp.f32 s0, s6 ; VFP2-NEXT: vmov r0, s4 ; VFP2-NEXT: itt lt ; VFP2-NEXT: movwlt r0, #32768 @@ -3161,11 +2954,11 @@ define i19 @test_signed_i19_f16(half %f) nounwind { ; VFP2-NEXT: bl __aeabi_h2f ; VFP2-NEXT: vmov s0, r0 ; VFP2-NEXT: vldr s2, .LCPI24_0 +; VFP2-NEXT: vldr s6, .LCPI24_1 ; VFP2-NEXT: vcvt.s32.f32 s4, s0 ; VFP2-NEXT: vcmp.f32 s0, s2 -; VFP2-NEXT: vldr s2, .LCPI24_1 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s0, s2 +; VFP2-NEXT: vcmp.f32 s0, s6 ; VFP2-NEXT: vmov r0, s4 ; VFP2-NEXT: itt lt ; VFP2-NEXT: movlt r0, #0 @@ -3357,40 +3150,32 @@ define i50 @test_signed_i50_f16(half %f) nounwind { ; ; VFP2-LABEL: test_signed_i50_f16: ; VFP2: @ %bb.0: -; VFP2-NEXT: .save {r7, lr} -; VFP2-NEXT: push {r7, lr} -; VFP2-NEXT: .vsave {d8} -; VFP2-NEXT: vpush {d8} +; VFP2-NEXT: .save {r4, lr} +; VFP2-NEXT: push {r4, lr} ; VFP2-NEXT: bl __aeabi_h2f -; VFP2-NEXT: vmov s16, r0 +; VFP2-NEXT: mov r4, r0 ; VFP2-NEXT: bl __aeabi_f2lz ; VFP2-NEXT: vldr s0, .LCPI26_0 -; VFP2-NEXT: vldr s2, .LCPI26_1 -; VFP2-NEXT: vcmp.f32 s16, s0 +; VFP2-NEXT: vmov s2, r4 +; VFP2-NEXT: vldr s4, .LCPI26_1 +; VFP2-NEXT: vcmp.f32 s2, s0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, s2 -; VFP2-NEXT: itt lt +; VFP2-NEXT: ittt lt ; VFP2-NEXT: movlt r1, #0 ; VFP2-NEXT: movtlt r1, #65534 +; VFP2-NEXT: movlt r0, #0 +; VFP2-NEXT: vcmp.f32 s2, s4 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, s0 -; VFP2-NEXT: itt gt +; VFP2-NEXT: ittt gt ; VFP2-NEXT: movwgt r1, #65535 ; VFP2-NEXT: movtgt r1, #1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vcmp.f32 s16, s2 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vcmp.f32 s16, s16 +; VFP2-NEXT: vcmp.f32 s2, s2 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr ; VFP2-NEXT: itt vs ; VFP2-NEXT: movvs r0, #0 ; VFP2-NEXT: movvs r1, #0 -; VFP2-NEXT: vpop {d8} -; VFP2-NEXT: pop {r7, pc} +; VFP2-NEXT: pop {r4, pc} ; VFP2-NEXT: .p2align 2 ; VFP2-NEXT: @ %bb.1: ; VFP2-NEXT: .LCPI26_0: @@ -3412,21 +3197,15 @@ define i50 @test_signed_i50_f16(half %f) nounwind { ; FP16-NEXT: vldr s2, .LCPI26_1 ; FP16-NEXT: vcmp.f32 s16, s0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s2 -; FP16-NEXT: itt lt +; FP16-NEXT: ittt lt ; FP16-NEXT: movlt r1, #0 ; FP16-NEXT: movtlt r1, #65534 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: itt gt -; FP16-NEXT: movwgt r1, #65535 -; FP16-NEXT: movtgt r1, #1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: movlt r0, #0 ; FP16-NEXT: vcmp.f32 s16, s2 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt +; FP16-NEXT: ittt gt +; FP16-NEXT: movwgt r1, #65535 +; FP16-NEXT: movtgt r1, #1 ; FP16-NEXT: movgt.w r0, #-1 ; FP16-NEXT: vcmp.f32 s16, s16 ; FP16-NEXT: vmrs APSR_nzcv, fpscr @@ -3531,27 +3310,18 @@ define i64 @test_signed_i64_f16(half %f) nounwind { ; VFP2-NEXT: vldr s4, .LCPI27_1 ; VFP2-NEXT: vcmp.f32 s2, s0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s4 -; VFP2-NEXT: it lt +; VFP2-NEXT: itt lt ; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s2 -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s0 -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r0, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt ; VFP2-NEXT: movlt.w r1, #-2147483648 ; VFP2-NEXT: vcmp.f32 s2, s4 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt +; VFP2-NEXT: itt gt ; VFP2-NEXT: mvngt r1, #-2147483648 +; VFP2-NEXT: movgt.w r0, #-1 ; VFP2-NEXT: vcmp.f32 s2, s2 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs +; VFP2-NEXT: itt vs +; VFP2-NEXT: movvs r0, #0 ; VFP2-NEXT: movvs r1, #0 ; VFP2-NEXT: pop {r4, pc} ; VFP2-NEXT: .p2align 2 @@ -3575,27 +3345,18 @@ define i64 @test_signed_i64_f16(half %f) nounwind { ; FP16-NEXT: vldr s2, .LCPI27_1 ; FP16-NEXT: vcmp.f32 s16, s0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s2 -; FP16-NEXT: it lt +; FP16-NEXT: itt lt ; FP16-NEXT: movlt r0, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s16 -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it vs -; FP16-NEXT: movvs r0, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: movlt.w r1, #-2147483648 ; FP16-NEXT: vcmp.f32 s16, s2 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt +; FP16-NEXT: itt gt ; FP16-NEXT: mvngt r1, #-2147483648 +; FP16-NEXT: movgt.w r0, #-1 ; FP16-NEXT: vcmp.f32 s16, s16 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs +; FP16-NEXT: itt vs +; FP16-NEXT: movvs r0, #0 ; FP16-NEXT: movvs r1, #0 ; FP16-NEXT: vpop {d8} ; FP16-NEXT: pop {r7, pc} @@ -3737,51 +3498,24 @@ define i100 @test_signed_i100_f16(half %f) nounwind { ; VFP2-NEXT: vldr s4, .LCPI28_1 ; VFP2-NEXT: vcmp.f32 s2, s0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s4 -; VFP2-NEXT: it lt +; VFP2-NEXT: itttt lt ; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s2 -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s0 -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r0, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s4 -; VFP2-NEXT: it lt ; VFP2-NEXT: movlt r1, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s2 -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r1, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s0 -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r1, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s4 -; VFP2-NEXT: it lt ; VFP2-NEXT: movlt r2, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s2 -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r2, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s0 -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r2, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt ; VFP2-NEXT: mvnlt r3, #7 ; VFP2-NEXT: vcmp.f32 s2, s4 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt +; VFP2-NEXT: itttt gt ; VFP2-NEXT: movgt r3, #7 +; VFP2-NEXT: movgt.w r2, #-1 +; VFP2-NEXT: movgt.w r1, #-1 +; VFP2-NEXT: movgt.w r0, #-1 ; VFP2-NEXT: vcmp.f32 s2, s2 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs +; VFP2-NEXT: itttt vs +; VFP2-NEXT: movvs r0, #0 +; VFP2-NEXT: movvs r1, #0 +; VFP2-NEXT: movvs r2, #0 ; VFP2-NEXT: movvs r3, #0 ; VFP2-NEXT: pop {r4, pc} ; VFP2-NEXT: .p2align 2 @@ -3805,51 +3539,24 @@ define i100 @test_signed_i100_f16(half %f) nounwind { ; FP16-NEXT: vldr s2, .LCPI28_1 ; FP16-NEXT: vcmp.f32 s16, s0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s2 -; FP16-NEXT: it lt +; FP16-NEXT: itttt lt ; FP16-NEXT: movlt r0, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s16 -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it vs -; FP16-NEXT: movvs r0, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s2 -; FP16-NEXT: it lt ; FP16-NEXT: movlt r1, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s16 -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r1, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it vs -; FP16-NEXT: movvs r1, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s2 -; FP16-NEXT: it lt ; FP16-NEXT: movlt r2, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s16 -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r2, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it vs -; FP16-NEXT: movvs r2, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: mvnlt r3, #7 ; FP16-NEXT: vcmp.f32 s16, s2 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt +; FP16-NEXT: itttt gt ; FP16-NEXT: movgt r3, #7 +; FP16-NEXT: movgt.w r2, #-1 +; FP16-NEXT: movgt.w r1, #-1 +; FP16-NEXT: movgt.w r0, #-1 ; FP16-NEXT: vcmp.f32 s16, s16 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs +; FP16-NEXT: itttt vs +; FP16-NEXT: movvs r0, #0 +; FP16-NEXT: movvs r1, #0 +; FP16-NEXT: movvs r2, #0 ; FP16-NEXT: movvs r3, #0 ; FP16-NEXT: vpop {d8} ; FP16-NEXT: pop {r7, pc} @@ -3994,51 +3701,24 @@ define i128 @test_signed_i128_f16(half %f) nounwind { ; VFP2-NEXT: vldr s4, .LCPI29_1 ; VFP2-NEXT: vcmp.f32 s2, s0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s4 -; VFP2-NEXT: it lt +; VFP2-NEXT: itttt lt ; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s2 -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s0 -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r0, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s4 -; VFP2-NEXT: it lt ; VFP2-NEXT: movlt r1, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s2 -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r1, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s0 -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r1, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s4 -; VFP2-NEXT: it lt ; VFP2-NEXT: movlt r2, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s2 -; VFP2-NEXT: it gt -; VFP2-NEXT: movgt.w r2, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s2, s0 -; VFP2-NEXT: it vs -; VFP2-NEXT: movvs r2, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt ; VFP2-NEXT: movlt.w r3, #-2147483648 ; VFP2-NEXT: vcmp.f32 s2, s4 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt +; VFP2-NEXT: itttt gt ; VFP2-NEXT: mvngt r3, #-2147483648 +; VFP2-NEXT: movgt.w r2, #-1 +; VFP2-NEXT: movgt.w r1, #-1 +; VFP2-NEXT: movgt.w r0, #-1 ; VFP2-NEXT: vcmp.f32 s2, s2 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it vs +; VFP2-NEXT: itttt vs +; VFP2-NEXT: movvs r0, #0 +; VFP2-NEXT: movvs r1, #0 +; VFP2-NEXT: movvs r2, #0 ; VFP2-NEXT: movvs r3, #0 ; VFP2-NEXT: pop {r4, pc} ; VFP2-NEXT: .p2align 2 @@ -4062,51 +3742,24 @@ define i128 @test_signed_i128_f16(half %f) nounwind { ; FP16-NEXT: vldr s2, .LCPI29_1 ; FP16-NEXT: vcmp.f32 s16, s0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s2 -; FP16-NEXT: it lt +; FP16-NEXT: itttt lt ; FP16-NEXT: movlt r0, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s16 -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it vs -; FP16-NEXT: movvs r0, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s2 -; FP16-NEXT: it lt ; FP16-NEXT: movlt r1, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s16 -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r1, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it vs -; FP16-NEXT: movvs r1, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s2 -; FP16-NEXT: it lt ; FP16-NEXT: movlt r2, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s16 -; FP16-NEXT: it gt -; FP16-NEXT: movgt.w r2, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it vs -; FP16-NEXT: movvs r2, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt ; FP16-NEXT: movlt.w r3, #-2147483648 ; FP16-NEXT: vcmp.f32 s16, s2 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt +; FP16-NEXT: itttt gt ; FP16-NEXT: mvngt r3, #-2147483648 +; FP16-NEXT: movgt.w r2, #-1 +; FP16-NEXT: movgt.w r1, #-1 +; FP16-NEXT: movgt.w r0, #-1 ; FP16-NEXT: vcmp.f32 s16, s16 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it vs +; FP16-NEXT: itttt vs +; FP16-NEXT: movvs r0, #0 +; FP16-NEXT: movvs r1, #0 +; FP16-NEXT: movvs r2, #0 ; FP16-NEXT: movvs r3, #0 ; FP16-NEXT: vpop {d8} ; FP16-NEXT: pop {r7, pc} diff --git a/llvm/test/CodeGen/ARM/fptoui-sat-scalar.ll b/llvm/test/CodeGen/ARM/fptoui-sat-scalar.ll index 3438fb113015c..14eb67104edda 100644 --- a/llvm/test/CodeGen/ARM/fptoui-sat-scalar.ll +++ b/llvm/test/CodeGen/ARM/fptoui-sat-scalar.ll @@ -503,8 +503,8 @@ define i50 @test_signed_i50_f32(float %f) nounwind { ; VFP-NEXT: vcmp.f32 s16, #0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr ; VFP-NEXT: itt lt -; VFP-NEXT: movlt r0, #0 ; VFP-NEXT: movlt r1, #0 +; VFP-NEXT: movlt r0, #0 ; VFP-NEXT: vcmp.f32 s16, s0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr ; VFP-NEXT: ittt gt @@ -586,19 +586,13 @@ define i64 @test_signed_i64_f32(float %f) nounwind { ; VFP-NEXT: vldr s0, .LCPI7_0 ; VFP-NEXT: vcmp.f32 s16, #0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it lt +; VFP-NEXT: itt lt +; VFP-NEXT: movlt r1, #0 ; VFP-NEXT: movlt r0, #0 ; VFP-NEXT: vcmp.f32 s16, s0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, #0 -; VFP-NEXT: it gt +; VFP-NEXT: itt gt ; VFP-NEXT: movgt.w r0, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it lt -; VFP-NEXT: movlt r1, #0 -; VFP-NEXT: vcmp.f32 s16, s0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it gt ; VFP-NEXT: movgt.w r1, #-1 ; VFP-NEXT: vpop {d8} ; VFP-NEXT: pop {r7, pc} @@ -701,35 +695,17 @@ define i100 @test_signed_i100_f32(float %f) nounwind { ; VFP-NEXT: vldr s0, .LCPI8_0 ; VFP-NEXT: vcmp.f32 s16, #0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it lt +; VFP-NEXT: itttt lt +; VFP-NEXT: movlt r3, #0 +; VFP-NEXT: movlt r2, #0 +; VFP-NEXT: movlt r1, #0 ; VFP-NEXT: movlt r0, #0 ; VFP-NEXT: vcmp.f32 s16, s0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, #0 -; VFP-NEXT: it gt +; VFP-NEXT: itttt gt ; VFP-NEXT: movgt.w r0, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, s0 -; VFP-NEXT: it lt -; VFP-NEXT: movlt r1, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, #0 -; VFP-NEXT: it gt ; VFP-NEXT: movgt.w r1, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, s0 -; VFP-NEXT: it lt -; VFP-NEXT: movlt r2, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, #0 -; VFP-NEXT: it gt ; VFP-NEXT: movgt.w r2, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it lt -; VFP-NEXT: movlt r3, #0 -; VFP-NEXT: vcmp.f32 s16, s0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it gt ; VFP-NEXT: movgt r3, #15 ; VFP-NEXT: vpop {d8} ; VFP-NEXT: pop {r7, pc} @@ -831,35 +807,17 @@ define i128 @test_signed_i128_f32(float %f) nounwind { ; VFP-NEXT: vldr s0, .LCPI9_0 ; VFP-NEXT: vcmp.f32 s16, #0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it lt +; VFP-NEXT: itttt lt +; VFP-NEXT: movlt r3, #0 +; VFP-NEXT: movlt r2, #0 +; VFP-NEXT: movlt r1, #0 ; VFP-NEXT: movlt r0, #0 ; VFP-NEXT: vcmp.f32 s16, s0 ; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, #0 -; VFP-NEXT: it gt +; VFP-NEXT: itttt gt ; VFP-NEXT: movgt.w r0, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, s0 -; VFP-NEXT: it lt -; VFP-NEXT: movlt r1, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, #0 -; VFP-NEXT: it gt ; VFP-NEXT: movgt.w r1, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, s0 -; VFP-NEXT: it lt -; VFP-NEXT: movlt r2, #0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: vcmp.f32 s16, #0 -; VFP-NEXT: it gt ; VFP-NEXT: movgt.w r2, #-1 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it lt -; VFP-NEXT: movlt r3, #0 -; VFP-NEXT: vcmp.f32 s16, s0 -; VFP-NEXT: vmrs APSR_nzcv, fpscr -; VFP-NEXT: it gt ; VFP-NEXT: movgt.w r3, #-1 ; VFP-NEXT: vpop {d8} ; VFP-NEXT: pop {r7, pc} @@ -1453,8 +1411,8 @@ define i50 @test_signed_i50_f64(double %f) nounwind { ; VFP2-NEXT: vldr d16, .LCPI16_0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr ; VFP2-NEXT: itt lt -; VFP2-NEXT: movlt r0, #0 ; VFP2-NEXT: movlt r1, #0 +; VFP2-NEXT: movlt r0, #0 ; VFP2-NEXT: vcmp.f64 d8, d16 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr ; VFP2-NEXT: ittt gt @@ -1564,19 +1522,13 @@ define i64 @test_signed_i64_f64(double %f) nounwind { ; VFP2-NEXT: vcmp.f64 d8, #0 ; VFP2-NEXT: vldr d16, .LCPI17_0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt +; VFP2-NEXT: itt lt +; VFP2-NEXT: movlt r1, #0 ; VFP2-NEXT: movlt r0, #0 ; VFP2-NEXT: vcmp.f64 d8, d16 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt +; VFP2-NEXT: itt gt ; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vcmp.f64 d8, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r1, #0 -; VFP2-NEXT: vcmp.f64 d8, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r1, #-1 ; VFP2-NEXT: vpop {d8} ; VFP2-NEXT: pop {r7, pc} @@ -1597,19 +1549,13 @@ define i64 @test_signed_i64_f64(double %f) nounwind { ; FP16-NEXT: vcmp.f64 d8, #0 ; FP16-NEXT: vldr d0, .LCPI17_0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt +; FP16-NEXT: itt lt +; FP16-NEXT: movlt r1, #0 ; FP16-NEXT: movlt r0, #0 ; FP16-NEXT: vcmp.f64 d8, d0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt +; FP16-NEXT: itt gt ; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vcmp.f64 d8, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt -; FP16-NEXT: movlt r1, #0 -; FP16-NEXT: vcmp.f64 d8, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r1, #-1 ; FP16-NEXT: vpop {d8} ; FP16-NEXT: pop {r7, pc} @@ -1719,35 +1665,17 @@ define i100 @test_signed_i100_f64(double %f) nounwind { ; VFP2-NEXT: vcmp.f64 d8, #0 ; VFP2-NEXT: vldr d16, .LCPI18_0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt +; VFP2-NEXT: itttt lt +; VFP2-NEXT: movlt r3, #0 +; VFP2-NEXT: movlt r2, #0 +; VFP2-NEXT: movlt r1, #0 ; VFP2-NEXT: movlt r0, #0 ; VFP2-NEXT: vcmp.f64 d8, d16 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt +; VFP2-NEXT: itttt gt ; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vcmp.f64 d8, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r1, #0 -; VFP2-NEXT: vcmp.f64 d8, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r1, #-1 -; VFP2-NEXT: vcmp.f64 d8, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r2, #0 -; VFP2-NEXT: vcmp.f64 d8, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r2, #-1 -; VFP2-NEXT: vcmp.f64 d8, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r3, #0 -; VFP2-NEXT: vcmp.f64 d8, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt r3, #15 ; VFP2-NEXT: vpop {d8} ; VFP2-NEXT: pop {r7, pc} @@ -1768,35 +1696,17 @@ define i100 @test_signed_i100_f64(double %f) nounwind { ; FP16-NEXT: vcmp.f64 d8, #0 ; FP16-NEXT: vldr d0, .LCPI18_0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt +; FP16-NEXT: itttt lt +; FP16-NEXT: movlt r3, #0 +; FP16-NEXT: movlt r2, #0 +; FP16-NEXT: movlt r1, #0 ; FP16-NEXT: movlt r0, #0 ; FP16-NEXT: vcmp.f64 d8, d0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt +; FP16-NEXT: itttt gt ; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vcmp.f64 d8, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt -; FP16-NEXT: movlt r1, #0 -; FP16-NEXT: vcmp.f64 d8, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r1, #-1 -; FP16-NEXT: vcmp.f64 d8, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt -; FP16-NEXT: movlt r2, #0 -; FP16-NEXT: vcmp.f64 d8, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r2, #-1 -; FP16-NEXT: vcmp.f64 d8, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt -; FP16-NEXT: movlt r3, #0 -; FP16-NEXT: vcmp.f64 d8, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt ; FP16-NEXT: movgt r3, #15 ; FP16-NEXT: vpop {d8} ; FP16-NEXT: pop {r7, pc} @@ -1902,35 +1812,17 @@ define i128 @test_signed_i128_f64(double %f) nounwind { ; VFP2-NEXT: vcmp.f64 d8, #0 ; VFP2-NEXT: vldr d16, .LCPI19_0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt +; VFP2-NEXT: itttt lt +; VFP2-NEXT: movlt r3, #0 +; VFP2-NEXT: movlt r2, #0 +; VFP2-NEXT: movlt r1, #0 ; VFP2-NEXT: movlt r0, #0 ; VFP2-NEXT: vcmp.f64 d8, d16 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt +; VFP2-NEXT: itttt gt ; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vcmp.f64 d8, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r1, #0 -; VFP2-NEXT: vcmp.f64 d8, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r1, #-1 -; VFP2-NEXT: vcmp.f64 d8, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r2, #0 -; VFP2-NEXT: vcmp.f64 d8, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r2, #-1 -; VFP2-NEXT: vcmp.f64 d8, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r3, #0 -; VFP2-NEXT: vcmp.f64 d8, d16 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r3, #-1 ; VFP2-NEXT: vpop {d8} ; VFP2-NEXT: pop {r7, pc} @@ -1951,35 +1843,17 @@ define i128 @test_signed_i128_f64(double %f) nounwind { ; FP16-NEXT: vcmp.f64 d8, #0 ; FP16-NEXT: vldr d0, .LCPI19_0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt +; FP16-NEXT: itttt lt +; FP16-NEXT: movlt r3, #0 +; FP16-NEXT: movlt r2, #0 +; FP16-NEXT: movlt r1, #0 ; FP16-NEXT: movlt r0, #0 ; FP16-NEXT: vcmp.f64 d8, d0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt +; FP16-NEXT: itttt gt ; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vcmp.f64 d8, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt -; FP16-NEXT: movlt r1, #0 -; FP16-NEXT: vcmp.f64 d8, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r1, #-1 -; FP16-NEXT: vcmp.f64 d8, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt -; FP16-NEXT: movlt r2, #0 -; FP16-NEXT: vcmp.f64 d8, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r2, #-1 -; FP16-NEXT: vcmp.f64 d8, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt -; FP16-NEXT: movlt r3, #0 -; FP16-NEXT: vcmp.f64 d8, d0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r3, #-1 ; FP16-NEXT: vpop {d8} ; FP16-NEXT: pop {r7, pc} @@ -2526,27 +2400,25 @@ define i50 @test_signed_i50_f16(half %f) nounwind { ; ; VFP2-LABEL: test_signed_i50_f16: ; VFP2: @ %bb.0: -; VFP2-NEXT: .save {r7, lr} -; VFP2-NEXT: push {r7, lr} -; VFP2-NEXT: .vsave {d8} -; VFP2-NEXT: vpush {d8} +; VFP2-NEXT: .save {r4, lr} +; VFP2-NEXT: push {r4, lr} ; VFP2-NEXT: bl __aeabi_h2f -; VFP2-NEXT: vmov s16, r0 +; VFP2-NEXT: mov r4, r0 ; VFP2-NEXT: bl __aeabi_f2ulz -; VFP2-NEXT: vldr s0, .LCPI26_0 -; VFP2-NEXT: vcmp.f32 s16, #0 +; VFP2-NEXT: vmov s0, r4 +; VFP2-NEXT: vldr s2, .LCPI26_0 +; VFP2-NEXT: vcmp.f32 s0, #0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr ; VFP2-NEXT: itt lt -; VFP2-NEXT: movlt r0, #0 ; VFP2-NEXT: movlt r1, #0 -; VFP2-NEXT: vcmp.f32 s16, s0 +; VFP2-NEXT: movlt r0, #0 +; VFP2-NEXT: vcmp.f32 s0, s2 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr ; VFP2-NEXT: ittt gt ; VFP2-NEXT: movwgt r1, #65535 ; VFP2-NEXT: movtgt r1, #3 ; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vpop {d8} -; VFP2-NEXT: pop {r7, pc} +; VFP2-NEXT: pop {r4, pc} ; VFP2-NEXT: .p2align 2 ; VFP2-NEXT: @ %bb.1: ; VFP2-NEXT: .LCPI26_0: @@ -2566,8 +2438,8 @@ define i50 @test_signed_i50_f16(half %f) nounwind { ; FP16-NEXT: vcmp.f32 s16, #0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr ; FP16-NEXT: itt lt -; FP16-NEXT: movlt r0, #0 ; FP16-NEXT: movlt r1, #0 +; FP16-NEXT: movlt r0, #0 ; FP16-NEXT: vcmp.f32 s16, s0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr ; FP16-NEXT: ittt gt @@ -2642,32 +2514,24 @@ define i64 @test_signed_i64_f16(half %f) nounwind { ; ; VFP2-LABEL: test_signed_i64_f16: ; VFP2: @ %bb.0: -; VFP2-NEXT: .save {r7, lr} -; VFP2-NEXT: push {r7, lr} -; VFP2-NEXT: .vsave {d8} -; VFP2-NEXT: vpush {d8} +; VFP2-NEXT: .save {r4, lr} +; VFP2-NEXT: push {r4, lr} ; VFP2-NEXT: bl __aeabi_h2f -; VFP2-NEXT: vmov s16, r0 +; VFP2-NEXT: mov r4, r0 ; VFP2-NEXT: bl __aeabi_f2ulz -; VFP2-NEXT: vldr s0, .LCPI27_0 -; VFP2-NEXT: vcmp.f32 s16, #0 +; VFP2-NEXT: vmov s0, r4 +; VFP2-NEXT: vldr s2, .LCPI27_0 +; VFP2-NEXT: vcmp.f32 s0, #0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt +; VFP2-NEXT: itt lt +; VFP2-NEXT: movlt r1, #0 ; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vcmp.f32 s16, s0 +; VFP2-NEXT: vcmp.f32 s0, s2 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, #0 -; VFP2-NEXT: it gt +; VFP2-NEXT: itt gt ; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r1, #0 -; VFP2-NEXT: vcmp.f32 s16, s0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r1, #-1 -; VFP2-NEXT: vpop {d8} -; VFP2-NEXT: pop {r7, pc} +; VFP2-NEXT: pop {r4, pc} ; VFP2-NEXT: .p2align 2 ; VFP2-NEXT: @ %bb.1: ; VFP2-NEXT: .LCPI27_0: @@ -2686,19 +2550,13 @@ define i64 @test_signed_i64_f16(half %f) nounwind { ; FP16-NEXT: vldr s0, .LCPI27_0 ; FP16-NEXT: vcmp.f32 s16, #0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt +; FP16-NEXT: itt lt +; FP16-NEXT: movlt r1, #0 ; FP16-NEXT: movlt r0, #0 ; FP16-NEXT: vcmp.f32 s16, s0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, #0 -; FP16-NEXT: it gt +; FP16-NEXT: itt gt ; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt -; FP16-NEXT: movlt r1, #0 -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r1, #-1 ; FP16-NEXT: vpop {d8} ; FP16-NEXT: pop {r7, pc} @@ -2794,48 +2652,28 @@ define i100 @test_signed_i100_f16(half %f) nounwind { ; ; VFP2-LABEL: test_signed_i100_f16: ; VFP2: @ %bb.0: -; VFP2-NEXT: .save {r7, lr} -; VFP2-NEXT: push {r7, lr} -; VFP2-NEXT: .vsave {d8} -; VFP2-NEXT: vpush {d8} +; VFP2-NEXT: .save {r4, lr} +; VFP2-NEXT: push {r4, lr} ; VFP2-NEXT: bl __aeabi_h2f -; VFP2-NEXT: vmov s16, r0 +; VFP2-NEXT: mov r4, r0 ; VFP2-NEXT: bl __fixunssfti -; VFP2-NEXT: vldr s0, .LCPI28_0 -; VFP2-NEXT: vcmp.f32 s16, #0 +; VFP2-NEXT: vmov s0, r4 +; VFP2-NEXT: vldr s2, .LCPI28_0 +; VFP2-NEXT: vcmp.f32 s0, #0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt +; VFP2-NEXT: itttt lt +; VFP2-NEXT: movlt r3, #0 +; VFP2-NEXT: movlt r2, #0 +; VFP2-NEXT: movlt r1, #0 ; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vcmp.f32 s16, s0 +; VFP2-NEXT: vcmp.f32 s0, s2 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, #0 -; VFP2-NEXT: it gt +; VFP2-NEXT: itttt gt ; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, s0 -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r1, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, #0 -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r1, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, s0 -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r2, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, #0 -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r2, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r3, #0 -; VFP2-NEXT: vcmp.f32 s16, s0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt r3, #15 -; VFP2-NEXT: vpop {d8} -; VFP2-NEXT: pop {r7, pc} +; VFP2-NEXT: pop {r4, pc} ; VFP2-NEXT: .p2align 2 ; VFP2-NEXT: @ %bb.1: ; VFP2-NEXT: .LCPI28_0: @@ -2854,35 +2692,17 @@ define i100 @test_signed_i100_f16(half %f) nounwind { ; FP16-NEXT: vldr s0, .LCPI28_0 ; FP16-NEXT: vcmp.f32 s16, #0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt +; FP16-NEXT: itttt lt +; FP16-NEXT: movlt r3, #0 +; FP16-NEXT: movlt r2, #0 +; FP16-NEXT: movlt r1, #0 ; FP16-NEXT: movlt r0, #0 ; FP16-NEXT: vcmp.f32 s16, s0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, #0 -; FP16-NEXT: it gt +; FP16-NEXT: itttt gt ; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it lt -; FP16-NEXT: movlt r1, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, #0 -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r1, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it lt -; FP16-NEXT: movlt r2, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, #0 -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r2, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt -; FP16-NEXT: movlt r3, #0 -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt ; FP16-NEXT: movgt r3, #15 ; FP16-NEXT: vpop {d8} ; FP16-NEXT: pop {r7, pc} @@ -2977,48 +2797,28 @@ define i128 @test_signed_i128_f16(half %f) nounwind { ; ; VFP2-LABEL: test_signed_i128_f16: ; VFP2: @ %bb.0: -; VFP2-NEXT: .save {r7, lr} -; VFP2-NEXT: push {r7, lr} -; VFP2-NEXT: .vsave {d8} -; VFP2-NEXT: vpush {d8} +; VFP2-NEXT: .save {r4, lr} +; VFP2-NEXT: push {r4, lr} ; VFP2-NEXT: bl __aeabi_h2f -; VFP2-NEXT: vmov s16, r0 +; VFP2-NEXT: mov r4, r0 ; VFP2-NEXT: bl __fixunssfti -; VFP2-NEXT: vldr s0, .LCPI29_0 -; VFP2-NEXT: vcmp.f32 s16, #0 +; VFP2-NEXT: vmov s0, r4 +; VFP2-NEXT: vldr s2, .LCPI29_0 +; VFP2-NEXT: vcmp.f32 s0, #0 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt +; VFP2-NEXT: itttt lt +; VFP2-NEXT: movlt r3, #0 +; VFP2-NEXT: movlt r2, #0 +; VFP2-NEXT: movlt r1, #0 ; VFP2-NEXT: movlt r0, #0 -; VFP2-NEXT: vcmp.f32 s16, s0 +; VFP2-NEXT: vcmp.f32 s0, s2 ; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, #0 -; VFP2-NEXT: it gt +; VFP2-NEXT: itttt gt ; VFP2-NEXT: movgt.w r0, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, s0 -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r1, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, #0 -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r1, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, s0 -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r2, #0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: vcmp.f32 s16, #0 -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r2, #-1 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it lt -; VFP2-NEXT: movlt r3, #0 -; VFP2-NEXT: vcmp.f32 s16, s0 -; VFP2-NEXT: vmrs APSR_nzcv, fpscr -; VFP2-NEXT: it gt ; VFP2-NEXT: movgt.w r3, #-1 -; VFP2-NEXT: vpop {d8} -; VFP2-NEXT: pop {r7, pc} +; VFP2-NEXT: pop {r4, pc} ; VFP2-NEXT: .p2align 2 ; VFP2-NEXT: @ %bb.1: ; VFP2-NEXT: .LCPI29_0: @@ -3037,35 +2837,17 @@ define i128 @test_signed_i128_f16(half %f) nounwind { ; FP16-NEXT: vldr s0, .LCPI29_0 ; FP16-NEXT: vcmp.f32 s16, #0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt +; FP16-NEXT: itttt lt +; FP16-NEXT: movlt r3, #0 +; FP16-NEXT: movlt r2, #0 +; FP16-NEXT: movlt r1, #0 ; FP16-NEXT: movlt r0, #0 ; FP16-NEXT: vcmp.f32 s16, s0 ; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, #0 -; FP16-NEXT: it gt +; FP16-NEXT: itttt gt ; FP16-NEXT: movgt.w r0, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it lt -; FP16-NEXT: movlt r1, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, #0 -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r1, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: it lt -; FP16-NEXT: movlt r2, #0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: vcmp.f32 s16, #0 -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r2, #-1 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it lt -; FP16-NEXT: movlt r3, #0 -; FP16-NEXT: vcmp.f32 s16, s0 -; FP16-NEXT: vmrs APSR_nzcv, fpscr -; FP16-NEXT: it gt ; FP16-NEXT: movgt.w r3, #-1 ; FP16-NEXT: vpop {d8} ; FP16-NEXT: pop {r7, pc} diff --git a/llvm/test/CodeGen/ARM/minnum-maxnum-intrinsics.ll b/llvm/test/CodeGen/ARM/minnum-maxnum-intrinsics.ll index feb790821e875..4003af5d44be8 100644 --- a/llvm/test/CodeGen/ARM/minnum-maxnum-intrinsics.ll +++ b/llvm/test/CodeGen/ARM/minnum-maxnum-intrinsics.ll @@ -16,12 +16,12 @@ declare <2 x double> @llvm.maxnum.v2f64(<2 x double>, <2 x double>) define float @fminnum32_intrinsic(float %x, float %y) { ; ARMV7-LABEL: fminnum32_intrinsic: ; ARMV7: @ %bb.0: -; ARMV7-NEXT: vmov s0, r0 -; ARMV7-NEXT: vmov s2, r1 -; ARMV7-NEXT: vcmp.f32 s0, s2 +; ARMV7-NEXT: vmov s0, r1 +; ARMV7-NEXT: vmov s2, r0 +; ARMV7-NEXT: vcmp.f32 s2, s0 ; ARMV7-NEXT: vmrs APSR_nzcv, fpscr -; ARMV7-NEXT: vmovlt.f32 s2, s0 -; ARMV7-NEXT: vmov r0, s2 +; ARMV7-NEXT: vmovlt.f32 s0, s2 +; ARMV7-NEXT: vmov r0, s0 ; ARMV7-NEXT: bx lr ; ; ARMV8-LABEL: fminnum32_intrinsic: @@ -102,12 +102,12 @@ define float @fminnum32_non_zero_intrinsic(float %x) { define float @fmaxnum32_intrinsic(float %x, float %y) { ; ARMV7-LABEL: fmaxnum32_intrinsic: ; ARMV7: @ %bb.0: -; ARMV7-NEXT: vmov s0, r0 -; ARMV7-NEXT: vmov s2, r1 -; ARMV7-NEXT: vcmp.f32 s0, s2 +; ARMV7-NEXT: vmov s0, r1 +; ARMV7-NEXT: vmov s2, r0 +; ARMV7-NEXT: vcmp.f32 s2, s0 ; ARMV7-NEXT: vmrs APSR_nzcv, fpscr -; ARMV7-NEXT: vmovgt.f32 s2, s0 -; ARMV7-NEXT: vmov r0, s2 +; ARMV7-NEXT: vmovgt.f32 s0, s2 +; ARMV7-NEXT: vmov r0, s0 ; ARMV7-NEXT: bx lr ; ; ARMV8-LABEL: fmaxnum32_intrinsic: @@ -160,12 +160,12 @@ define float @fmaxnum32_nsz_intrinsic(float %x, float %y) { define float @fmaxnum32_zero_intrinsic(float %x) { ; ARMV7-LABEL: fmaxnum32_zero_intrinsic: ; ARMV7: @ %bb.0: -; ARMV7-NEXT: vmov s2, r0 -; ARMV7-NEXT: vldr s0, .LCPI5_0 -; ARMV7-NEXT: vcmp.f32 s2, #0 +; ARMV7-NEXT: vmov s0, r0 +; ARMV7-NEXT: vldr s2, .LCPI5_0 +; ARMV7-NEXT: vcmp.f32 s0, #0 ; ARMV7-NEXT: vmrs APSR_nzcv, fpscr -; ARMV7-NEXT: vmovgt.f32 s0, s2 -; ARMV7-NEXT: vmov r0, s0 +; ARMV7-NEXT: vmovgt.f32 s2, s0 +; ARMV7-NEXT: vmov r0, s2 ; ARMV7-NEXT: bx lr ; ARMV7-NEXT: .p2align 2 ; ARMV7-NEXT: @ %bb.1: @@ -425,12 +425,12 @@ define double@fmaxnum64_nsz_intrinsic(double %x, double %y) { define double @fmaxnum64_zero_intrinsic(double %x) { ; ARMV7-LABEL: fmaxnum64_zero_intrinsic: ; ARMV7: @ %bb.0: -; ARMV7-NEXT: vmov d17, r0, r1 -; ARMV7-NEXT: vcmp.f64 d17, #0 +; ARMV7-NEXT: vmov d16, r0, r1 +; ARMV7-NEXT: vcmp.f64 d16, #0 ; ARMV7-NEXT: vmrs APSR_nzcv, fpscr -; ARMV7-NEXT: vmov.i32 d16, #0x0 -; ARMV7-NEXT: vmovgt.f64 d16, d17 -; ARMV7-NEXT: vmov r0, r1, d16 +; ARMV7-NEXT: vmov.i32 d17, #0x0 +; ARMV7-NEXT: vmovgt.f64 d17, d16 +; ARMV7-NEXT: vmov r0, r1, d17 ; ARMV7-NEXT: bx lr ; ; ARMV8-LABEL: fmaxnum64_zero_intrinsic: @@ -1065,18 +1065,18 @@ define <2 x double> @fminnumv264_one_zero_intrinsic(<2 x double> %x) { ; ; ARMV8M-LABEL: fminnumv264_one_zero_intrinsic: ; ARMV8M: @ %bb.0: -; ARMV8M-NEXT: vmov d3, r2, r3 -; ARMV8M-NEXT: vldr d1, .LCPI27_0 -; ARMV8M-NEXT: vcmp.f64 d3, #0 +; ARMV8M-NEXT: vmov d1, r2, r3 +; ARMV8M-NEXT: vldr d0, .LCPI27_0 +; ARMV8M-NEXT: vcmp.f64 d1, #0 ; ARMV8M-NEXT: vmrs APSR_nzcv, fpscr ; ARMV8M-NEXT: vmov d2, r0, r1 -; ARMV8M-NEXT: vmov.f64 d0, #-1.000000e+00 -; ARMV8M-NEXT: vcmp.f64 d0, d2 -; ARMV8M-NEXT: vmovlt.f64 d1, d3 +; ARMV8M-NEXT: vmov.f64 d3, #-1.000000e+00 +; ARMV8M-NEXT: vcmp.f64 d3, d2 +; ARMV8M-NEXT: vmovlt.f64 d0, d1 ; ARMV8M-NEXT: vmrs APSR_nzcv, fpscr -; ARMV8M-NEXT: vmov r2, r3, d1 -; ARMV8M-NEXT: vselgt.f64 d0, d2, d0 -; ARMV8M-NEXT: vmov r0, r1, d0 +; ARMV8M-NEXT: vmov r2, r3, d0 +; ARMV8M-NEXT: vselgt.f64 d1, d2, d3 +; ARMV8M-NEXT: vmov r0, r1, d1 ; ARMV8M-NEXT: bx lr ; ARMV8M-NEXT: .p2align 3 ; ARMV8M-NEXT: @ %bb.1: @@ -1186,18 +1186,18 @@ define <2 x double> @fmaxnumv264_nsz_intrinsic(<2 x double> %x, <2 x double> %y) define <2 x double> @fmaxnumv264_zero_intrinsic(<2 x double> %x) { ; ARMV7-LABEL: fmaxnumv264_zero_intrinsic: ; ARMV7: @ %bb.0: -; ARMV7-NEXT: vldr d17, .LCPI30_0 -; ARMV7-NEXT: vmov d18, r2, r3 -; ARMV7-NEXT: vmov d19, r0, r1 -; ARMV7-NEXT: vcmp.f64 d18, d17 +; ARMV7-NEXT: vmov d18, r0, r1 +; ARMV7-NEXT: vldr d16, .LCPI30_0 +; ARMV7-NEXT: vcmp.f64 d18, #0 ; ARMV7-NEXT: vmrs APSR_nzcv, fpscr -; ARMV7-NEXT: vmov.i32 d16, #0x0 -; ARMV7-NEXT: vcmp.f64 d19, #0 +; ARMV7-NEXT: vmov d19, r2, r3 +; ARMV7-NEXT: vcmp.f64 d19, d16 +; ARMV7-NEXT: vmov.i32 d17, #0x0 ; ARMV7-NEXT: vmovgt.f64 d17, d18 ; ARMV7-NEXT: vmrs APSR_nzcv, fpscr -; ARMV7-NEXT: vmov r2, r3, d17 +; ARMV7-NEXT: vmov r0, r1, d17 ; ARMV7-NEXT: vmovgt.f64 d16, d19 -; ARMV7-NEXT: vmov r0, r1, d16 +; ARMV7-NEXT: vmov r2, r3, d16 ; ARMV7-NEXT: bx lr ; ARMV7-NEXT: .p2align 3 ; ARMV7-NEXT: @ %bb.1: @@ -1225,26 +1225,26 @@ define <2 x double> @fmaxnumv264_zero_intrinsic(<2 x double> %x) { ; ARMV8M-LABEL: fmaxnumv264_zero_intrinsic: ; ARMV8M: @ %bb.0: ; ARMV8M-NEXT: vmov d2, r0, r1 -; ARMV8M-NEXT: vldr d0, .LCPI30_0 +; ARMV8M-NEXT: vldr d1, .LCPI30_1 ; ARMV8M-NEXT: vcmp.f64 d2, #0 ; ARMV8M-NEXT: vmrs APSR_nzcv, fpscr ; ARMV8M-NEXT: vmov d3, r2, r3 -; ARMV8M-NEXT: vcmp.f64 d3, d0 -; ARMV8M-NEXT: vldr d1, .LCPI30_1 -; ARMV8M-NEXT: vselgt.f64 d1, d2, d1 +; ARMV8M-NEXT: vcmp.f64 d3, d1 +; ARMV8M-NEXT: vldr d0, .LCPI30_0 +; ARMV8M-NEXT: vselgt.f64 d0, d2, d0 ; ARMV8M-NEXT: vmrs APSR_nzcv, fpscr -; ARMV8M-NEXT: vmov r0, r1, d1 -; ARMV8M-NEXT: vselgt.f64 d0, d3, d0 -; ARMV8M-NEXT: vmov r2, r3, d0 +; ARMV8M-NEXT: vmov r0, r1, d0 +; ARMV8M-NEXT: vselgt.f64 d1, d3, d1 +; ARMV8M-NEXT: vmov r2, r3, d1 ; ARMV8M-NEXT: bx lr ; ARMV8M-NEXT: .p2align 3 ; ARMV8M-NEXT: @ %bb.1: ; ARMV8M-NEXT: .LCPI30_0: -; ARMV8M-NEXT: .long 0 @ double -0 -; ARMV8M-NEXT: .long 2147483648 -; ARMV8M-NEXT: .LCPI30_1: ; ARMV8M-NEXT: .long 0 @ double 0 ; ARMV8M-NEXT: .long 0 +; ARMV8M-NEXT: .LCPI30_1: +; ARMV8M-NEXT: .long 0 @ double -0 +; ARMV8M-NEXT: .long 2147483648 %a = call nnan <2 x double> @llvm.maxnum.v2f64(<2 x double> %x, <2 x double>) ret <2 x double> %a } diff --git a/llvm/test/CodeGen/ARM/select.ll b/llvm/test/CodeGen/ARM/select.ll index 24ca9aeac7f2d..496a6c0f5acbb 100644 --- a/llvm/test/CodeGen/ARM/select.ll +++ b/llvm/test/CodeGen/ARM/select.ll @@ -164,13 +164,13 @@ define double @f7(double %a, double %b) { ; CHECK-VFP-LABEL: f7: ; CHECK-VFP: @ %bb.0: ; CHECK-VFP-NEXT: vldr d17, .LCPI6_0 -; CHECK-VFP-NEXT: vmov d19, r0, r1 +; CHECK-VFP-NEXT: vmov d18, r0, r1 ; CHECK-VFP-NEXT: vmov.f64 d16, #-1.000000e+00 -; CHECK-VFP-NEXT: vcmp.f64 d19, d17 +; CHECK-VFP-NEXT: vcmp.f64 d18, d17 ; CHECK-VFP-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-VFP-NEXT: vmov d18, r2, r3 -; CHECK-VFP-NEXT: vmovmi.f64 d18, d16 -; CHECK-VFP-NEXT: vmov r0, r1, d18 +; CHECK-VFP-NEXT: vmov d19, r2, r3 +; CHECK-VFP-NEXT: vmovmi.f64 d19, d16 +; CHECK-VFP-NEXT: vmov r0, r1, d19 ; CHECK-VFP-NEXT: bx lr ; CHECK-VFP-NEXT: .p2align 3 ; CHECK-VFP-NEXT: @ %bb.1: @@ -181,14 +181,14 @@ define double @f7(double %a, double %b) { ; CHECK-NEON-LABEL: f7: ; CHECK-NEON: @ %bb.0: ; CHECK-NEON-NEXT: vldr d17, LCPI6_0 -; CHECK-NEON-NEXT: vmov d19, r0, r1 -; CHECK-NEON-NEXT: vmov d18, r2, r3 -; CHECK-NEON-NEXT: vcmp.f64 d19, d17 +; CHECK-NEON-NEXT: vmov d18, r0, r1 +; CHECK-NEON-NEXT: vmov d19, r2, r3 +; CHECK-NEON-NEXT: vcmp.f64 d18, d17 ; CHECK-NEON-NEXT: vmov.f64 d16, #-1.000000e+00 ; CHECK-NEON-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEON-NEXT: it mi -; CHECK-NEON-NEXT: vmovmi.f64 d18, d16 -; CHECK-NEON-NEXT: vmov r0, r1, d18 +; CHECK-NEON-NEXT: vmovmi.f64 d19, d16 +; CHECK-NEON-NEXT: vmov r0, r1, d19 ; CHECK-NEON-NEXT: bx lr ; CHECK-NEON-NEXT: .p2align 3 ; CHECK-NEON-NEXT: @ %bb.1: diff --git a/llvm/test/CodeGen/Thumb2/mve-fmas.ll b/llvm/test/CodeGen/Thumb2/mve-fmas.ll index 8016b940b8d51..377440e1bbc93 100644 --- a/llvm/test/CodeGen/Thumb2/mve-fmas.ll +++ b/llvm/test/CodeGen/Thumb2/mve-fmas.ll @@ -896,17 +896,17 @@ define arm_aapcs_vfpcc <4 x float> @vfma32_v1_pred(<4 x float> %src1, <4 x float ; CHECK-MVE-NEXT: vcmp.f32 s5, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, #0 -; CHECK-MVE-NEXT: vmov.f32 s14, s0 -; CHECK-MVE-NEXT: vmov.f32 s12, s1 -; CHECK-MVE-NEXT: vmla.f32 s14, s4, s8 -; CHECK-MVE-NEXT: vmov.f32 s4, s3 -; CHECK-MVE-NEXT: vmov.f32 s8, s2 -; CHECK-MVE-NEXT: vmla.f32 s12, s5, s9 -; CHECK-MVE-NEXT: vmla.f32 s4, s7, s11 -; CHECK-MVE-NEXT: vmla.f32 s8, s6, s10 +; CHECK-MVE-NEXT: vmov.f32 s14, s2 +; CHECK-MVE-NEXT: vmov.f32 s12, s3 +; CHECK-MVE-NEXT: vmla.f32 s14, s6, s10 +; CHECK-MVE-NEXT: vmov.f32 s10, s1 +; CHECK-MVE-NEXT: vmla.f32 s12, s7, s11 +; CHECK-MVE-NEXT: vmla.f32 s10, s5, s9 +; CHECK-MVE-NEXT: vmov.f32 s9, s0 ; CHECK-MVE-NEXT: cset r0, mi ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s7, #0 +; CHECK-MVE-NEXT: vmla.f32 s9, s4, s8 ; CHECK-MVE-NEXT: cset r1, mi ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s6, #0 @@ -914,13 +914,13 @@ define arm_aapcs_vfpcc <4 x float> @vfma32_v1_pred(<4 x float> %src1, <4 x float ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: cset r3, mi ; CHECK-MVE-NEXT: cmp r2, #0 -; CHECK-MVE-NEXT: vseleq.f32 s3, s3, s4 +; CHECK-MVE-NEXT: vseleq.f32 s3, s3, s12 ; CHECK-MVE-NEXT: cmp r3, #0 -; CHECK-MVE-NEXT: vseleq.f32 s2, s2, s8 +; CHECK-MVE-NEXT: vseleq.f32 s2, s2, s14 ; CHECK-MVE-NEXT: cmp r0, #0 -; CHECK-MVE-NEXT: vseleq.f32 s1, s1, s12 +; CHECK-MVE-NEXT: vseleq.f32 s1, s1, s10 ; CHECK-MVE-NEXT: cmp r1, #0 -; CHECK-MVE-NEXT: vseleq.f32 s0, s0, s14 +; CHECK-MVE-NEXT: vseleq.f32 s0, s0, s9 ; CHECK-MVE-NEXT: bx lr entry: %0 = fmul <4 x float> %src2, %src3 @@ -949,17 +949,17 @@ define arm_aapcs_vfpcc <4 x float> @vfma32_v2_pred(<4 x float> %src1, <4 x float ; CHECK-MVE-NEXT: vcmp.f32 s5, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, #0 -; CHECK-MVE-NEXT: vmov.f32 s14, s0 -; CHECK-MVE-NEXT: vmov.f32 s12, s1 -; CHECK-MVE-NEXT: vmla.f32 s14, s4, s8 -; CHECK-MVE-NEXT: vmov.f32 s4, s3 -; CHECK-MVE-NEXT: vmov.f32 s8, s2 -; CHECK-MVE-NEXT: vmla.f32 s12, s5, s9 -; CHECK-MVE-NEXT: vmla.f32 s4, s7, s11 -; CHECK-MVE-NEXT: vmla.f32 s8, s6, s10 +; CHECK-MVE-NEXT: vmov.f32 s14, s2 +; CHECK-MVE-NEXT: vmov.f32 s12, s3 +; CHECK-MVE-NEXT: vmla.f32 s14, s6, s10 +; CHECK-MVE-NEXT: vmov.f32 s10, s1 +; CHECK-MVE-NEXT: vmla.f32 s12, s7, s11 +; CHECK-MVE-NEXT: vmla.f32 s10, s5, s9 +; CHECK-MVE-NEXT: vmov.f32 s9, s0 ; CHECK-MVE-NEXT: cset r0, mi ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s7, #0 +; CHECK-MVE-NEXT: vmla.f32 s9, s4, s8 ; CHECK-MVE-NEXT: cset r1, mi ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s6, #0 @@ -967,13 +967,13 @@ define arm_aapcs_vfpcc <4 x float> @vfma32_v2_pred(<4 x float> %src1, <4 x float ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: cset r3, mi ; CHECK-MVE-NEXT: cmp r2, #0 -; CHECK-MVE-NEXT: vseleq.f32 s3, s3, s4 +; CHECK-MVE-NEXT: vseleq.f32 s3, s3, s12 ; CHECK-MVE-NEXT: cmp r3, #0 -; CHECK-MVE-NEXT: vseleq.f32 s2, s2, s8 +; CHECK-MVE-NEXT: vseleq.f32 s2, s2, s14 ; CHECK-MVE-NEXT: cmp r0, #0 -; CHECK-MVE-NEXT: vseleq.f32 s1, s1, s12 +; CHECK-MVE-NEXT: vseleq.f32 s1, s1, s10 ; CHECK-MVE-NEXT: cmp r1, #0 -; CHECK-MVE-NEXT: vseleq.f32 s0, s0, s14 +; CHECK-MVE-NEXT: vseleq.f32 s0, s0, s9 ; CHECK-MVE-NEXT: bx lr entry: %0 = fmul <4 x float> %src2, %src3 @@ -1002,17 +1002,17 @@ define arm_aapcs_vfpcc <4 x float> @vfms32_pred(<4 x float> %src1, <4 x float> % ; CHECK-MVE-NEXT: vcmp.f32 s5, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, #0 -; CHECK-MVE-NEXT: vmov.f32 s14, s0 -; CHECK-MVE-NEXT: vmov.f32 s12, s1 -; CHECK-MVE-NEXT: vmls.f32 s14, s4, s8 -; CHECK-MVE-NEXT: vmov.f32 s4, s3 -; CHECK-MVE-NEXT: vmov.f32 s8, s2 -; CHECK-MVE-NEXT: vmls.f32 s12, s5, s9 -; CHECK-MVE-NEXT: vmls.f32 s4, s7, s11 -; CHECK-MVE-NEXT: vmls.f32 s8, s6, s10 +; CHECK-MVE-NEXT: vmov.f32 s14, s2 +; CHECK-MVE-NEXT: vmov.f32 s12, s3 +; CHECK-MVE-NEXT: vmls.f32 s14, s6, s10 +; CHECK-MVE-NEXT: vmov.f32 s10, s1 +; CHECK-MVE-NEXT: vmls.f32 s12, s7, s11 +; CHECK-MVE-NEXT: vmls.f32 s10, s5, s9 +; CHECK-MVE-NEXT: vmov.f32 s9, s0 ; CHECK-MVE-NEXT: cset r0, mi ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s7, #0 +; CHECK-MVE-NEXT: vmls.f32 s9, s4, s8 ; CHECK-MVE-NEXT: cset r1, mi ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s6, #0 @@ -1020,13 +1020,13 @@ define arm_aapcs_vfpcc <4 x float> @vfms32_pred(<4 x float> %src1, <4 x float> % ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: cset r3, mi ; CHECK-MVE-NEXT: cmp r2, #0 -; CHECK-MVE-NEXT: vseleq.f32 s3, s3, s4 +; CHECK-MVE-NEXT: vseleq.f32 s3, s3, s12 ; CHECK-MVE-NEXT: cmp r3, #0 -; CHECK-MVE-NEXT: vseleq.f32 s2, s2, s8 +; CHECK-MVE-NEXT: vseleq.f32 s2, s2, s14 ; CHECK-MVE-NEXT: cmp r0, #0 -; CHECK-MVE-NEXT: vseleq.f32 s1, s1, s12 +; CHECK-MVE-NEXT: vseleq.f32 s1, s1, s10 ; CHECK-MVE-NEXT: cmp r1, #0 -; CHECK-MVE-NEXT: vseleq.f32 s0, s0, s14 +; CHECK-MVE-NEXT: vseleq.f32 s0, s0, s9 ; CHECK-MVE-NEXT: bx lr entry: %0 = fmul <4 x float> %src2, %src3 @@ -1058,14 +1058,14 @@ define arm_aapcs_vfpcc <4 x float> @vfmar32_pred(<4 x float> %src1, <4 x float> ; CHECK-MVE-NEXT: vcmp.f32 s5, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, #0 -; CHECK-MVE-NEXT: vmov.f32 s12, s0 -; CHECK-MVE-NEXT: vmov.f32 s14, s2 -; CHECK-MVE-NEXT: vmov.f32 s10, s1 -; CHECK-MVE-NEXT: vmla.f32 s12, s4, s8 -; CHECK-MVE-NEXT: vmov.f32 s4, s3 -; CHECK-MVE-NEXT: vmla.f32 s14, s6, s8 -; CHECK-MVE-NEXT: vmla.f32 s10, s5, s8 -; CHECK-MVE-NEXT: vmla.f32 s4, s7, s8 +; CHECK-MVE-NEXT: vmov.f32 s10, s3 +; CHECK-MVE-NEXT: vmov.f32 s12, s2 +; CHECK-MVE-NEXT: vmov.f32 s14, s1 +; CHECK-MVE-NEXT: vmov.f32 s9, s0 +; CHECK-MVE-NEXT: vmla.f32 s10, s7, s8 +; CHECK-MVE-NEXT: vmla.f32 s12, s6, s8 +; CHECK-MVE-NEXT: vmla.f32 s14, s5, s8 +; CHECK-MVE-NEXT: vmla.f32 s9, s4, s8 ; CHECK-MVE-NEXT: cset r0, mi ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s7, #0 @@ -1076,13 +1076,13 @@ define arm_aapcs_vfpcc <4 x float> @vfmar32_pred(<4 x float> %src1, <4 x float> ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: cset r3, mi ; CHECK-MVE-NEXT: cmp r2, #0 -; CHECK-MVE-NEXT: vseleq.f32 s3, s3, s4 +; CHECK-MVE-NEXT: vseleq.f32 s3, s3, s10 ; CHECK-MVE-NEXT: cmp r3, #0 -; CHECK-MVE-NEXT: vseleq.f32 s2, s2, s14 +; CHECK-MVE-NEXT: vseleq.f32 s2, s2, s12 ; CHECK-MVE-NEXT: cmp r0, #0 -; CHECK-MVE-NEXT: vseleq.f32 s1, s1, s10 +; CHECK-MVE-NEXT: vseleq.f32 s1, s1, s14 ; CHECK-MVE-NEXT: cmp r1, #0 -; CHECK-MVE-NEXT: vseleq.f32 s0, s0, s12 +; CHECK-MVE-NEXT: vseleq.f32 s0, s0, s9 ; CHECK-MVE-NEXT: bx lr entry: %i = insertelement <4 x float> undef, float %src3, i32 0 @@ -1115,13 +1115,13 @@ define arm_aapcs_vfpcc <4 x float> @vfmas32_pred(<4 x float> %src1, <4 x float> ; CHECK-MVE-NEXT: vcmp.f32 s5, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, #0 -; CHECK-MVE-NEXT: vmov.f32 s12, s8 ; CHECK-MVE-NEXT: vmov.f32 s10, s8 -; CHECK-MVE-NEXT: vmla.f32 s12, s0, s4 -; CHECK-MVE-NEXT: vmov.f32 s4, s8 -; CHECK-MVE-NEXT: vmla.f32 s8, s2, s6 -; CHECK-MVE-NEXT: vmla.f32 s10, s1, s5 -; CHECK-MVE-NEXT: vmla.f32 s4, s3, s7 +; CHECK-MVE-NEXT: vmov.f32 s12, s8 +; CHECK-MVE-NEXT: vmov.f32 s14, s8 +; CHECK-MVE-NEXT: vmla.f32 s8, s0, s4 +; CHECK-MVE-NEXT: vmla.f32 s10, s3, s7 +; CHECK-MVE-NEXT: vmla.f32 s12, s2, s6 +; CHECK-MVE-NEXT: vmla.f32 s14, s1, s5 ; CHECK-MVE-NEXT: cset r0, mi ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s7, #0 @@ -1132,13 +1132,13 @@ define arm_aapcs_vfpcc <4 x float> @vfmas32_pred(<4 x float> %src1, <4 x float> ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: cset r3, mi ; CHECK-MVE-NEXT: cmp r2, #0 -; CHECK-MVE-NEXT: vseleq.f32 s3, s3, s4 +; CHECK-MVE-NEXT: vseleq.f32 s3, s3, s10 ; CHECK-MVE-NEXT: cmp r3, #0 -; CHECK-MVE-NEXT: vseleq.f32 s2, s2, s8 +; CHECK-MVE-NEXT: vseleq.f32 s2, s2, s12 ; CHECK-MVE-NEXT: cmp r0, #0 -; CHECK-MVE-NEXT: vseleq.f32 s1, s1, s10 +; CHECK-MVE-NEXT: vseleq.f32 s1, s1, s14 ; CHECK-MVE-NEXT: cmp r1, #0 -; CHECK-MVE-NEXT: vseleq.f32 s0, s0, s12 +; CHECK-MVE-NEXT: vseleq.f32 s0, s0, s8 ; CHECK-MVE-NEXT: bx lr entry: %i = insertelement <4 x float> undef, float %src3, i32 0 diff --git a/llvm/test/CodeGen/Thumb2/mve-fptosi-sat-vector.ll b/llvm/test/CodeGen/Thumb2/mve-fptosi-sat-vector.ll index 81b6a6940a7d6..75b6cb3e1272b 100644 --- a/llvm/test/CodeGen/Thumb2/mve-fptosi-sat-vector.ll +++ b/llvm/test/CodeGen/Thumb2/mve-fptosi-sat-vector.ll @@ -42,54 +42,36 @@ define arm_aapcs_vfpcc <2 x i32> @test_signed_v2f32_v2i32(<2 x float> %f) { ; CHECK-NEXT: vldr s20, .LCPI1_1 ; CHECK-NEXT: vcmp.f32 s17, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r4, #-1 ; CHECK-NEXT: movlt.w r5, #-2147483648 ; CHECK-NEXT: vcmp.f32 s17, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt ; CHECK-NEXT: mvngt r5, #-2147483648 +; CHECK-NEXT: movgt r4, #0 ; CHECK-NEXT: vcmp.f32 s17, s17 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r4, #0 ; CHECK-NEXT: movvs r5, #0 ; CHECK-NEXT: bl __aeabi_f2lz ; CHECK-NEXT: vcmp.f32 s16, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt.w r0, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r0, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s18 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r4, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s17 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s18 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt ; CHECK-NEXT: movlt.w r1, #-1 ; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt ; CHECK-NEXT: movgt r1, #0 +; CHECK-NEXT: mvngt r0, #-2147483648 ; CHECK-NEXT: vcmp.f32 s16, s16 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r0, #0 ; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 ; CHECK-NEXT: vmov q0[3], q0[1], r1, r4 ; CHECK-NEXT: vpop {d8, d9, d10} ; CHECK-NEXT: pop {r4, r5, r7, pc} @@ -1275,54 +1257,36 @@ define arm_aapcs_vfpcc <2 x i32> @test_signed_v2f16_v2i32(<2 x half> %f) { ; CHECK-NEXT: mov r4, r1 ; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r4, #-1 ; CHECK-NEXT: movlt.w r5, #-2147483648 ; CHECK-NEXT: vcmp.f32 s18, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt ; CHECK-NEXT: mvngt r5, #-2147483648 +; CHECK-NEXT: movgt r4, #0 ; CHECK-NEXT: vcmp.f32 s18, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r4, #0 ; CHECK-NEXT: movvs r5, #0 ; CHECK-NEXT: bl __aeabi_f2lz ; CHECK-NEXT: vcmp.f32 s16, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt.w r0, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r0, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r4, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt ; CHECK-NEXT: movlt.w r1, #-1 ; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt ; CHECK-NEXT: movgt r1, #0 +; CHECK-NEXT: mvngt r0, #-2147483648 ; CHECK-NEXT: vcmp.f32 s16, s16 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r0, #0 ; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 ; CHECK-NEXT: vmov q0[3], q0[1], r1, r4 ; CHECK-NEXT: vpop {d8, d9, d10, d11} ; CHECK-NEXT: pop {r4, r5, r7, pc} @@ -1568,9 +1532,9 @@ define arm_aapcs_vfpcc <4 x i8> @test_signed_v4f32_v4i8(<4 x float> %f) { ; CHECK-MVE-NEXT: vmaxnm.f32 s12, s2, s4 ; CHECK-MVE-NEXT: vmaxnm.f32 s10, s0, s4 ; CHECK-MVE-NEXT: vminnm.f32 s12, s12, s6 -; CHECK-MVE-NEXT: vmaxnm.f32 s8, s1, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s8, s3, s4 ; CHECK-MVE-NEXT: vminnm.f32 s10, s10, s6 -; CHECK-MVE-NEXT: vmaxnm.f32 s4, s3, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s4, s1, s4 ; CHECK-MVE-NEXT: vcvt.s32.f32 s12, s12 ; CHECK-MVE-NEXT: vminnm.f32 s8, s8, s6 ; CHECK-MVE-NEXT: vminnm.f32 s4, s4, s6 @@ -1588,10 +1552,10 @@ define arm_aapcs_vfpcc <4 x i8> @test_signed_v4f32_v4i8(<4 x float> %f) { ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r1, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r2, s4 +; CHECK-MVE-NEXT: vmov r2, s8 ; CHECK-MVE-NEXT: vcmp.f32 s1, s1 ; CHECK-MVE-NEXT: vmov q0[2], q0[0], r1, r0 -; CHECK-MVE-NEXT: vmov r3, s8 +; CHECK-MVE-NEXT: vmov r3, s4 ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r2, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr @@ -1627,9 +1591,9 @@ define arm_aapcs_vfpcc <4 x i13> @test_signed_v4f32_v4i13(<4 x float> %f) { ; CHECK-MVE-NEXT: vmaxnm.f32 s12, s2, s4 ; CHECK-MVE-NEXT: vmaxnm.f32 s10, s0, s4 ; CHECK-MVE-NEXT: vminnm.f32 s12, s12, s6 -; CHECK-MVE-NEXT: vmaxnm.f32 s8, s1, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s8, s3, s4 ; CHECK-MVE-NEXT: vminnm.f32 s10, s10, s6 -; CHECK-MVE-NEXT: vmaxnm.f32 s4, s3, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s4, s1, s4 ; CHECK-MVE-NEXT: vcvt.s32.f32 s12, s12 ; CHECK-MVE-NEXT: vminnm.f32 s8, s8, s6 ; CHECK-MVE-NEXT: vminnm.f32 s4, s4, s6 @@ -1647,10 +1611,10 @@ define arm_aapcs_vfpcc <4 x i13> @test_signed_v4f32_v4i13(<4 x float> %f) { ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r1, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r2, s4 +; CHECK-MVE-NEXT: vmov r2, s8 ; CHECK-MVE-NEXT: vcmp.f32 s1, s1 ; CHECK-MVE-NEXT: vmov q0[2], q0[0], r1, r0 -; CHECK-MVE-NEXT: vmov r3, s8 +; CHECK-MVE-NEXT: vmov r3, s4 ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r2, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr @@ -1686,9 +1650,9 @@ define arm_aapcs_vfpcc <4 x i16> @test_signed_v4f32_v4i16(<4 x float> %f) { ; CHECK-MVE-NEXT: vmaxnm.f32 s12, s2, s4 ; CHECK-MVE-NEXT: vmaxnm.f32 s10, s0, s4 ; CHECK-MVE-NEXT: vminnm.f32 s12, s12, s6 -; CHECK-MVE-NEXT: vmaxnm.f32 s8, s1, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s8, s3, s4 ; CHECK-MVE-NEXT: vminnm.f32 s10, s10, s6 -; CHECK-MVE-NEXT: vmaxnm.f32 s4, s3, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s4, s1, s4 ; CHECK-MVE-NEXT: vcvt.s32.f32 s12, s12 ; CHECK-MVE-NEXT: vminnm.f32 s8, s8, s6 ; CHECK-MVE-NEXT: vminnm.f32 s4, s4, s6 @@ -1706,10 +1670,10 @@ define arm_aapcs_vfpcc <4 x i16> @test_signed_v4f32_v4i16(<4 x float> %f) { ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r1, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r2, s4 +; CHECK-MVE-NEXT: vmov r2, s8 ; CHECK-MVE-NEXT: vcmp.f32 s1, s1 ; CHECK-MVE-NEXT: vmov q0[2], q0[0], r1, r0 -; CHECK-MVE-NEXT: vmov r3, s8 +; CHECK-MVE-NEXT: vmov r3, s4 ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r2, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr @@ -1743,9 +1707,9 @@ define arm_aapcs_vfpcc <4 x i19> @test_signed_v4f32_v4i19(<4 x float> %f) { ; CHECK-MVE-NEXT: vmaxnm.f32 s12, s2, s4 ; CHECK-MVE-NEXT: vmaxnm.f32 s10, s0, s4 ; CHECK-MVE-NEXT: vminnm.f32 s12, s12, s6 -; CHECK-MVE-NEXT: vmaxnm.f32 s8, s1, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s8, s3, s4 ; CHECK-MVE-NEXT: vminnm.f32 s10, s10, s6 -; CHECK-MVE-NEXT: vmaxnm.f32 s4, s3, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s4, s1, s4 ; CHECK-MVE-NEXT: vcvt.s32.f32 s12, s12 ; CHECK-MVE-NEXT: vminnm.f32 s8, s8, s6 ; CHECK-MVE-NEXT: vminnm.f32 s4, s4, s6 @@ -1763,10 +1727,10 @@ define arm_aapcs_vfpcc <4 x i19> @test_signed_v4f32_v4i19(<4 x float> %f) { ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r1, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r2, s4 +; CHECK-MVE-NEXT: vmov r2, s8 ; CHECK-MVE-NEXT: vcmp.f32 s1, s1 ; CHECK-MVE-NEXT: vmov q0[2], q0[0], r1, r0 -; CHECK-MVE-NEXT: vmov r3, s8 +; CHECK-MVE-NEXT: vmov r3, s4 ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r2, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr @@ -1821,146 +1785,122 @@ define arm_aapcs_vfpcc <4 x i32> @test_signed_v4f32_v4i32_duplicate(<4 x float> define arm_aapcs_vfpcc <4 x i50> @test_signed_v4f32_v4i50(<4 x float> %f) { ; CHECK-LABEL: test_signed_v4f32_v4i50: ; CHECK: @ %bb.0: -; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, lr} -; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, lr} +; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r11, lr} +; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r11, lr} ; CHECK-NEXT: .vsave {d8, d9, d10, d11} ; CHECK-NEXT: vpush {d8, d9, d10, d11} ; CHECK-NEXT: vmov q4, q0 ; CHECK-NEXT: mov r8, r0 +; CHECK-NEXT: vmov r0, s19 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: mov r4, r0 ; CHECK-NEXT: vmov r0, s18 +; CHECK-NEXT: mov r7, r1 ; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: mov r9, r0 -; CHECK-NEXT: vmov r0, s19 -; CHECK-NEXT: vldr s20, .LCPI28_0 +; CHECK-NEXT: mov r6, r0 +; CHECK-NEXT: vmov r0, s17 ; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: vmov r6, s16 -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt lt -; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: movtlt r5, #65534 ; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vldr s22, .LCPI28_1 -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: mov r4, r0 +; CHECK-NEXT: mov r9, r0 +; CHECK-NEXT: vmov r0, s16 +; CHECK-NEXT: vldr s22, .LCPI28_0 +; CHECK-NEXT: mov r11, r1 +; CHECK-NEXT: vldr s20, .LCPI28_1 +; CHECK-NEXT: vcmp.f32 s17, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r7, r1 -; CHECK-NEXT: mov r0, r6 -; CHECK-NEXT: vcmp.f32 s18, s22 ; CHECK-NEXT: itt lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: movtlt r7, #65534 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r5, #65535 -; CHECK-NEXT: movtgt r5, #1 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: mov r10, r1 +; CHECK-NEXT: movwlt r11, #0 +; CHECK-NEXT: movtlt r11, #65534 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s22 +; CHECK-NEXT: vcmp.f32 s17, s20 ; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt.w r9, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r7, #65535 -; CHECK-NEXT: movtgt r7, #1 +; CHECK-NEXT: vcmp.f32 s17, s17 +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r11, #65535 +; CHECK-NEXT: movtgt r11, #1 +; CHECK-NEXT: movgt.w r9, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s16, s16 +; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs.w r9, #0 +; CHECK-NEXT: movvs.w r11, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str.w r0, [r8] -; CHECK-NEXT: vmov r0, s17 -; CHECK-NEXT: vcmp.f32 s19, s20 +; CHECK-NEXT: vcmp.f32 s18, s20 +; CHECK-NEXT: ittt lt +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movtlt r5, #65534 +; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: vcmp.f32 s18, s18 +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r5, #65535 +; CHECK-NEXT: movtgt r5, #1 +; CHECK-NEXT: movgt.w r6, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: it lt +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r6, #0 +; CHECK-NEXT: movvs r5, #0 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: ittt lt +; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: movtlt r7, #65534 ; CHECK-NEXT: movlt r4, #0 +; CHECK-NEXT: vcmp.f32 s19, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r7, #65535 +; CHECK-NEXT: movtgt r7, #1 ; CHECK-NEXT: movgt.w r4, #-1 +; CHECK-NEXT: vcmp.f32 s19, s19 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: itt vs ; CHECK-NEXT: movvs r4, #0 ; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r9, #0 -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r9, #-1 -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: mov r1, r7 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt vs -; CHECK-NEXT: movvs.w r9, #0 -; CHECK-NEXT: movvs r5, #0 -; CHECK-NEXT: bfc r1, #18, #14 -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: bfc r5, #18, #14 -; CHECK-NEXT: mov r6, r9 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: lsll r4, r1, #22 -; CHECK-NEXT: lsrl r6, r5, #28 -; CHECK-NEXT: itt lt -; CHECK-NEXT: movwlt r10, #0 -; CHECK-NEXT: movtlt r10, #65534 -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: orrs r1, r5 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r10, #65535 -; CHECK-NEXT: movtgt r10, #1 -; CHECK-NEXT: str.w r1, [r8, #20] ; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: orr.w r2, r6, r4 +; CHECK-NEXT: vcmp.f32 s16, s22 +; CHECK-NEXT: bfc r5, #18, #14 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt lt +; CHECK-NEXT: ittt lt ; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: movtlt r1, #65534 -; CHECK-NEXT: vcmp.f32 s17, s22 +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: vcmp.f32 s16, s20 +; CHECK-NEXT: mov r2, r6 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movwgt r1, #65535 ; CHECK-NEXT: movtgt r1, #1 -; CHECK-NEXT: str.w r2, [r8, #16] -; CHECK-NEXT: lsrs r2, r7, #10 -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: strb.w r2, [r8, #24] -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s17, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s17, s17 +; CHECK-NEXT: vcmp.f32 s16, s16 +; CHECK-NEXT: lsrl r2, r5, #28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt vs +; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: bfc r1, #18, #14 -; CHECK-NEXT: mov r2, r0 -; CHECK-NEXT: lsrl r2, r1, #14 -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: orr.w r1, r1, r9, lsl #4 +; CHECK-NEXT: str.w r0, [r8] +; CHECK-NEXT: lsrs r0, r7, #10 +; CHECK-NEXT: bfc r7, #18, #14 +; CHECK-NEXT: bfc r11, #18, #14 +; CHECK-NEXT: lsll r4, r7, #22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: strd r2, r1, [r8, #8] +; CHECK-NEXT: orr.w r3, r5, r7 +; CHECK-NEXT: str.w r3, [r8, #20] +; CHECK-NEXT: orr.w r2, r2, r4 +; CHECK-NEXT: str.w r2, [r8, #16] +; CHECK-NEXT: strb.w r0, [r8, #24] +; CHECK-NEXT: mov r0, r9 +; CHECK-NEXT: lsrl r0, r11, #14 +; CHECK-NEXT: orr.w r2, r11, r6, lsl #4 +; CHECK-NEXT: strd r0, r2, [r8, #8] ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r10, #0 -; CHECK-NEXT: bfc r10, #18, #14 -; CHECK-NEXT: orr.w r0, r10, r0, lsl #18 +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: bfc r1, #18, #14 +; CHECK-NEXT: orr.w r0, r1, r9, lsl #18 ; CHECK-NEXT: str.w r0, [r8, #4] ; CHECK-NEXT: vpop {d8, d9, d10, d11} -; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, pc} +; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r11, pc} ; CHECK-NEXT: .p2align 2 ; CHECK-NEXT: @ %bb.1: ; CHECK-NEXT: .LCPI28_0: @@ -1974,131 +1914,98 @@ define arm_aapcs_vfpcc <4 x i50> @test_signed_v4f32_v4i50(<4 x float> %f) { define arm_aapcs_vfpcc <4 x i64> @test_signed_v4f32_v4i64(<4 x float> %f) { ; CHECK-LABEL: test_signed_v4f32_v4i64: ; CHECK: @ %bb.0: -; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, lr} -; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, lr} +; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, r11, lr} +; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} +; CHECK-NEXT: .pad #4 +; CHECK-NEXT: sub sp, #4 ; CHECK-NEXT: .vsave {d8, d9, d10, d11} ; CHECK-NEXT: vpush {d8, d9, d10, d11} ; CHECK-NEXT: vmov q4, q0 ; CHECK-NEXT: vmov r0, s19 ; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: mov r10, r0 +; CHECK-NEXT: mov r11, r0 ; CHECK-NEXT: vmov r0, s18 ; CHECK-NEXT: vldr s22, .LCPI29_0 -; CHECK-NEXT: mov r9, r1 +; CHECK-NEXT: mov r10, r1 ; CHECK-NEXT: vldr s20, .LCPI29_1 -; CHECK-NEXT: vmov r8, s16 +; CHECK-NEXT: vmov r9, s17 ; CHECK-NEXT: vcmp.f32 s19, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r10, #0 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r10, #-2147483648 +; CHECK-NEXT: movlt.w r11, #0 ; CHECK-NEXT: vcmp.f32 s19, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r10, #-1 +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r11, #-1 +; CHECK-NEXT: mvngt r10, #-2147483648 ; CHECK-NEXT: vcmp.f32 s19, s19 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vmov r4, s17 -; CHECK-NEXT: it vs +; CHECK-NEXT: vmov r8, s16 +; CHECK-NEXT: itt vs ; CHECK-NEXT: movvs.w r10, #0 +; CHECK-NEXT: movvs.w r11, #0 ; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s18, s22 ; CHECK-NEXT: mov r7, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r7, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r9, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r9, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r6, r1 ; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r9, #0 +; CHECK-NEXT: mov r6, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r7, #0 ; CHECK-NEXT: movlt.w r6, #-2147483648 ; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r0, r4 -; CHECK-NEXT: it gt +; CHECK-NEXT: mov r0, r9 +; CHECK-NEXT: itt gt ; CHECK-NEXT: mvngt r6, #-2147483648 +; CHECK-NEXT: movgt.w r7, #-1 ; CHECK-NEXT: vcmp.f32 s18, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r7, #0 ; CHECK-NEXT: movvs r6, #0 ; CHECK-NEXT: bl __aeabi_f2lz ; CHECK-NEXT: mov r5, r0 ; CHECK-NEXT: vcmp.f32 s17, s22 +; CHECK-NEXT: mov r4, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r4, #-2147483648 ; CHECK-NEXT: movlt r5, #0 ; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: mov r0, r8 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: mov r0, r8 +; CHECK-NEXT: itt gt ; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: mvngt r4, #-2147483648 ; CHECK-NEXT: vcmp.f32 s17, s17 -; CHECK-NEXT: mov r4, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r4, #0 ; CHECK-NEXT: movvs r5, #0 ; CHECK-NEXT: bl __aeabi_f2lz ; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: vmov q1[2], q1[0], r7, r10 +; CHECK-NEXT: vmov q1[2], q1[0], r7, r11 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r4, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s17 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r4, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt ; CHECK-NEXT: movlt.w r1, #-2147483648 ; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 +; CHECK-NEXT: vmov q1[3], q1[1], r6, r10 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt ; CHECK-NEXT: mvngt r1, #-2147483648 +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: vmov q1[3], q1[1], r6, r9 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r0, #0 ; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 ; CHECK-NEXT: vmov q0[3], q0[1], r1, r4 ; CHECK-NEXT: vpop {d8, d9, d10, d11} -; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, pc} +; CHECK-NEXT: add sp, #4 +; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} ; CHECK-NEXT: .p2align 2 ; CHECK-NEXT: @ %bb.1: ; CHECK-NEXT: .LCPI29_0: @@ -2118,568 +2025,370 @@ define arm_aapcs_vfpcc <4 x i100> @test_signed_v4f32_v4i100(<4 x float> %f) { ; CHECK-NEXT: sub sp, #4 ; CHECK-NEXT: .vsave {d8, d9, d10, d11} ; CHECK-NEXT: vpush {d8, d9, d10, d11} +; CHECK-NEXT: .pad #8 +; CHECK-NEXT: sub sp, #8 ; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: mov r9, r0 -; CHECK-NEXT: vmov r0, s18 +; CHECK-NEXT: mov r4, r0 +; CHECK-NEXT: vmov r0, s17 ; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: mov r10, r3 -; CHECK-NEXT: vmov r3, s16 +; CHECK-NEXT: mov r7, r0 +; CHECK-NEXT: vmov r0, s18 ; CHECK-NEXT: vldr s22, .LCPI30_0 -; CHECK-NEXT: vmov r7, s17 +; CHECK-NEXT: mov r11, r1 ; CHECK-NEXT: vldr s20, .LCPI30_1 -; CHECK-NEXT: vmov r4, s19 -; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: mov r10, r2 +; CHECK-NEXT: vcmp.f32 s17, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: mvnlt r3, #7 +; CHECK-NEXT: movlt.w r10, #0 +; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: movlt.w r11, #0 +; CHECK-NEXT: vcmp.f32 s17, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r11, #-1 +; CHECK-NEXT: movgt.w r7, #-1 +; CHECK-NEXT: movgt.w r10, #-1 +; CHECK-NEXT: movgt r3, #7 +; CHECK-NEXT: vcmp.f32 s17, s17 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s22 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: str.w r2, [r9, #33] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: str r3, [sp, #4] @ 4-byte Spill +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs.w r10, #0 +; CHECK-NEXT: movvs r7, #0 +; CHECK-NEXT: str r7, [sp] @ 4-byte Spill ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: movvs.w r11, #0 +; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: mov r5, r3 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r9, #29] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: mvnlt r5, #7 ; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt r5, #7 ; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: vcmp.f32 s18, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: str.w r2, [r4, #33] +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: str.w r1, [r4, #29] +; CHECK-NEXT: vmov r1, s19 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str.w r0, [r9, #25] -; CHECK-NEXT: mov r0, r3 +; CHECK-NEXT: str.w r0, [r4, #25] +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r5, #0 +; CHECK-NEXT: mov r0, r1 ; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: mov r11, r3 +; CHECK-NEXT: mov r7, r0 +; CHECK-NEXT: vmov r0, s16 +; CHECK-NEXT: vcmp.f32 s19, s22 +; CHECK-NEXT: mov r9, r1 +; CHECK-NEXT: mov r6, r2 +; CHECK-NEXT: mov r8, r3 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: mvnlt r8, #7 +; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: movlt.w r9, #0 +; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: vcmp.f32 s19, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r7, #-1 +; CHECK-NEXT: movgt.w r9, #-1 +; CHECK-NEXT: movgt.w r6, #-1 +; CHECK-NEXT: movgt.w r8, #7 +; CHECK-NEXT: vcmp.f32 s19, s19 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: itttt vs +; CHECK-NEXT: movvs.w r8, #0 +; CHECK-NEXT: movvs r6, #0 +; CHECK-NEXT: movvs.w r9, #0 +; CHECK-NEXT: movvs r7, #0 +; CHECK-NEXT: bl __fixsfti ; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: str.w r2, [r9, #8] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r9, #4] -; CHECK-NEXT: it lt ; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: mvnlt r3, #7 ; CHECK-NEXT: vcmp.f32 s16, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt r3, #7 ; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: vcmp.f32 s16, s16 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: str r2, [r4, #8] +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: str r1, [r4, #4] +; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str.w r0, [r9] -; CHECK-NEXT: mov r0, r4 +; CHECK-NEXT: str r0, [r4] +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: lsrl r0, r9, #28 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: orr.w r1, r9, r6, lsl #4 +; CHECK-NEXT: str.w r1, [r4, #45] +; CHECK-NEXT: and r1, r8, #15 +; CHECK-NEXT: str.w r0, [r4, #41] +; CHECK-NEXT: and r0, r5, #15 +; CHECK-NEXT: lsrl r6, r1, #28 +; CHECK-NEXT: strb.w r6, [r4, #49] +; CHECK-NEXT: orr.w r0, r0, r7, lsl #4 +; CHECK-NEXT: str.w r0, [r4, #37] +; CHECK-NEXT: ldr r2, [sp] @ 4-byte Reload +; CHECK-NEXT: mov r0, r2 +; CHECK-NEXT: lsrl r0, r11, #28 +; CHECK-NEXT: orr.w r1, r11, r10, lsl #4 +; CHECK-NEXT: strd r0, r1, [r4, #16] +; CHECK-NEXT: ldr r0, [sp, #4] @ 4-byte Reload +; CHECK-NEXT: and r1, r0, #15 +; CHECK-NEXT: lsrl r10, r1, #28 +; CHECK-NEXT: strb.w r10, [r4, #24] +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: and r0, r3, #15 +; CHECK-NEXT: orr.w r0, r0, r2, lsl #4 +; CHECK-NEXT: str r0, [r4, #12] +; CHECK-NEXT: add sp, #8 +; CHECK-NEXT: vpop {d8, d9, d10, d11} +; CHECK-NEXT: add sp, #4 +; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} +; CHECK-NEXT: .p2align 2 +; CHECK-NEXT: @ %bb.1: +; CHECK-NEXT: .LCPI30_0: +; CHECK-NEXT: .long 0xf1000000 @ float -6.338253E+29 +; CHECK-NEXT: .LCPI30_1: +; CHECK-NEXT: .long 0x70ffffff @ float 6.33825262E+29 + %x = call <4 x i100> @llvm.fptosi.sat.v4f32.v4i100(<4 x float> %f) + ret <4 x i100> %x +} + +define arm_aapcs_vfpcc <4 x i128> @test_signed_v4f32_v4i128(<4 x float> %f) { +; CHECK-LABEL: test_signed_v4f32_v4i128: +; CHECK: @ %bb.0: +; CHECK-NEXT: .save {r4, r5, r6, r7, lr} +; CHECK-NEXT: push {r4, r5, r6, r7, lr} +; CHECK-NEXT: .pad #4 +; CHECK-NEXT: sub sp, #4 +; CHECK-NEXT: .vsave {d8, d9, d10, d11} +; CHECK-NEXT: vpush {d8, d9, d10, d11} +; CHECK-NEXT: vmov q4, q0 +; CHECK-NEXT: mov r4, r0 +; CHECK-NEXT: vmov r0, s19 ; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: mov r5, r0 +; CHECK-NEXT: vmov r0, s18 +; CHECK-NEXT: vldr s22, .LCPI31_0 +; CHECK-NEXT: vmov r7, s16 +; CHECK-NEXT: vldr s20, .LCPI31_1 +; CHECK-NEXT: vmov r6, s17 ; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: mov r6, r0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r5, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: mvngt r3, #-2147483648 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s22 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r6, #0 +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: ittt vs +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: movvs r5, #0 +; CHECK-NEXT: strd r5, r1, [r4, #48] +; CHECK-NEXT: strd r2, r3, [r4, #56] +; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: add.w r12, r4, #32 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r10, #7 ; CHECK-NEXT: vcmp.f32 s18, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r10, #7 ; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: mov r5, r1 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: mvngt r3, #-2147483648 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: ittt vs +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: movvs r1, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r10, #0 -; CHECK-NEXT: and r0, r10, #15 -; CHECK-NEXT: mov r4, r2 -; CHECK-NEXT: orr.w r0, r0, r6, lsl #4 -; CHECK-NEXT: str.w r0, [r9, #37] -; CHECK-NEXT: mov r0, r7 -; CHECK-NEXT: mov r8, r3 +; CHECK-NEXT: movvs r0, #0 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r6 ; CHECK-NEXT: bl __fixsfti ; CHECK-NEXT: vcmp.f32 s17, s22 +; CHECK-NEXT: add.w r12, r4, #16 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s17, s17 -; CHECK-NEXT: it gt +; CHECK-NEXT: itttt gt ; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: mvngt r3, #-2147483648 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: ittt vs +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: movvs r1, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s22 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: vcmp.f32 s16, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r11, #7 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r11, #7 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r11, #0 -; CHECK-NEXT: and r7, r11, #15 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: orr.w r7, r7, r0, lsl #4 -; CHECK-NEXT: str.w r7, [r9, #12] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r5, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r5, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r4, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: lsrl r6, r5, #28 -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r4, #0 -; CHECK-NEXT: orr.w r7, r5, r4, lsl #4 -; CHECK-NEXT: str.w r7, [r9, #45] -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r6, [r9, #41] -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r8, #7 -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r8, #7 -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r8, #0 -; CHECK-NEXT: and r5, r8, #15 -; CHECK-NEXT: vcmp.f32 s17, s22 -; CHECK-NEXT: lsrl r4, r5, #28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: strb.w r4, [r9, #49] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s17 -; CHECK-NEXT: it gt +; CHECK-NEXT: vcmp.f32 s16, s16 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: lsrl r0, r1, #28 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vcmp.f32 s17, s17 +; CHECK-NEXT: mvngt r3, #-2147483648 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: ittt vs +; CHECK-NEXT: movvs r3, #0 ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: orr.w r1, r1, r2, lsl #4 -; CHECK-NEXT: vcmp.f32 s17, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: strd r0, r1, [r9, #16] -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r3, #7 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r3, #7 -; CHECK-NEXT: vcmp.f32 s17, s17 +; CHECK-NEXT: movvs r1, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r3, #0 -; CHECK-NEXT: and r1, r3, #15 -; CHECK-NEXT: lsrl r2, r1, #28 -; CHECK-NEXT: strb.w r2, [r9, #24] +; CHECK-NEXT: movvs r0, #0 +; CHECK-NEXT: stm r4!, {r0, r1, r2, r3} ; CHECK-NEXT: vpop {d8, d9, d10, d11} ; CHECK-NEXT: add sp, #4 -; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} +; CHECK-NEXT: pop {r4, r5, r6, r7, pc} ; CHECK-NEXT: .p2align 2 ; CHECK-NEXT: @ %bb.1: -; CHECK-NEXT: .LCPI30_0: -; CHECK-NEXT: .long 0xf1000000 @ float -6.338253E+29 -; CHECK-NEXT: .LCPI30_1: -; CHECK-NEXT: .long 0x70ffffff @ float 6.33825262E+29 - %x = call <4 x i100> @llvm.fptosi.sat.v4f32.v4i100(<4 x float> %f) - ret <4 x i100> %x +; CHECK-NEXT: .LCPI31_0: +; CHECK-NEXT: .long 0xff000000 @ float -1.70141183E+38 +; CHECK-NEXT: .LCPI31_1: +; CHECK-NEXT: .long 0x7effffff @ float 1.70141173E+38 + %x = call <4 x i128> @llvm.fptosi.sat.v4f32.v4i128(<4 x float> %f) + ret <4 x i128> %x } -define arm_aapcs_vfpcc <4 x i128> @test_signed_v4f32_v4i128(<4 x float> %f) { -; CHECK-LABEL: test_signed_v4f32_v4i128: +; +; 2-Vector double to signed integer -- result size variation +; + +declare <2 x i1> @llvm.fptosi.sat.v2f64.v2i1 (<2 x double>) +declare <2 x i8> @llvm.fptosi.sat.v2f64.v2i8 (<2 x double>) +declare <2 x i13> @llvm.fptosi.sat.v2f64.v2i13 (<2 x double>) +declare <2 x i16> @llvm.fptosi.sat.v2f64.v2i16 (<2 x double>) +declare <2 x i19> @llvm.fptosi.sat.v2f64.v2i19 (<2 x double>) +declare <2 x i50> @llvm.fptosi.sat.v2f64.v2i50 (<2 x double>) +declare <2 x i64> @llvm.fptosi.sat.v2f64.v2i64 (<2 x double>) +declare <2 x i100> @llvm.fptosi.sat.v2f64.v2i100(<2 x double>) +declare <2 x i128> @llvm.fptosi.sat.v2f64.v2i128(<2 x double>) + +define arm_aapcs_vfpcc <2 x i1> @test_signed_v2f64_v2i1(<2 x double> %f) { +; CHECK-LABEL: test_signed_v2f64_v2i1: ; CHECK: @ %bb.0: -; CHECK-NEXT: .save {r4, r5, r6, r7, lr} -; CHECK-NEXT: push {r4, r5, r6, r7, lr} +; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, r11, lr} +; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} ; CHECK-NEXT: .pad #4 ; CHECK-NEXT: sub sp, #4 -; CHECK-NEXT: .vsave {d8, d9, d10, d11} -; CHECK-NEXT: vpush {d8, d9, d10, d11} +; CHECK-NEXT: .vsave {d8, d9} +; CHECK-NEXT: vpush {d8, d9} +; CHECK-NEXT: .pad #24 +; CHECK-NEXT: sub sp, #24 ; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: mov r4, r0 -; CHECK-NEXT: vmov r0, s19 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vmov r5, s18 -; CHECK-NEXT: vldr s22, .LCPI31_0 -; CHECK-NEXT: vldr s20, .LCPI31_1 -; CHECK-NEXT: vmov r7, s16 -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r3, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r3, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r3, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: str r3, [r4, #60] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: str r2, [r4, #56] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #52] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4, #48] -; CHECK-NEXT: mov r0, r5 -; CHECK-NEXT: vmov r6, s17 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r3, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r3, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r3, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: str r3, [r4, #44] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: str r2, [r4, #40] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #36] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4, #32] -; CHECK-NEXT: mov r0, r6 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s17, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r3, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s17 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r3, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r3, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: str r3, [r4, #28] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s17 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: str r2, [r4, #24] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s17 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #20] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s17, s17 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4, #16] -; CHECK-NEXT: mov r0, r7 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r3, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r3, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r3, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: str r3, [r4, #12] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: str r2, [r4, #8] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #4] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4] -; CHECK-NEXT: vpop {d8, d9, d10, d11} -; CHECK-NEXT: add sp, #4 -; CHECK-NEXT: pop {r4, r5, r6, r7, pc} -; CHECK-NEXT: .p2align 2 -; CHECK-NEXT: @ %bb.1: -; CHECK-NEXT: .LCPI31_0: -; CHECK-NEXT: .long 0xff000000 @ float -1.70141183E+38 -; CHECK-NEXT: .LCPI31_1: -; CHECK-NEXT: .long 0x7effffff @ float 1.70141173E+38 - %x = call <4 x i128> @llvm.fptosi.sat.v4f32.v4i128(<4 x float> %f) - ret <4 x i128> %x -} - -; -; 2-Vector double to signed integer -- result size variation -; - -declare <2 x i1> @llvm.fptosi.sat.v2f64.v2i1 (<2 x double>) -declare <2 x i8> @llvm.fptosi.sat.v2f64.v2i8 (<2 x double>) -declare <2 x i13> @llvm.fptosi.sat.v2f64.v2i13 (<2 x double>) -declare <2 x i16> @llvm.fptosi.sat.v2f64.v2i16 (<2 x double>) -declare <2 x i19> @llvm.fptosi.sat.v2f64.v2i19 (<2 x double>) -declare <2 x i50> @llvm.fptosi.sat.v2f64.v2i50 (<2 x double>) -declare <2 x i64> @llvm.fptosi.sat.v2f64.v2i64 (<2 x double>) -declare <2 x i100> @llvm.fptosi.sat.v2f64.v2i100(<2 x double>) -declare <2 x i128> @llvm.fptosi.sat.v2f64.v2i128(<2 x double>) - -define arm_aapcs_vfpcc <2 x i1> @test_signed_v2f64_v2i1(<2 x double> %f) { -; CHECK-LABEL: test_signed_v2f64_v2i1: -; CHECK: @ %bb.0: -; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, r11, lr} -; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} -; CHECK-NEXT: .pad #4 -; CHECK-NEXT: sub sp, #4 -; CHECK-NEXT: .vsave {d8, d9} -; CHECK-NEXT: vpush {d8, d9} -; CHECK-NEXT: .pad #24 -; CHECK-NEXT: sub sp, #24 -; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: vldr d0, .LCPI32_0 -; CHECK-NEXT: vmov r8, r7, d8 -; CHECK-NEXT: str r0, [sp, #20] @ 4-byte Spill -; CHECK-NEXT: vmov r2, r3, d0 -; CHECK-NEXT: mov r0, r8 -; CHECK-NEXT: mov r1, r7 -; CHECK-NEXT: strd r2, r3, [sp, #12] @ 8-byte Folded Spill -; CHECK-NEXT: bl __aeabi_dcmpgt -; CHECK-NEXT: vldr d0, .LCPI32_1 -; CHECK-NEXT: mov r9, r0 -; CHECK-NEXT: mov r0, r8 -; CHECK-NEXT: mov r1, r7 -; CHECK-NEXT: vmov r2, r3, d0 -; CHECK-NEXT: strd r2, r3, [sp, #4] @ 8-byte Folded Spill -; CHECK-NEXT: bl __aeabi_dcmpge -; CHECK-NEXT: mov r10, r0 -; CHECK-NEXT: mov r0, r8 -; CHECK-NEXT: mov r1, r7 -; CHECK-NEXT: bl __aeabi_d2iz -; CHECK-NEXT: mov r11, r0 -; CHECK-NEXT: cmp.w r10, #0 -; CHECK-NEXT: it eq -; CHECK-NEXT: moveq.w r11, #-1 -; CHECK-NEXT: mov r0, r8 -; CHECK-NEXT: mov r1, r7 -; CHECK-NEXT: mov r2, r8 -; CHECK-NEXT: mov r3, r7 -; CHECK-NEXT: cmp.w r9, #0 -; CHECK-NEXT: vmov r6, r5, d9 -; CHECK-NEXT: it ne -; CHECK-NEXT: movne.w r11, #0 -; CHECK-NEXT: bl __aeabi_dcmpun -; CHECK-NEXT: cmp r0, #0 -; CHECK-NEXT: it ne -; CHECK-NEXT: movne.w r11, #0 -; CHECK-NEXT: and r0, r11, #1 -; CHECK-NEXT: ldrd r2, r3, [sp, #12] @ 8-byte Folded Reload -; CHECK-NEXT: rsbs r0, r0, #0 -; CHECK-NEXT: movs r4, #0 -; CHECK-NEXT: bfi r4, r0, #0, #1 +; CHECK-NEXT: vldr d0, .LCPI32_0 +; CHECK-NEXT: vmov r8, r7, d8 +; CHECK-NEXT: str r0, [sp, #20] @ 4-byte Spill +; CHECK-NEXT: vmov r2, r3, d0 +; CHECK-NEXT: mov r0, r8 +; CHECK-NEXT: mov r1, r7 +; CHECK-NEXT: strd r2, r3, [sp, #12] @ 8-byte Folded Spill +; CHECK-NEXT: bl __aeabi_dcmpgt +; CHECK-NEXT: vldr d0, .LCPI32_1 +; CHECK-NEXT: mov r9, r0 +; CHECK-NEXT: mov r0, r8 +; CHECK-NEXT: mov r1, r7 +; CHECK-NEXT: vmov r2, r3, d0 +; CHECK-NEXT: strd r2, r3, [sp, #4] @ 8-byte Folded Spill +; CHECK-NEXT: bl __aeabi_dcmpge +; CHECK-NEXT: mov r10, r0 +; CHECK-NEXT: mov r0, r8 +; CHECK-NEXT: mov r1, r7 +; CHECK-NEXT: bl __aeabi_d2iz +; CHECK-NEXT: mov r11, r0 +; CHECK-NEXT: cmp.w r10, #0 +; CHECK-NEXT: it eq +; CHECK-NEXT: moveq.w r11, #-1 +; CHECK-NEXT: mov r0, r8 +; CHECK-NEXT: mov r1, r7 +; CHECK-NEXT: mov r2, r8 +; CHECK-NEXT: mov r3, r7 +; CHECK-NEXT: cmp.w r9, #0 +; CHECK-NEXT: vmov r6, r5, d9 +; CHECK-NEXT: it ne +; CHECK-NEXT: movne.w r11, #0 +; CHECK-NEXT: bl __aeabi_dcmpun +; CHECK-NEXT: cmp r0, #0 +; CHECK-NEXT: it ne +; CHECK-NEXT: movne.w r11, #0 +; CHECK-NEXT: and r0, r11, #1 +; CHECK-NEXT: ldrd r2, r3, [sp, #12] @ 8-byte Folded Reload +; CHECK-NEXT: rsbs r0, r0, #0 +; CHECK-NEXT: movs r4, #0 +; CHECK-NEXT: bfi r4, r0, #0, #1 ; CHECK-NEXT: mov r0, r6 ; CHECK-NEXT: mov r1, r5 ; CHECK-NEXT: bl __aeabi_dcmpgt @@ -4244,33 +3953,37 @@ define arm_aapcs_vfpcc <8 x i1> @test_signed_v8f16_v8i1(<8 x half> %f) { ; CHECK-NEXT: .vsave {d8} ; CHECK-NEXT: vpush {d8} ; CHECK-NEXT: vcvtb.f32.f16 s15, s0 -; CHECK-NEXT: vmov.f32 s5, #-1.000000e+00 -; CHECK-NEXT: vldr s7, .LCPI42_0 -; CHECK-NEXT: vmaxnm.f32 s16, s15, s5 +; CHECK-NEXT: vmov.f32 s7, #-1.000000e+00 +; CHECK-NEXT: vldr s5, .LCPI42_0 +; CHECK-NEXT: vmaxnm.f32 s16, s15, s7 ; CHECK-NEXT: vcvtt.f32.f16 s12, s2 ; CHECK-NEXT: vcvtt.f32.f16 s9, s1 -; CHECK-NEXT: vminnm.f32 s16, s16, s7 +; CHECK-NEXT: vminnm.f32 s16, s16, s5 ; CHECK-NEXT: vcvtt.f32.f16 s4, s3 ; CHECK-NEXT: vcvt.s32.f32 s16, s16 ; CHECK-NEXT: vcvtb.f32.f16 s8, s3 ; CHECK-NEXT: vcvtb.f32.f16 s2, s2 ; CHECK-NEXT: vcvtb.f32.f16 s1, s1 ; CHECK-NEXT: vcvtt.f32.f16 s0, s0 -; CHECK-NEXT: vmaxnm.f32 s6, s4, s5 -; CHECK-NEXT: vmaxnm.f32 s10, s8, s5 -; CHECK-NEXT: vmaxnm.f32 s14, s12, s5 -; CHECK-NEXT: vmaxnm.f32 s3, s2, s5 -; CHECK-NEXT: vmaxnm.f32 s11, s9, s5 -; CHECK-NEXT: vmaxnm.f32 s13, s1, s5 -; CHECK-NEXT: vmaxnm.f32 s5, s0, s5 -; CHECK-NEXT: vminnm.f32 s5, s5, s7 -; CHECK-NEXT: vminnm.f32 s13, s13, s7 -; CHECK-NEXT: vcvt.s32.f32 s5, s5 -; CHECK-NEXT: movs r1, #0 +; CHECK-NEXT: vmaxnm.f32 s6, s4, s7 +; CHECK-NEXT: vmaxnm.f32 s10, s8, s7 +; CHECK-NEXT: vmaxnm.f32 s14, s12, s7 +; CHECK-NEXT: vmaxnm.f32 s3, s2, s7 +; CHECK-NEXT: vmaxnm.f32 s11, s9, s7 +; CHECK-NEXT: vmaxnm.f32 s13, s1, s7 +; CHECK-NEXT: vmaxnm.f32 s7, s0, s7 +; CHECK-NEXT: vminnm.f32 s6, s6, s5 +; CHECK-NEXT: vminnm.f32 s10, s10, s5 +; CHECK-NEXT: vminnm.f32 s14, s14, s5 +; CHECK-NEXT: vminnm.f32 s3, s3, s5 +; CHECK-NEXT: vminnm.f32 s11, s11, s5 +; CHECK-NEXT: vminnm.f32 s13, s13, s5 +; CHECK-NEXT: vminnm.f32 s5, s7, s5 ; CHECK-NEXT: vcmp.f32 s15, s15 -; CHECK-NEXT: vminnm.f32 s11, s11, s7 -; CHECK-NEXT: vmov r2, s16 +; CHECK-NEXT: vcvt.s32.f32 s5, s5 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: vmov r2, s16 +; CHECK-NEXT: mov.w r1, #0 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 ; CHECK-NEXT: vcvt.s32.f32 s13, s13 @@ -4280,60 +3993,56 @@ define arm_aapcs_vfpcc <8 x i1> @test_signed_v8f16_v8i1(<8 x half> %f) { ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: bfi r1, r2, #0, #1 ; CHECK-NEXT: vcvt.s32.f32 s11, s11 +; CHECK-NEXT: vcmp.f32 s1, s1 ; CHECK-NEXT: vmov r2, s5 -; CHECK-NEXT: vminnm.f32 s3, s3, s7 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vcmp.f32 s1, s1 -; CHECK-NEXT: and r2, r2, #1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: rsb.w r2, r2, #0 +; CHECK-NEXT: and r2, r2, #1 ; CHECK-NEXT: vcvt.s32.f32 s3, s3 +; CHECK-NEXT: rsb.w r2, r2, #0 +; CHECK-NEXT: vcmp.f32 s9, s9 ; CHECK-NEXT: bfi r1, r2, #1, #1 ; CHECK-NEXT: vmov r2, s13 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vminnm.f32 s14, s14, s7 -; CHECK-NEXT: and r2, r2, #1 -; CHECK-NEXT: vcmp.f32 s9, s9 -; CHECK-NEXT: rsbs r2, r2, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: and r2, r2, #1 +; CHECK-NEXT: vcvt.s32.f32 s14, s14 +; CHECK-NEXT: rsb.w r2, r2, #0 +; CHECK-NEXT: vcmp.f32 s2, s2 ; CHECK-NEXT: bfi r1, r2, #2, #1 ; CHECK-NEXT: vmov r2, s11 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vcvt.s32.f32 s14, s14 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: and r2, r2, #1 -; CHECK-NEXT: vminnm.f32 s10, s10, s7 -; CHECK-NEXT: rsbs r2, r2, #0 -; CHECK-NEXT: vcmp.f32 s2, s2 +; CHECK-NEXT: vcvt.s32.f32 s10, s10 +; CHECK-NEXT: rsb.w r2, r2, #0 +; CHECK-NEXT: vcmp.f32 s12, s12 ; CHECK-NEXT: bfi r1, r2, #3, #1 ; CHECK-NEXT: vmov r2, s3 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vcvt.s32.f32 s10, s10 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: and r2, r2, #1 -; CHECK-NEXT: rsbs r2, r2, #0 -; CHECK-NEXT: vminnm.f32 s6, s6, s7 +; CHECK-NEXT: vcvt.s32.f32 s6, s6 +; CHECK-NEXT: rsb.w r2, r2, #0 +; CHECK-NEXT: vcmp.f32 s8, s8 ; CHECK-NEXT: bfi r1, r2, #4, #1 -; CHECK-NEXT: vcmp.f32 s12, s12 ; CHECK-NEXT: vmov r2, s14 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vcvt.s32.f32 s6, s6 -; CHECK-NEXT: and r2, r2, #1 -; CHECK-NEXT: vcmp.f32 s8, s8 -; CHECK-NEXT: rsbs r2, r2, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: and r2, r2, #1 +; CHECK-NEXT: vcmp.f32 s4, s4 +; CHECK-NEXT: rsb.w r2, r2, #0 ; CHECK-NEXT: bfi r1, r2, #5, #1 ; CHECK-NEXT: vmov r2, s10 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vcmp.f32 s4, s4 -; CHECK-NEXT: and r2, r2, #1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: and r2, r2, #1 ; CHECK-NEXT: rsb.w r2, r2, #0 ; CHECK-NEXT: bfi r1, r2, #6, #1 ; CHECK-NEXT: vmov r2, s6 @@ -4360,40 +4069,40 @@ define arm_aapcs_vfpcc <8 x i8> @test_signed_v8f16_v8i8(<8 x half> %f) { ; CHECK-MVE-NEXT: push {r4, r5, r7, lr} ; CHECK-MVE-NEXT: .vsave {d8} ; CHECK-MVE-NEXT: vpush {d8} -; CHECK-MVE-NEXT: vldr s8, .LCPI43_1 +; CHECK-MVE-NEXT: vldr s6, .LCPI43_1 ; CHECK-MVE-NEXT: vcvtt.f32.f16 s13, s3 ; CHECK-MVE-NEXT: vcvtb.f32.f16 s3, s3 -; CHECK-MVE-NEXT: vldr s6, .LCPI43_0 -; CHECK-MVE-NEXT: vmaxnm.f32 s16, s3, s8 -; CHECK-MVE-NEXT: vcvtt.f32.f16 s4, s0 -; CHECK-MVE-NEXT: vcvtt.f32.f16 s12, s1 +; CHECK-MVE-NEXT: vldr s4, .LCPI43_0 +; CHECK-MVE-NEXT: vmaxnm.f32 s16, s3, s6 ; CHECK-MVE-NEXT: vcvtt.f32.f16 s7, s2 -; CHECK-MVE-NEXT: vmaxnm.f32 s15, s13, s8 -; CHECK-MVE-NEXT: vminnm.f32 s16, s16, s6 -; CHECK-MVE-NEXT: vcvtb.f32.f16 s0, s0 -; CHECK-MVE-NEXT: vcvtb.f32.f16 s1, s1 +; CHECK-MVE-NEXT: vmaxnm.f32 s15, s13, s6 +; CHECK-MVE-NEXT: vminnm.f32 s16, s16, s4 ; CHECK-MVE-NEXT: vcvtb.f32.f16 s2, s2 -; CHECK-MVE-NEXT: vmaxnm.f32 s10, s4, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s14, s12, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s5, s0, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s9, s7, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s11, s1, s8 -; CHECK-MVE-NEXT: vminnm.f32 s15, s15, s6 +; CHECK-MVE-NEXT: vminnm.f32 s15, s15, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s11, s2, s6 ; CHECK-MVE-NEXT: vcvt.s32.f32 s16, s16 -; CHECK-MVE-NEXT: vmaxnm.f32 s8, s2, s8 -; CHECK-MVE-NEXT: vminnm.f32 s10, s10, s6 -; CHECK-MVE-NEXT: vminnm.f32 s14, s14, s6 -; CHECK-MVE-NEXT: vminnm.f32 s5, s5, s6 -; CHECK-MVE-NEXT: vminnm.f32 s9, s9, s6 -; CHECK-MVE-NEXT: vminnm.f32 s11, s11, s6 -; CHECK-MVE-NEXT: vminnm.f32 s6, s8, s6 +; CHECK-MVE-NEXT: vcvtt.f32.f16 s12, s1 +; CHECK-MVE-NEXT: vmaxnm.f32 s9, s7, s6 +; CHECK-MVE-NEXT: vminnm.f32 s11, s11, s4 +; CHECK-MVE-NEXT: vcvtb.f32.f16 s1, s1 ; CHECK-MVE-NEXT: vcvt.s32.f32 s15, s15 -; CHECK-MVE-NEXT: vcvt.s32.f32 s6, s6 -; CHECK-MVE-NEXT: vcvt.s32.f32 s9, s9 +; CHECK-MVE-NEXT: vcvtb.f32.f16 s8, s0 +; CHECK-MVE-NEXT: vmaxnm.f32 s5, s1, s6 +; CHECK-MVE-NEXT: vminnm.f32 s9, s9, s4 ; CHECK-MVE-NEXT: vcvt.s32.f32 s11, s11 -; CHECK-MVE-NEXT: vcvt.s32.f32 s14, s14 +; CHECK-MVE-NEXT: vmaxnm.f32 s10, s8, s6 +; CHECK-MVE-NEXT: vmaxnm.f32 s14, s12, s6 +; CHECK-MVE-NEXT: vminnm.f32 s5, s5, s4 +; CHECK-MVE-NEXT: vcvt.s32.f32 s9, s9 +; CHECK-MVE-NEXT: vcvtt.f32.f16 s0, s0 +; CHECK-MVE-NEXT: vminnm.f32 s10, s10, s4 +; CHECK-MVE-NEXT: vminnm.f32 s14, s14, s4 ; CHECK-MVE-NEXT: vcvt.s32.f32 s5, s5 +; CHECK-MVE-NEXT: vmaxnm.f32 s6, s0, s6 +; CHECK-MVE-NEXT: vminnm.f32 s4, s6, s4 ; CHECK-MVE-NEXT: vcvt.s32.f32 s10, s10 +; CHECK-MVE-NEXT: vcvt.s32.f32 s14, s14 +; CHECK-MVE-NEXT: vcvt.s32.f32 s4, s4 ; CHECK-MVE-NEXT: vcmp.f32 s3, s3 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vmov r12, s16 @@ -4406,7 +4115,7 @@ define arm_aapcs_vfpcc <8 x i8> @test_signed_v8f16_v8i8(<8 x half> %f) { ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs.w lr, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r2, s6 +; CHECK-MVE-NEXT: vmov r2, s11 ; CHECK-MVE-NEXT: vcmp.f32 s7, s7 ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r2, #0 @@ -4416,22 +4125,22 @@ define arm_aapcs_vfpcc <8 x i8> @test_signed_v8f16_v8i8(<8 x half> %f) { ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r3, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r0, s11 +; CHECK-MVE-NEXT: vmov r0, s5 ; CHECK-MVE-NEXT: vcmp.f32 s12, s12 ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r0, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vmov r1, s14 -; CHECK-MVE-NEXT: vmov r4, s5 +; CHECK-MVE-NEXT: vmov r4, s10 ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r1, #0 -; CHECK-MVE-NEXT: vcmp.f32 s0, s0 +; CHECK-MVE-NEXT: vcmp.f32 s8, s8 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r4, #0 -; CHECK-MVE-NEXT: vcmp.f32 s4, s4 +; CHECK-MVE-NEXT: vcmp.f32 s0, s0 ; CHECK-MVE-NEXT: vmov.16 q0[0], r4 -; CHECK-MVE-NEXT: vmov r5, s10 +; CHECK-MVE-NEXT: vmov r5, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r5, #0 @@ -4468,40 +4177,40 @@ define arm_aapcs_vfpcc <8 x i13> @test_signed_v8f16_v8i13(<8 x half> %f) { ; CHECK-MVE-NEXT: push {r4, r5, r7, lr} ; CHECK-MVE-NEXT: .vsave {d8} ; CHECK-MVE-NEXT: vpush {d8} -; CHECK-MVE-NEXT: vldr s8, .LCPI44_1 +; CHECK-MVE-NEXT: vldr s6, .LCPI44_1 ; CHECK-MVE-NEXT: vcvtt.f32.f16 s13, s3 ; CHECK-MVE-NEXT: vcvtb.f32.f16 s3, s3 -; CHECK-MVE-NEXT: vldr s6, .LCPI44_0 -; CHECK-MVE-NEXT: vmaxnm.f32 s16, s3, s8 -; CHECK-MVE-NEXT: vcvtt.f32.f16 s4, s0 -; CHECK-MVE-NEXT: vcvtt.f32.f16 s12, s1 +; CHECK-MVE-NEXT: vldr s4, .LCPI44_0 +; CHECK-MVE-NEXT: vmaxnm.f32 s16, s3, s6 ; CHECK-MVE-NEXT: vcvtt.f32.f16 s7, s2 -; CHECK-MVE-NEXT: vmaxnm.f32 s15, s13, s8 -; CHECK-MVE-NEXT: vminnm.f32 s16, s16, s6 -; CHECK-MVE-NEXT: vcvtb.f32.f16 s0, s0 -; CHECK-MVE-NEXT: vcvtb.f32.f16 s1, s1 +; CHECK-MVE-NEXT: vmaxnm.f32 s15, s13, s6 +; CHECK-MVE-NEXT: vminnm.f32 s16, s16, s4 ; CHECK-MVE-NEXT: vcvtb.f32.f16 s2, s2 -; CHECK-MVE-NEXT: vmaxnm.f32 s10, s4, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s14, s12, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s5, s0, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s9, s7, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s11, s1, s8 -; CHECK-MVE-NEXT: vminnm.f32 s15, s15, s6 +; CHECK-MVE-NEXT: vminnm.f32 s15, s15, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s11, s2, s6 ; CHECK-MVE-NEXT: vcvt.s32.f32 s16, s16 -; CHECK-MVE-NEXT: vmaxnm.f32 s8, s2, s8 -; CHECK-MVE-NEXT: vminnm.f32 s10, s10, s6 -; CHECK-MVE-NEXT: vminnm.f32 s14, s14, s6 -; CHECK-MVE-NEXT: vminnm.f32 s5, s5, s6 -; CHECK-MVE-NEXT: vminnm.f32 s9, s9, s6 -; CHECK-MVE-NEXT: vminnm.f32 s11, s11, s6 -; CHECK-MVE-NEXT: vminnm.f32 s6, s8, s6 +; CHECK-MVE-NEXT: vcvtt.f32.f16 s12, s1 +; CHECK-MVE-NEXT: vmaxnm.f32 s9, s7, s6 +; CHECK-MVE-NEXT: vminnm.f32 s11, s11, s4 +; CHECK-MVE-NEXT: vcvtb.f32.f16 s1, s1 ; CHECK-MVE-NEXT: vcvt.s32.f32 s15, s15 -; CHECK-MVE-NEXT: vcvt.s32.f32 s6, s6 -; CHECK-MVE-NEXT: vcvt.s32.f32 s9, s9 +; CHECK-MVE-NEXT: vcvtb.f32.f16 s8, s0 +; CHECK-MVE-NEXT: vmaxnm.f32 s5, s1, s6 +; CHECK-MVE-NEXT: vminnm.f32 s9, s9, s4 ; CHECK-MVE-NEXT: vcvt.s32.f32 s11, s11 -; CHECK-MVE-NEXT: vcvt.s32.f32 s14, s14 +; CHECK-MVE-NEXT: vmaxnm.f32 s10, s8, s6 +; CHECK-MVE-NEXT: vmaxnm.f32 s14, s12, s6 +; CHECK-MVE-NEXT: vminnm.f32 s5, s5, s4 +; CHECK-MVE-NEXT: vcvt.s32.f32 s9, s9 +; CHECK-MVE-NEXT: vcvtt.f32.f16 s0, s0 +; CHECK-MVE-NEXT: vminnm.f32 s10, s10, s4 +; CHECK-MVE-NEXT: vminnm.f32 s14, s14, s4 ; CHECK-MVE-NEXT: vcvt.s32.f32 s5, s5 +; CHECK-MVE-NEXT: vmaxnm.f32 s6, s0, s6 +; CHECK-MVE-NEXT: vminnm.f32 s4, s6, s4 ; CHECK-MVE-NEXT: vcvt.s32.f32 s10, s10 +; CHECK-MVE-NEXT: vcvt.s32.f32 s14, s14 +; CHECK-MVE-NEXT: vcvt.s32.f32 s4, s4 ; CHECK-MVE-NEXT: vcmp.f32 s3, s3 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vmov r12, s16 @@ -4514,7 +4223,7 @@ define arm_aapcs_vfpcc <8 x i13> @test_signed_v8f16_v8i13(<8 x half> %f) { ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs.w lr, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r2, s6 +; CHECK-MVE-NEXT: vmov r2, s11 ; CHECK-MVE-NEXT: vcmp.f32 s7, s7 ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r2, #0 @@ -4524,22 +4233,22 @@ define arm_aapcs_vfpcc <8 x i13> @test_signed_v8f16_v8i13(<8 x half> %f) { ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r3, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r0, s11 +; CHECK-MVE-NEXT: vmov r0, s5 ; CHECK-MVE-NEXT: vcmp.f32 s12, s12 ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r0, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vmov r1, s14 -; CHECK-MVE-NEXT: vmov r4, s5 +; CHECK-MVE-NEXT: vmov r4, s10 ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r1, #0 -; CHECK-MVE-NEXT: vcmp.f32 s0, s0 +; CHECK-MVE-NEXT: vcmp.f32 s8, s8 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r4, #0 -; CHECK-MVE-NEXT: vcmp.f32 s4, s4 +; CHECK-MVE-NEXT: vcmp.f32 s0, s0 ; CHECK-MVE-NEXT: vmov.16 q0[0], r4 -; CHECK-MVE-NEXT: vmov r5, s10 +; CHECK-MVE-NEXT: vmov r5, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: it vs ; CHECK-MVE-NEXT: movvs r5, #0 @@ -4578,838 +4287,273 @@ define arm_aapcs_vfpcc <8 x i16> @test_signed_v8f16_v8i16(<8 x half> %f) { ; CHECK-MVE-NEXT: push {r4, r5, r7, lr} ; CHECK-MVE-NEXT: .vsave {d8} ; CHECK-MVE-NEXT: vpush {d8} -; CHECK-MVE-NEXT: vldr s8, .LCPI45_1 +; CHECK-MVE-NEXT: vldr s6, .LCPI45_1 ; CHECK-MVE-NEXT: vcvtt.f32.f16 s13, s3 -; CHECK-MVE-NEXT: vcvtb.f32.f16 s3, s3 -; CHECK-MVE-NEXT: vldr s6, .LCPI45_0 -; CHECK-MVE-NEXT: vmaxnm.f32 s16, s3, s8 -; CHECK-MVE-NEXT: vcvtt.f32.f16 s4, s0 -; CHECK-MVE-NEXT: vcvtt.f32.f16 s12, s1 -; CHECK-MVE-NEXT: vcvtt.f32.f16 s7, s2 -; CHECK-MVE-NEXT: vmaxnm.f32 s15, s13, s8 -; CHECK-MVE-NEXT: vminnm.f32 s16, s16, s6 -; CHECK-MVE-NEXT: vcvtb.f32.f16 s0, s0 -; CHECK-MVE-NEXT: vcvtb.f32.f16 s1, s1 -; CHECK-MVE-NEXT: vcvtb.f32.f16 s2, s2 -; CHECK-MVE-NEXT: vmaxnm.f32 s10, s4, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s14, s12, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s5, s0, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s9, s7, s8 -; CHECK-MVE-NEXT: vmaxnm.f32 s11, s1, s8 -; CHECK-MVE-NEXT: vminnm.f32 s15, s15, s6 -; CHECK-MVE-NEXT: vcvt.s32.f32 s16, s16 -; CHECK-MVE-NEXT: vmaxnm.f32 s8, s2, s8 -; CHECK-MVE-NEXT: vminnm.f32 s10, s10, s6 -; CHECK-MVE-NEXT: vminnm.f32 s14, s14, s6 -; CHECK-MVE-NEXT: vminnm.f32 s5, s5, s6 -; CHECK-MVE-NEXT: vminnm.f32 s9, s9, s6 -; CHECK-MVE-NEXT: vminnm.f32 s11, s11, s6 -; CHECK-MVE-NEXT: vminnm.f32 s6, s8, s6 -; CHECK-MVE-NEXT: vcvt.s32.f32 s15, s15 -; CHECK-MVE-NEXT: vcvt.s32.f32 s6, s6 -; CHECK-MVE-NEXT: vcvt.s32.f32 s9, s9 -; CHECK-MVE-NEXT: vcvt.s32.f32 s11, s11 -; CHECK-MVE-NEXT: vcvt.s32.f32 s14, s14 -; CHECK-MVE-NEXT: vcvt.s32.f32 s5, s5 -; CHECK-MVE-NEXT: vcvt.s32.f32 s10, s10 -; CHECK-MVE-NEXT: vcmp.f32 s3, s3 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r12, s16 -; CHECK-MVE-NEXT: vcmp.f32 s13, s13 -; CHECK-MVE-NEXT: it vs -; CHECK-MVE-NEXT: movvs.w r12, #0 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov lr, s15 -; CHECK-MVE-NEXT: vcmp.f32 s2, s2 -; CHECK-MVE-NEXT: it vs -; CHECK-MVE-NEXT: movvs.w lr, #0 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r2, s6 -; CHECK-MVE-NEXT: vcmp.f32 s7, s7 -; CHECK-MVE-NEXT: it vs -; CHECK-MVE-NEXT: movvs r2, #0 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r3, s9 -; CHECK-MVE-NEXT: vcmp.f32 s1, s1 -; CHECK-MVE-NEXT: it vs -; CHECK-MVE-NEXT: movvs r3, #0 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r0, s11 -; CHECK-MVE-NEXT: vcmp.f32 s12, s12 -; CHECK-MVE-NEXT: it vs -; CHECK-MVE-NEXT: movvs r0, #0 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmov r1, s14 -; CHECK-MVE-NEXT: vmov r4, s5 -; CHECK-MVE-NEXT: it vs -; CHECK-MVE-NEXT: movvs r1, #0 -; CHECK-MVE-NEXT: vcmp.f32 s0, s0 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: it vs -; CHECK-MVE-NEXT: movvs r4, #0 -; CHECK-MVE-NEXT: vcmp.f32 s4, s4 -; CHECK-MVE-NEXT: vmov.16 q0[0], r4 -; CHECK-MVE-NEXT: vmov r5, s10 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: it vs -; CHECK-MVE-NEXT: movvs r5, #0 -; CHECK-MVE-NEXT: vmov.16 q0[1], r5 -; CHECK-MVE-NEXT: vmov.16 q0[2], r0 -; CHECK-MVE-NEXT: vmov.16 q0[3], r1 -; CHECK-MVE-NEXT: vmov.16 q0[4], r2 -; CHECK-MVE-NEXT: vmov.16 q0[5], r3 -; CHECK-MVE-NEXT: vmov.16 q0[6], r12 -; CHECK-MVE-NEXT: vmov.16 q0[7], lr -; CHECK-MVE-NEXT: vpop {d8} -; CHECK-MVE-NEXT: pop {r4, r5, r7, pc} -; CHECK-MVE-NEXT: .p2align 2 -; CHECK-MVE-NEXT: @ %bb.1: -; CHECK-MVE-NEXT: .LCPI45_0: -; CHECK-MVE-NEXT: .long 0x46fffe00 @ float 32767 -; CHECK-MVE-NEXT: .LCPI45_1: -; CHECK-MVE-NEXT: .long 0xc7000000 @ float -32768 -; -; CHECK-MVEFP-LABEL: test_signed_v8f16_v8i16: -; CHECK-MVEFP: @ %bb.0: -; CHECK-MVEFP-NEXT: vcvt.s16.f16 q0, q0 -; CHECK-MVEFP-NEXT: bx lr - %x = call <8 x i16> @llvm.fptosi.sat.v8f16.v8i16(<8 x half> %f) - ret <8 x i16> %x -} - -define arm_aapcs_vfpcc <8 x i19> @test_signed_v8f16_v8i19(<8 x half> %f) { -; CHECK-LABEL: test_signed_v8f16_v8i19: -; CHECK: @ %bb.0: -; CHECK-NEXT: .save {r4, r5, r7, r9, r11, lr} -; CHECK-NEXT: push.w {r4, r5, r7, r9, r11, lr} -; CHECK-NEXT: vldr s4, .LCPI46_0 -; CHECK-NEXT: vcvtb.f32.f16 s8, s1 -; CHECK-NEXT: vcvtt.f32.f16 s12, s1 -; CHECK-NEXT: vcvtt.f32.f16 s1, s0 -; CHECK-NEXT: vldr s6, .LCPI46_1 -; CHECK-NEXT: vmaxnm.f32 s5, s1, s4 -; CHECK-NEXT: vcvtb.f32.f16 s0, s0 -; CHECK-NEXT: vmaxnm.f32 s14, s12, s4 -; CHECK-NEXT: vminnm.f32 s5, s5, s6 -; CHECK-NEXT: vmaxnm.f32 s7, s0, s4 -; CHECK-NEXT: vminnm.f32 s7, s7, s6 -; CHECK-NEXT: vcvt.s32.f32 s5, s5 -; CHECK-NEXT: vcvt.s32.f32 s7, s7 -; CHECK-NEXT: vminnm.f32 s14, s14, s6 -; CHECK-NEXT: vcvt.s32.f32 s14, s14 -; CHECK-NEXT: vmaxnm.f32 s10, s8, s4 -; CHECK-NEXT: vminnm.f32 s10, s10, s6 -; CHECK-NEXT: vcmp.f32 s1, s1 -; CHECK-NEXT: vcvt.s32.f32 s10, s10 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s0, s0 -; CHECK-NEXT: mov.w r7, #0 -; CHECK-NEXT: vcvtb.f32.f16 s0, s2 -; CHECK-NEXT: mov.w r9, #0 -; CHECK-NEXT: vmov r2, s5 -; CHECK-NEXT: mov.w r5, #0 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmov r1, s7 -; CHECK-NEXT: bfc r2, #19, #13 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vcmp.f32 s12, s12 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s8, s8 -; CHECK-NEXT: lsll r2, r7, #19 -; CHECK-NEXT: bfc r1, #19, #13 -; CHECK-NEXT: vmov r12, s14 -; CHECK-NEXT: vmaxnm.f32 s8, s0, s4 -; CHECK-NEXT: orr.w r1, r1, r2 -; CHECK-NEXT: str r1, [r0] -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r12, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s0, s0 -; CHECK-NEXT: vcvtt.f32.f16 s0, s2 -; CHECK-NEXT: vmaxnm.f32 s2, s0, s4 -; CHECK-NEXT: vminnm.f32 s8, s8, s6 -; CHECK-NEXT: vminnm.f32 s2, s2, s6 -; CHECK-NEXT: vmov r3, s10 -; CHECK-NEXT: vcvt.s32.f32 s2, s2 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r3, #0 -; CHECK-NEXT: vcvt.s32.f32 s8, s8 -; CHECK-NEXT: bfc r3, #19, #13 -; CHECK-NEXT: mov r2, r12 -; CHECK-NEXT: movs r1, #0 -; CHECK-NEXT: bfc r2, #19, #13 -; CHECK-NEXT: mov r4, r3 -; CHECK-NEXT: lsrl r2, r1, #7 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s0, s0 -; CHECK-NEXT: lsrl r4, r9, #26 -; CHECK-NEXT: vcvtt.f32.f16 s0, s3 -; CHECK-NEXT: mov lr, r1 -; CHECK-NEXT: orr.w r1, r4, r2 -; CHECK-NEXT: vmov r4, s2 -; CHECK-NEXT: vmaxnm.f32 s2, s0, s4 -; CHECK-NEXT: vmov r2, s8 -; CHECK-NEXT: vminnm.f32 s2, s2, s6 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vcvt.s32.f32 s2, s2 -; CHECK-NEXT: bfc r2, #19, #13 -; CHECK-NEXT: lsll r2, r5, #12 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r4, #0 -; CHECK-NEXT: orrs r2, r1 -; CHECK-NEXT: bfc r4, #19, #13 -; CHECK-NEXT: movs r1, #0 -; CHECK-NEXT: lsll r4, r1, #31 -; CHECK-NEXT: vcmp.f32 s0, s0 -; CHECK-NEXT: orrs r2, r4 -; CHECK-NEXT: str r2, [r0, #8] -; CHECK-NEXT: orr.w r2, r7, r3, lsl #6 -; CHECK-NEXT: vcvtb.f32.f16 s0, s3 -; CHECK-NEXT: orr.w r3, r2, r12, lsl #25 -; CHECK-NEXT: vmov r2, s2 -; CHECK-NEXT: vmaxnm.f32 s2, s0, s4 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vminnm.f32 s2, s2, s6 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vcvt.s32.f32 s2, s2 -; CHECK-NEXT: bfc r2, #19, #13 -; CHECK-NEXT: movs r7, #0 -; CHECK-NEXT: vcmp.f32 s0, s0 -; CHECK-NEXT: lsll r2, r7, #5 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov.w r11, #0 -; CHECK-NEXT: vmov r7, s2 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: mov r4, r7 -; CHECK-NEXT: bfc r4, #19, #13 -; CHECK-NEXT: lsrl r4, r11, #14 -; CHECK-NEXT: orrs r2, r4 -; CHECK-NEXT: strh r2, [r0, #16] -; CHECK-NEXT: str r3, [r0, #4] -; CHECK-NEXT: lsrs r2, r2, #16 -; CHECK-NEXT: strb r2, [r0, #18] -; CHECK-NEXT: orr.w r2, r9, lr -; CHECK-NEXT: orrs r2, r5 -; CHECK-NEXT: orrs r1, r2 -; CHECK-NEXT: orr.w r1, r1, r7, lsl #18 -; CHECK-NEXT: str r1, [r0, #12] -; CHECK-NEXT: pop.w {r4, r5, r7, r9, r11, pc} -; CHECK-NEXT: .p2align 2 -; CHECK-NEXT: @ %bb.1: -; CHECK-NEXT: .LCPI46_0: -; CHECK-NEXT: .long 0xc8800000 @ float -262144 -; CHECK-NEXT: .LCPI46_1: -; CHECK-NEXT: .long 0x487fffc0 @ float 262143 - %x = call <8 x i19> @llvm.fptosi.sat.v8f16.v8i19(<8 x half> %f) - ret <8 x i19> %x -} - -define arm_aapcs_vfpcc <8 x i32> @test_signed_v8f16_v8i32_duplicate(<8 x half> %f) { -; CHECK-LABEL: test_signed_v8f16_v8i32_duplicate: -; CHECK: @ %bb.0: -; CHECK-NEXT: vmovx.f16 s4, s3 -; CHECK-NEXT: vmovx.f16 s6, s0 -; CHECK-NEXT: vcvt.s32.f16 s8, s4 -; CHECK-NEXT: vmovx.f16 s4, s2 -; CHECK-NEXT: vcvt.s32.f16 s10, s4 -; CHECK-NEXT: vmovx.f16 s4, s1 -; CHECK-NEXT: vcvt.s32.f16 s14, s2 -; CHECK-NEXT: vcvt.s32.f16 s2, s1 -; CHECK-NEXT: vcvt.s32.f16 s0, s0 -; CHECK-NEXT: vcvt.s32.f16 s4, s4 -; CHECK-NEXT: vcvt.s32.f16 s6, s6 -; CHECK-NEXT: vmov r0, s2 -; CHECK-NEXT: vmov r1, s0 -; CHECK-NEXT: vcvt.s32.f16 s12, s3 -; CHECK-NEXT: vmov q0[2], q0[0], r1, r0 -; CHECK-NEXT: vmov r0, s4 -; CHECK-NEXT: vmov r1, s6 -; CHECK-NEXT: vmov q0[3], q0[1], r1, r0 -; CHECK-NEXT: vmov r0, s12 -; CHECK-NEXT: vmov r1, s14 -; CHECK-NEXT: vmov q1[2], q1[0], r1, r0 -; CHECK-NEXT: vmov r0, s8 -; CHECK-NEXT: vmov r1, s10 -; CHECK-NEXT: vmov q1[3], q1[1], r1, r0 -; CHECK-NEXT: bx lr - %x = call <8 x i32> @llvm.fptosi.sat.v8f16.v8i32(<8 x half> %f) - ret <8 x i32> %x -} - -define arm_aapcs_vfpcc <8 x i50> @test_signed_v8f16_v8i50(<8 x half> %f) { -; CHECK-LABEL: test_signed_v8f16_v8i50: -; CHECK: @ %bb.0: -; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, r11, lr} -; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} -; CHECK-NEXT: .pad #4 -; CHECK-NEXT: sub sp, #4 -; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: mov r9, r0 -; CHECK-NEXT: vcvtt.f32.f16 s30, s19 -; CHECK-NEXT: vmov r0, s30 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcvtb.f32.f16 s26, s18 -; CHECK-NEXT: mov r4, r0 -; CHECK-NEXT: vmov r0, s26 -; CHECK-NEXT: vldr s24, .LCPI48_1 -; CHECK-NEXT: vcvtb.f32.f16 s20, s16 -; CHECK-NEXT: vcvtb.f32.f16 s28, s19 -; CHECK-NEXT: vcmp.f32 s30, s24 -; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vmov r7, s20 -; CHECK-NEXT: vldr s22, .LCPI48_0 -; CHECK-NEXT: vmov r6, s28 -; CHECK-NEXT: itt lt -; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: movtlt r5, #65534 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s26, s24 -; CHECK-NEXT: mov r10, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s22 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s22 -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r5, #65535 -; CHECK-NEXT: movtgt r5, #1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s26, s26 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str.w r0, [r9, #25] -; CHECK-NEXT: mov r0, r7 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s20, s24 -; CHECK-NEXT: mov r8, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s22 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s20 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s24 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str.w r0, [r9] -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r4, #0 -; CHECK-NEXT: vcmp.f32 s30, s22 -; CHECK-NEXT: mov r0, r6 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r4, #-1 -; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt vs -; CHECK-NEXT: movvs r4, #0 -; CHECK-NEXT: movvs r5, #0 -; CHECK-NEXT: mov r7, r5 -; CHECK-NEXT: bfc r7, #18, #14 -; CHECK-NEXT: lsll r4, r7, #22 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s28, s24 -; CHECK-NEXT: mov r6, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s24 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: movtlt r1, #65534 -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r1, #65535 -; CHECK-NEXT: movtgt r1, #1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: mov r2, r6 -; CHECK-NEXT: bfc r1, #18, #14 -; CHECK-NEXT: vcvtt.f32.f16 s28, s18 -; CHECK-NEXT: lsrl r2, r1, #28 -; CHECK-NEXT: orr.w r0, r1, r7 -; CHECK-NEXT: str.w r0, [r9, #45] -; CHECK-NEXT: vmov r0, s28 -; CHECK-NEXT: orrs r4, r2 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s28, s24 -; CHECK-NEXT: mov r7, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: movtlt r1, #65534 -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: vcvtb.f32.f16 s18, s17 -; CHECK-NEXT: lsrs r0, r5, #10 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r1, #65535 -; CHECK-NEXT: movtgt r1, #1 -; CHECK-NEXT: str.w r4, [r9, #41] -; CHECK-NEXT: strb.w r0, [r9, #49] -; CHECK-NEXT: vmov r0, s18 -; CHECK-NEXT: vcmp.f32 s28, s24 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r7, #-1 -; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt vs -; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: bfc r1, #18, #14 -; CHECK-NEXT: mov r4, r7 -; CHECK-NEXT: lsrl r4, r1, #14 -; CHECK-NEXT: orr.w r6, r1, r6, lsl #4 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcvtt.f32.f16 s28, s17 -; CHECK-NEXT: mov r11, r0 -; CHECK-NEXT: vmov r0, s28 -; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: vcmp.f32 s18, s24 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt lt -; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: movtlt r5, #65534 -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r5, #65535 -; CHECK-NEXT: movtgt r5, #1 -; CHECK-NEXT: str.w r6, [r9, #37] -; CHECK-NEXT: str.w r4, [r9, #33] -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s28, s24 -; CHECK-NEXT: mov r4, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: itt lt -; CHECK-NEXT: movlt r4, #0 -; CHECK-NEXT: movtlt r4, #65534 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s24 -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r4, #65535 -; CHECK-NEXT: movtgt r4, #1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt lt -; CHECK-NEXT: movwlt r10, #0 -; CHECK-NEXT: movtlt r10, #65534 -; CHECK-NEXT: vcmp.f32 s26, s22 -; CHECK-NEXT: mov r6, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r10, #65535 -; CHECK-NEXT: movtgt r10, #1 -; CHECK-NEXT: vcmp.f32 s26, s26 -; CHECK-NEXT: vcvtt.f32.f16 s16, s16 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r10, #0 -; CHECK-NEXT: bfc r10, #18, #14 -; CHECK-NEXT: vcmp.f32 s28, s24 -; CHECK-NEXT: orr.w r0, r10, r7, lsl #18 -; CHECK-NEXT: str.w r0, [r9, #29] -; CHECK-NEXT: vmov r0, s16 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s24 -; CHECK-NEXT: itt vs -; CHECK-NEXT: movvs r6, #0 -; CHECK-NEXT: movvs r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r11, #0 -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: mov r1, r4 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r11, #-1 -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: bfc r1, #18, #14 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt vs -; CHECK-NEXT: movvs.w r11, #0 -; CHECK-NEXT: movvs r5, #0 -; CHECK-NEXT: vcmp.f32 s20, s24 -; CHECK-NEXT: bfc r5, #18, #14 -; CHECK-NEXT: mov r10, r11 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: lsll r6, r1, #22 -; CHECK-NEXT: lsrl r10, r5, #28 -; CHECK-NEXT: itt lt -; CHECK-NEXT: movwlt r8, #0 -; CHECK-NEXT: movtlt r8, #65534 -; CHECK-NEXT: vcmp.f32 s20, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r8, #65535 -; CHECK-NEXT: movtgt r8, #1 -; CHECK-NEXT: orrs r1, r5 -; CHECK-NEXT: str.w r1, [r9, #20] -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s16, s24 -; CHECK-NEXT: orr.w r2, r10, r6 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: movtlt r1, #65534 -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r1, #65535 -; CHECK-NEXT: movtgt r1, #1 -; CHECK-NEXT: str.w r2, [r9, #16] -; CHECK-NEXT: lsrs r2, r4, #10 -; CHECK-NEXT: vcmp.f32 s16, s24 -; CHECK-NEXT: strb.w r2, [r9, #24] -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s16, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: bfc r1, #18, #14 -; CHECK-NEXT: mov r2, r0 -; CHECK-NEXT: lsrl r2, r1, #14 -; CHECK-NEXT: vcmp.f32 s20, s20 -; CHECK-NEXT: orr.w r1, r1, r11, lsl #4 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: strd r2, r1, [r9, #8] -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r8, #0 -; CHECK-NEXT: bfc r8, #18, #14 -; CHECK-NEXT: orr.w r0, r8, r0, lsl #18 -; CHECK-NEXT: str.w r0, [r9, #4] -; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: add sp, #4 -; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} -; CHECK-NEXT: .p2align 2 -; CHECK-NEXT: @ %bb.1: -; CHECK-NEXT: .LCPI48_0: -; CHECK-NEXT: .long 0x57ffffff @ float 5.6294992E+14 -; CHECK-NEXT: .LCPI48_1: -; CHECK-NEXT: .long 0xd8000000 @ float -5.62949953E+14 - %x = call <8 x i50> @llvm.fptosi.sat.v8f16.v8i50(<8 x half> %f) - ret <8 x i50> %x -} - -define arm_aapcs_vfpcc <8 x i64> @test_signed_v8f16_v8i64(<8 x half> %f) { -; CHECK-LABEL: test_signed_v8f16_v8i64: -; CHECK: @ %bb.0: -; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, r11, lr} -; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} -; CHECK-NEXT: .pad #4 -; CHECK-NEXT: sub sp, #4 -; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: vcvtt.f32.f16 s20, s19 -; CHECK-NEXT: vmov r0, s20 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcvtb.f32.f16 s22, s19 -; CHECK-NEXT: mov r9, r0 -; CHECK-NEXT: vmov r0, s22 -; CHECK-NEXT: vldr s30, .LCPI49_1 -; CHECK-NEXT: vldr s28, .LCPI49_0 -; CHECK-NEXT: vcvtb.f32.f16 s24, s16 -; CHECK-NEXT: vcmp.f32 s20, s30 -; CHECK-NEXT: vcvtt.f32.f16 s16, s16 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r9, #0 -; CHECK-NEXT: vcmp.f32 s20, s28 -; CHECK-NEXT: mov r8, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r9, #-1 -; CHECK-NEXT: vcmp.f32 s20, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vmov r4, s24 -; CHECK-NEXT: vmov r5, s16 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r9, #0 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s22, s30 -; CHECK-NEXT: mov r11, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r11, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, s22 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r11, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s30 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r11, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r8, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s20 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r8, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r10, r1 -; CHECK-NEXT: vcmp.f32 s22, s30 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r8, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r10, #-2147483648 -; CHECK-NEXT: vcmp.f32 s22, s28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r0, r5 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r10, #-2147483648 -; CHECK-NEXT: vcmp.f32 s22, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r10, #0 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: mov r6, r0 -; CHECK-NEXT: vcmp.f32 s16, s30 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vcmp.f32 s16, s28 -; CHECK-NEXT: mov r0, r4 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r6, #0 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcvtt.f32.f16 s19, s17 -; CHECK-NEXT: mov r7, r1 -; CHECK-NEXT: vmov r1, s19 -; CHECK-NEXT: vcmp.f32 s24, s30 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s24, s28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s24, s24 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: vmov q5[2], q5[0], r0, r6 -; CHECK-NEXT: mov r0, r1 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcvtb.f32.f16 s17, s17 -; CHECK-NEXT: mov r6, r0 -; CHECK-NEXT: vmov r0, s17 -; CHECK-NEXT: mov r4, r1 -; CHECK-NEXT: vcmp.f32 s19, s30 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 +; CHECK-MVE-NEXT: vcvtb.f32.f16 s3, s3 +; CHECK-MVE-NEXT: vldr s4, .LCPI45_0 +; CHECK-MVE-NEXT: vmaxnm.f32 s16, s3, s6 +; CHECK-MVE-NEXT: vcvtt.f32.f16 s7, s2 +; CHECK-MVE-NEXT: vmaxnm.f32 s15, s13, s6 +; CHECK-MVE-NEXT: vminnm.f32 s16, s16, s4 +; CHECK-MVE-NEXT: vcvtb.f32.f16 s2, s2 +; CHECK-MVE-NEXT: vminnm.f32 s15, s15, s4 +; CHECK-MVE-NEXT: vmaxnm.f32 s11, s2, s6 +; CHECK-MVE-NEXT: vcvt.s32.f32 s16, s16 +; CHECK-MVE-NEXT: vcvtt.f32.f16 s12, s1 +; CHECK-MVE-NEXT: vmaxnm.f32 s9, s7, s6 +; CHECK-MVE-NEXT: vminnm.f32 s11, s11, s4 +; CHECK-MVE-NEXT: vcvtb.f32.f16 s1, s1 +; CHECK-MVE-NEXT: vcvt.s32.f32 s15, s15 +; CHECK-MVE-NEXT: vcvtb.f32.f16 s8, s0 +; CHECK-MVE-NEXT: vmaxnm.f32 s5, s1, s6 +; CHECK-MVE-NEXT: vminnm.f32 s9, s9, s4 +; CHECK-MVE-NEXT: vcvt.s32.f32 s11, s11 +; CHECK-MVE-NEXT: vmaxnm.f32 s10, s8, s6 +; CHECK-MVE-NEXT: vmaxnm.f32 s14, s12, s6 +; CHECK-MVE-NEXT: vminnm.f32 s5, s5, s4 +; CHECK-MVE-NEXT: vcvt.s32.f32 s9, s9 +; CHECK-MVE-NEXT: vcvtt.f32.f16 s0, s0 +; CHECK-MVE-NEXT: vminnm.f32 s10, s10, s4 +; CHECK-MVE-NEXT: vminnm.f32 s14, s14, s4 +; CHECK-MVE-NEXT: vcvt.s32.f32 s5, s5 +; CHECK-MVE-NEXT: vmaxnm.f32 s6, s0, s6 +; CHECK-MVE-NEXT: vminnm.f32 s4, s6, s4 +; CHECK-MVE-NEXT: vcvt.s32.f32 s10, s10 +; CHECK-MVE-NEXT: vcvt.s32.f32 s14, s14 +; CHECK-MVE-NEXT: vcvt.s32.f32 s4, s4 +; CHECK-MVE-NEXT: vcmp.f32 s3, s3 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vmov r12, s16 +; CHECK-MVE-NEXT: vcmp.f32 s13, s13 +; CHECK-MVE-NEXT: it vs +; CHECK-MVE-NEXT: movvs.w r12, #0 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vmov lr, s15 +; CHECK-MVE-NEXT: vcmp.f32 s2, s2 +; CHECK-MVE-NEXT: it vs +; CHECK-MVE-NEXT: movvs.w lr, #0 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vmov r2, s11 +; CHECK-MVE-NEXT: vcmp.f32 s7, s7 +; CHECK-MVE-NEXT: it vs +; CHECK-MVE-NEXT: movvs r2, #0 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vmov r3, s9 +; CHECK-MVE-NEXT: vcmp.f32 s1, s1 +; CHECK-MVE-NEXT: it vs +; CHECK-MVE-NEXT: movvs r3, #0 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vmov r0, s5 +; CHECK-MVE-NEXT: vcmp.f32 s12, s12 +; CHECK-MVE-NEXT: it vs +; CHECK-MVE-NEXT: movvs r0, #0 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vmov r1, s14 +; CHECK-MVE-NEXT: vmov r4, s10 +; CHECK-MVE-NEXT: it vs +; CHECK-MVE-NEXT: movvs r1, #0 +; CHECK-MVE-NEXT: vcmp.f32 s8, s8 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: it vs +; CHECK-MVE-NEXT: movvs r4, #0 +; CHECK-MVE-NEXT: vcmp.f32 s0, s0 +; CHECK-MVE-NEXT: vmov.16 q0[0], r4 +; CHECK-MVE-NEXT: vmov r5, s4 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: it vs +; CHECK-MVE-NEXT: movvs r5, #0 +; CHECK-MVE-NEXT: vmov.16 q0[1], r5 +; CHECK-MVE-NEXT: vmov.16 q0[2], r0 +; CHECK-MVE-NEXT: vmov.16 q0[3], r1 +; CHECK-MVE-NEXT: vmov.16 q0[4], r2 +; CHECK-MVE-NEXT: vmov.16 q0[5], r3 +; CHECK-MVE-NEXT: vmov.16 q0[6], r12 +; CHECK-MVE-NEXT: vmov.16 q0[7], lr +; CHECK-MVE-NEXT: vpop {d8} +; CHECK-MVE-NEXT: pop {r4, r5, r7, pc} +; CHECK-MVE-NEXT: .p2align 2 +; CHECK-MVE-NEXT: @ %bb.1: +; CHECK-MVE-NEXT: .LCPI45_0: +; CHECK-MVE-NEXT: .long 0x46fffe00 @ float 32767 +; CHECK-MVE-NEXT: .LCPI45_1: +; CHECK-MVE-NEXT: .long 0xc7000000 @ float -32768 +; +; CHECK-MVEFP-LABEL: test_signed_v8f16_v8i16: +; CHECK-MVEFP: @ %bb.0: +; CHECK-MVEFP-NEXT: vcvt.s16.f16 q0, q0 +; CHECK-MVEFP-NEXT: bx lr + %x = call <8 x i16> @llvm.fptosi.sat.v8f16.v8i16(<8 x half> %f) + ret <8 x i16> %x +} + +define arm_aapcs_vfpcc <8 x i19> @test_signed_v8f16_v8i19(<8 x half> %f) { +; CHECK-LABEL: test_signed_v8f16_v8i19: +; CHECK: @ %bb.0: +; CHECK-NEXT: .save {r4, r5, r7, r9, r11, lr} +; CHECK-NEXT: push.w {r4, r5, r7, r9, r11, lr} +; CHECK-NEXT: vldr s6, .LCPI46_1 +; CHECK-NEXT: vcvtb.f32.f16 s12, s0 +; CHECK-NEXT: vcvtt.f32.f16 s0, s0 +; CHECK-NEXT: vldr s4, .LCPI46_0 +; CHECK-NEXT: vmaxnm.f32 s5, s0, s6 +; CHECK-NEXT: vmaxnm.f32 s14, s12, s6 +; CHECK-NEXT: vminnm.f32 s5, s5, s4 +; CHECK-NEXT: vcvtt.f32.f16 s8, s1 +; CHECK-NEXT: vminnm.f32 s14, s14, s4 +; CHECK-NEXT: vcvt.s32.f32 s5, s5 +; CHECK-NEXT: vmaxnm.f32 s10, s8, s6 +; CHECK-NEXT: vcvt.s32.f32 s14, s14 +; CHECK-NEXT: vminnm.f32 s10, s10, s4 +; CHECK-NEXT: vcvtb.f32.f16 s1, s1 +; CHECK-NEXT: vcvt.s32.f32 s10, s10 +; CHECK-NEXT: vmaxnm.f32 s7, s1, s6 +; CHECK-NEXT: vminnm.f32 s7, s7, s4 +; CHECK-NEXT: vcmp.f32 s0, s0 +; CHECK-NEXT: vcvt.s32.f32 s7, s7 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s30 +; CHECK-NEXT: vcmp.f32 s12, s12 +; CHECK-NEXT: mov.w r7, #0 +; CHECK-NEXT: vmov r2, s5 +; CHECK-NEXT: vcvtb.f32.f16 s0, s2 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r5, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r5, #-2147483648 +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: vmov r1, s14 +; CHECK-NEXT: bfc r2, #19, #13 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s30 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r5, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r7, #-2147483648 -; CHECK-NEXT: vcmp.f32 s24, s28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r7, #-2147483648 -; CHECK-NEXT: vcmp.f32 s24, s24 +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: vcmp.f32 s8, s8 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: lsll r2, r7, #19 +; CHECK-NEXT: bfc r1, #19, #13 +; CHECK-NEXT: vmov r12, s10 +; CHECK-NEXT: vcmp.f32 s1, s1 +; CHECK-NEXT: vmaxnm.f32 s8, s0, s6 +; CHECK-NEXT: orr.w r1, r1, r2 +; CHECK-NEXT: str r1, [r0] ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: vmov q5[3], q5[1], r7, r5 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcvtt.f32.f16 s16, s18 -; CHECK-NEXT: mov r7, r1 -; CHECK-NEXT: vmov r1, s16 -; CHECK-NEXT: vcmp.f32 s17, s30 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s17, s28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s17, s17 +; CHECK-NEXT: movvs.w r12, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: vcmp.f32 s0, s0 +; CHECK-NEXT: vcvtt.f32.f16 s0, s2 +; CHECK-NEXT: vmaxnm.f32 s2, s0, s6 +; CHECK-NEXT: vminnm.f32 s8, s8, s4 +; CHECK-NEXT: vminnm.f32 s2, s2, s4 +; CHECK-NEXT: vmov r3, s7 +; CHECK-NEXT: vcvt.s32.f32 s2, s2 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: vmov q6[2], q6[0], r0, r6 -; CHECK-NEXT: mov r0, r1 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcvtb.f32.f16 s18, s18 -; CHECK-NEXT: mov r6, r0 -; CHECK-NEXT: vmov r0, s18 -; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: vcmp.f32 s16, s30 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: vcvt.s32.f32 s8, s8 +; CHECK-NEXT: bfc r3, #19, #13 +; CHECK-NEXT: mov r2, r12 +; CHECK-NEXT: movs r1, #0 +; CHECK-NEXT: bfc r2, #19, #13 +; CHECK-NEXT: mov r4, r3 +; CHECK-NEXT: mov.w r9, #0 +; CHECK-NEXT: lsrl r2, r1, #7 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s30 +; CHECK-NEXT: vcmp.f32 s0, s0 +; CHECK-NEXT: lsrl r4, r9, #26 +; CHECK-NEXT: vcvtt.f32.f16 s0, s3 +; CHECK-NEXT: mov lr, r1 +; CHECK-NEXT: orr.w r1, r4, r2 +; CHECK-NEXT: vmov r4, s2 +; CHECK-NEXT: vmaxnm.f32 s2, s0, s6 +; CHECK-NEXT: vmov r2, s8 +; CHECK-NEXT: vminnm.f32 s2, s2, s4 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r4, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r4, #-2147483648 +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: vcvt.s32.f32 s2, s2 +; CHECK-NEXT: bfc r2, #19, #13 +; CHECK-NEXT: movs r5, #0 +; CHECK-NEXT: lsll r2, r5, #12 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s30 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r4, #0 +; CHECK-NEXT: orrs r2, r1 +; CHECK-NEXT: bfc r4, #19, #13 +; CHECK-NEXT: movs r1, #0 +; CHECK-NEXT: lsll r4, r1, #31 +; CHECK-NEXT: vcmp.f32 s0, s0 +; CHECK-NEXT: orrs r2, r4 +; CHECK-NEXT: str r2, [r0, #8] +; CHECK-NEXT: orr.w r2, r7, r3, lsl #6 +; CHECK-NEXT: vcvtb.f32.f16 s0, s3 +; CHECK-NEXT: orr.w r3, r2, r12, lsl #25 +; CHECK-NEXT: vmov r2, s2 +; CHECK-NEXT: vmaxnm.f32 s2, s0, s6 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r7, #-2147483648 -; CHECK-NEXT: vcmp.f32 s17, s28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r7, #-2147483648 -; CHECK-NEXT: vcmp.f32 s17, s17 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: vmov q6[3], q6[1], r7, r4 -; CHECK-NEXT: bl __aeabi_f2lz -; CHECK-NEXT: vcmp.f32 s18, s30 -; CHECK-NEXT: vmov q3[2], q3[0], r11, r9 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s30 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r5, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r5, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s30 +; CHECK-NEXT: vminnm.f32 s2, s2, s4 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r5, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r1, #-2147483648 -; CHECK-NEXT: vcmp.f32 s18, s28 -; CHECK-NEXT: vmov q2[2], q2[0], r0, r6 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r1, #-2147483648 -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: vmov q3[3], q3[1], r10, r8 +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: vcvt.s32.f32 s2, s2 +; CHECK-NEXT: bfc r2, #19, #13 +; CHECK-NEXT: movs r7, #0 +; CHECK-NEXT: vcmp.f32 s0, s0 +; CHECK-NEXT: lsll r2, r7, #5 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: mov.w r11, #0 +; CHECK-NEXT: vmov r7, s2 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmov q2[3], q2[1], r1, r5 -; CHECK-NEXT: vmov q0, q5 -; CHECK-NEXT: vmov q1, q6 -; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: add sp, #4 -; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} +; CHECK-NEXT: movvs r7, #0 +; CHECK-NEXT: mov r4, r7 +; CHECK-NEXT: bfc r4, #19, #13 +; CHECK-NEXT: lsrl r4, r11, #14 +; CHECK-NEXT: orrs r2, r4 +; CHECK-NEXT: strh r2, [r0, #16] +; CHECK-NEXT: str r3, [r0, #4] +; CHECK-NEXT: lsrs r2, r2, #16 +; CHECK-NEXT: strb r2, [r0, #18] +; CHECK-NEXT: orr.w r2, r9, lr +; CHECK-NEXT: orrs r2, r5 +; CHECK-NEXT: orrs r1, r2 +; CHECK-NEXT: orr.w r1, r1, r7, lsl #18 +; CHECK-NEXT: str r1, [r0, #12] +; CHECK-NEXT: pop.w {r4, r5, r7, r9, r11, pc} ; CHECK-NEXT: .p2align 2 ; CHECK-NEXT: @ %bb.1: -; CHECK-NEXT: .LCPI49_0: -; CHECK-NEXT: .long 0x5effffff @ float 9.22337149E+18 -; CHECK-NEXT: .LCPI49_1: -; CHECK-NEXT: .long 0xdf000000 @ float -9.22337203E+18 - %x = call <8 x i64> @llvm.fptosi.sat.v8f16.v8i64(<8 x half> %f) - ret <8 x i64> %x +; CHECK-NEXT: .LCPI46_0: +; CHECK-NEXT: .long 0x487fffc0 @ float 262143 +; CHECK-NEXT: .LCPI46_1: +; CHECK-NEXT: .long 0xc8800000 @ float -262144 + %x = call <8 x i19> @llvm.fptosi.sat.v8f16.v8i19(<8 x half> %f) + ret <8 x i19> %x } -define arm_aapcs_vfpcc <8 x i100> @test_signed_v8f16_v8i100(<8 x half> %f) { -; CHECK-LABEL: test_signed_v8f16_v8i100: +define arm_aapcs_vfpcc <8 x i32> @test_signed_v8f16_v8i32_duplicate(<8 x half> %f) { +; CHECK-LABEL: test_signed_v8f16_v8i32_duplicate: +; CHECK: @ %bb.0: +; CHECK-NEXT: vmovx.f16 s4, s3 +; CHECK-NEXT: vmovx.f16 s6, s0 +; CHECK-NEXT: vcvt.s32.f16 s8, s4 +; CHECK-NEXT: vmovx.f16 s4, s2 +; CHECK-NEXT: vcvt.s32.f16 s10, s4 +; CHECK-NEXT: vmovx.f16 s4, s1 +; CHECK-NEXT: vcvt.s32.f16 s14, s2 +; CHECK-NEXT: vcvt.s32.f16 s2, s1 +; CHECK-NEXT: vcvt.s32.f16 s0, s0 +; CHECK-NEXT: vcvt.s32.f16 s4, s4 +; CHECK-NEXT: vcvt.s32.f16 s6, s6 +; CHECK-NEXT: vmov r0, s2 +; CHECK-NEXT: vmov r1, s0 +; CHECK-NEXT: vcvt.s32.f16 s12, s3 +; CHECK-NEXT: vmov q0[2], q0[0], r1, r0 +; CHECK-NEXT: vmov r0, s4 +; CHECK-NEXT: vmov r1, s6 +; CHECK-NEXT: vmov q0[3], q0[1], r1, r0 +; CHECK-NEXT: vmov r0, s12 +; CHECK-NEXT: vmov r1, s14 +; CHECK-NEXT: vmov q1[2], q1[0], r1, r0 +; CHECK-NEXT: vmov r0, s8 +; CHECK-NEXT: vmov r1, s10 +; CHECK-NEXT: vmov q1[3], q1[1], r1, r0 +; CHECK-NEXT: bx lr + %x = call <8 x i32> @llvm.fptosi.sat.v8f16.v8i32(<8 x half> %f) + ret <8 x i32> %x +} + +define arm_aapcs_vfpcc <8 x i50> @test_signed_v8f16_v8i50(<8 x half> %f) { +; CHECK-LABEL: test_signed_v8f16_v8i50: ; CHECK: @ %bb.0: ; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, r11, lr} ; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} @@ -5417,996 +4561,1039 @@ define arm_aapcs_vfpcc <8 x i100> @test_signed_v8f16_v8i100(<8 x half> %f) { ; CHECK-NEXT: sub sp, #4 ; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15} ; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: .pad #32 -; CHECK-NEXT: sub sp, #32 +; CHECK-NEXT: .pad #24 +; CHECK-NEXT: sub sp, #24 ; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: mov r11, r0 -; CHECK-NEXT: vcvtb.f32.f16 s21, s19 -; CHECK-NEXT: vcvtt.f32.f16 s24, s19 -; CHECK-NEXT: vmov r0, s21 -; CHECK-NEXT: vcvtb.f32.f16 s26, s16 +; CHECK-NEXT: mov r7, r0 +; CHECK-NEXT: vcvtt.f32.f16 s24, s17 +; CHECK-NEXT: vmov r0, s24 +; CHECK-NEXT: bl __aeabi_f2lz ; CHECK-NEXT: vcvtb.f32.f16 s28, s17 -; CHECK-NEXT: vcvtb.f32.f16 s30, s18 -; CHECK-NEXT: vldr s20, .LCPI50_2 -; CHECK-NEXT: vmov r8, s24 -; CHECK-NEXT: vmov r9, s26 -; CHECK-NEXT: vcvtt.f32.f16 s22, s18 -; CHECK-NEXT: vmov r6, s28 -; CHECK-NEXT: vmov r7, s30 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vldr s18, .LCPI50_3 -; CHECK-NEXT: mov r5, r3 -; CHECK-NEXT: vcmp.f32 s21, s18 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s21, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s21, s21 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s21, s18 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: mov r8, r0 +; CHECK-NEXT: vmov r0, s28 +; CHECK-NEXT: mov r6, r1 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcvtt.f32.f16 s30, s16 +; CHECK-NEXT: mov r4, r0 +; CHECK-NEXT: vmov r0, s30 +; CHECK-NEXT: mov r5, r1 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcvtb.f32.f16 s26, s18 +; CHECK-NEXT: mov r2, r0 +; CHECK-NEXT: vmov r0, s26 +; CHECK-NEXT: vldr s20, .LCPI48_0 +; CHECK-NEXT: vldr s22, .LCPI48_1 +; CHECK-NEXT: vcmp.f32 s30, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s21, s20 -; CHECK-NEXT: str.w r2, [r11, #83] -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movtlt r1, #65534 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s21, s21 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s21, s18 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r11, #79] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s21, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s21, s21 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str.w r0, [r11, #75] -; CHECK-NEXT: mov r0, r7 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s30, s18 -; CHECK-NEXT: mov r7, r3 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s20 +; CHECK-NEXT: vcmp.f32 s30, s22 ; CHECK-NEXT: it lt ; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r1, #65535 +; CHECK-NEXT: movtgt r1, #1 ; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s18 +; CHECK-NEXT: vcmp.f32 s28, s20 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: str.w r2, [r11, #58] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s18 +; CHECK-NEXT: str r2, [sp, #20] @ 4-byte Spill ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r1, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r11, #54] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str.w r0, [r11, #50] -; CHECK-NEXT: mov r0, r6 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s28, s18 -; CHECK-NEXT: str r3, [sp, #24] @ 4-byte Spill -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: vcmp.f32 s28, s22 +; CHECK-NEXT: str r1, [sp, #16] @ 4-byte Spill +; CHECK-NEXT: ittt lt +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movtlt r5, #65534 +; CHECK-NEXT: movlt r4, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r5, #65535 +; CHECK-NEXT: movtgt r5, #1 +; CHECK-NEXT: movgt.w r4, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s18 +; CHECK-NEXT: vcmp.f32 s24, s20 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: str.w r2, [r11, #33] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s18 +; CHECK-NEXT: movvs r4, #0 +; CHECK-NEXT: str r4, [sp, #12] @ 4-byte Spill ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: movvs r5, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r11, #29] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s28, s20 +; CHECK-NEXT: str r5, [sp, #8] @ 4-byte Spill +; CHECK-NEXT: ittt lt +; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: movtlt r6, #65534 +; CHECK-NEXT: movlt.w r8, #0 +; CHECK-NEXT: vcmp.f32 s24, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s28, s28 +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r6, #65535 +; CHECK-NEXT: movtgt r6, #1 +; CHECK-NEXT: movgt.w r8, #-1 +; CHECK-NEXT: vcmp.f32 s24, s24 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str.w r0, [r11, #25] -; CHECK-NEXT: mov r0, r9 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s26, s18 -; CHECK-NEXT: str r3, [sp, #12] @ 4-byte Spill -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s26 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s18 +; CHECK-NEXT: movvs.w r8, #0 +; CHECK-NEXT: str.w r8, [sp, #4] @ 4-byte Spill ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movvs r6, #0 +; CHECK-NEXT: str r6, [sp] @ 4-byte Spill +; CHECK-NEXT: bl __aeabi_f2lz ; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: str.w r2, [r11, #8] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s26 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s18 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: mov r6, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r11, #4] -; CHECK-NEXT: it lt +; CHECK-NEXT: ittt lt +; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: movtlt r6, #65534 ; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s26, s20 +; CHECK-NEXT: vcmp.f32 s26, s22 +; CHECK-NEXT: vcvtt.f32.f16 s18, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r6, #65535 +; CHECK-NEXT: movtgt r6, #1 ; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: vcmp.f32 s26, s26 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str.w r0, [r11] -; CHECK-NEXT: mov r0, r8 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s24, s18 -; CHECK-NEXT: mov r6, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: str r3, [sp, #8] @ 4-byte Spill -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s24 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s21, s18 +; CHECK-NEXT: str.w r0, [r7, #25] +; CHECK-NEXT: vmov r0, s18 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r6, #0 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcmp.f32 s18, s20 +; CHECK-NEXT: mov r8, r0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r5, #7 -; CHECK-NEXT: vcmp.f32 s21, s20 ; CHECK-NEXT: mov r9, r1 +; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: ittt lt +; CHECK-NEXT: movlt.w r8, #0 +; CHECK-NEXT: movwlt r9, #0 +; CHECK-NEXT: movtlt r9, #65534 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r5, #7 -; CHECK-NEXT: vcmp.f32 s21, s21 -; CHECK-NEXT: mov r8, r2 +; CHECK-NEXT: vcmp.f32 s18, s18 +; CHECK-NEXT: vcvtb.f32.f16 s18, s19 +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r9, #65535 +; CHECK-NEXT: movtgt r9, #1 +; CHECK-NEXT: movgt.w r8, #-1 +; CHECK-NEXT: vmov r0, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r5, #0 -; CHECK-NEXT: and r0, r5, #15 -; CHECK-NEXT: orr.w r1, r0, r6, lsl #4 -; CHECK-NEXT: vmov r0, s22 -; CHECK-NEXT: str.w r1, [r11, #87] -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s22, s18 +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs.w r8, #0 +; CHECK-NEXT: movvs.w r9, #0 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: mov r10, r0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: str r2, [sp, #4] @ 4-byte Spill -; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: str r3, [sp, #20] @ 4-byte Spill -; CHECK-NEXT: it lt +; CHECK-NEXT: mov r11, r1 +; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: ittt lt ; CHECK-NEXT: movlt.w r10, #0 +; CHECK-NEXT: movwlt r11, #0 +; CHECK-NEXT: movtlt r11, #65534 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, s22 -; CHECK-NEXT: it gt +; CHECK-NEXT: vcmp.f32 s18, s18 +; CHECK-NEXT: vcvtt.f32.f16 s18, s19 +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r11, #65535 +; CHECK-NEXT: movtgt r11, #1 ; CHECK-NEXT: movgt.w r10, #-1 +; CHECK-NEXT: vmov r0, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s18 -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs ; CHECK-NEXT: movvs.w r10, #0 +; CHECK-NEXT: movvs.w r11, #0 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcvtb.f32.f16 s16, s16 +; CHECK-NEXT: mov r4, r0 +; CHECK-NEXT: vmov r0, s16 +; CHECK-NEXT: mov r5, r1 +; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r7, #7 -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r7, #7 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: and r0, r7, #15 -; CHECK-NEXT: orr.w r0, r0, r10, lsl #4 -; CHECK-NEXT: vcvtt.f32.f16 s30, s17 -; CHECK-NEXT: str.w r0, [r11, #62] -; CHECK-NEXT: vmov r0, s30 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s30, s18 -; CHECK-NEXT: str r2, [sp, #16] @ 4-byte Spill -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r3, [sp, #28] @ 4-byte Spill -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: mov r4, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: vcmp.f32 s28, s18 -; CHECK-NEXT: mov r1, r0 -; CHECK-NEXT: str r0, [sp] @ 4-byte Spill -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: ldr r0, [sp, #24] @ 4-byte Reload -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r0, #7 +; CHECK-NEXT: ittt lt +; CHECK-NEXT: movlt r4, #0 +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movtlt r5, #65534 +; CHECK-NEXT: vcmp.f32 s18, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r0, #7 -; CHECK-NEXT: vcmp.f32 s28, s28 +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r5, #65535 +; CHECK-NEXT: movtgt r5, #1 +; CHECK-NEXT: movgt.w r4, #-1 +; CHECK-NEXT: vcmp.f32 s18, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: and r0, r0, #15 -; CHECK-NEXT: vcvtt.f32.f16 s16, s16 -; CHECK-NEXT: orr.w r0, r0, r1, lsl #4 -; CHECK-NEXT: str.w r0, [r11, #37] -; CHECK-NEXT: vmov r0, s16 -; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s16, s18 +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r4, #0 +; CHECK-NEXT: movvs r5, #0 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcmp.f32 s16, s20 +; CHECK-NEXT: bfc r11, #18, #14 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: ittt lt +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movtlt r1, #65534 ; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s16, s20 +; CHECK-NEXT: vcmp.f32 s16, s22 +; CHECK-NEXT: mov r2, r10 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r1, #65535 +; CHECK-NEXT: movtgt r1, #1 ; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: vcmp.f32 s16, s16 +; CHECK-NEXT: lsrl r2, r11, #28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: vcmp.f32 s26, s18 -; CHECK-NEXT: ldr r7, [sp, #12] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r7, #7 -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r7, #7 -; CHECK-NEXT: vcmp.f32 s26, s26 +; CHECK-NEXT: str r0, [r7] +; CHECK-NEXT: lsrs r0, r5, #10 +; CHECK-NEXT: bfc r5, #18, #14 +; CHECK-NEXT: bfc r9, #18, #14 +; CHECK-NEXT: lsll r4, r5, #22 +; CHECK-NEXT: bfc r6, #18, #14 +; CHECK-NEXT: orr.w r3, r11, r5 +; CHECK-NEXT: str.w r3, [r7, #45] +; CHECK-NEXT: orrs r2, r4 +; CHECK-NEXT: str.w r2, [r7, #41] +; CHECK-NEXT: strb.w r0, [r7, #49] +; CHECK-NEXT: mov r0, r8 +; CHECK-NEXT: lsrl r0, r9, #14 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: orr.w r2, r9, r10, lsl #4 +; CHECK-NEXT: str.w r2, [r7, #37] +; CHECK-NEXT: str.w r0, [r7, #33] +; CHECK-NEXT: orr.w r0, r6, r8, lsl #18 +; CHECK-NEXT: str.w r0, [r7, #29] +; CHECK-NEXT: ldr r5, [sp, #8] @ 4-byte Reload +; CHECK-NEXT: ldr r3, [sp] @ 4-byte Reload +; CHECK-NEXT: ldr r4, [sp, #12] @ 4-byte Reload +; CHECK-NEXT: ldr r6, [sp, #4] @ 4-byte Reload +; CHECK-NEXT: bfc r5, #18, #14 +; CHECK-NEXT: lsr.w r0, r3, #10 +; CHECK-NEXT: bfc r3, #18, #14 +; CHECK-NEXT: mov r2, r4 +; CHECK-NEXT: lsll r6, r3, #22 +; CHECK-NEXT: lsrl r2, r5, #28 +; CHECK-NEXT: orr.w r3, r3, r5 +; CHECK-NEXT: str r3, [r7, #20] +; CHECK-NEXT: orr.w r2, r2, r6 +; CHECK-NEXT: str r2, [r7, #16] +; CHECK-NEXT: strb r0, [r7, #24] +; CHECK-NEXT: ldr r3, [sp, #16] @ 4-byte Reload +; CHECK-NEXT: ldr r6, [sp, #20] @ 4-byte Reload +; CHECK-NEXT: bfc r3, #18, #14 +; CHECK-NEXT: mov r0, r6 +; CHECK-NEXT: lsrl r0, r3, #14 +; CHECK-NEXT: orr.w r2, r3, r4, lsl #4 +; CHECK-NEXT: strd r0, r2, [r7, #8] ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: vcmp.f32 s24, s18 -; CHECK-NEXT: and r7, r7, #15 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: orr.w r7, r7, r0, lsl #4 -; CHECK-NEXT: str.w r7, [r11, #12] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r9, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: b.w .LBB50_3 +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: bfc r1, #18, #14 +; CHECK-NEXT: orr.w r0, r1, r6, lsl #18 +; CHECK-NEXT: str r0, [r7, #4] +; CHECK-NEXT: add sp, #24 +; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: add sp, #4 +; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} ; CHECK-NEXT: .p2align 2 ; CHECK-NEXT: @ %bb.1: -; CHECK-NEXT: .LCPI50_2: -; CHECK-NEXT: .long 0x70ffffff @ float 6.33825262E+29 -; CHECK-NEXT: .p2align 2 -; CHECK-NEXT: @ %bb.2: -; CHECK-NEXT: .LCPI50_3: -; CHECK-NEXT: .long 0xf1000000 @ float -6.338253E+29 -; CHECK-NEXT: .p2align 1 -; CHECK-NEXT: .LBB50_3: -; CHECK-NEXT: vcmp.f32 s24, s24 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r9, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s18 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs.w r9, #0 +; CHECK-NEXT: .LCPI48_0: +; CHECK-NEXT: .long 0xd8000000 @ float -5.62949953E+14 +; CHECK-NEXT: .LCPI48_1: +; CHECK-NEXT: .long 0x57ffffff @ float 5.6294992E+14 + %x = call <8 x i50> @llvm.fptosi.sat.v8f16.v8i50(<8 x half> %f) + ret <8 x i50> %x +} + +define arm_aapcs_vfpcc <8 x i64> @test_signed_v8f16_v8i64(<8 x half> %f) { +; CHECK-LABEL: test_signed_v8f16_v8i64: +; CHECK: @ %bb.0: +; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, r11, lr} +; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} +; CHECK-NEXT: .pad #4 +; CHECK-NEXT: sub sp, #4 +; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: vmov q4, q0 +; CHECK-NEXT: vcvtt.f32.f16 s24, s19 +; CHECK-NEXT: vmov r0, s24 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcvtb.f32.f16 s26, s19 +; CHECK-NEXT: mov r9, r0 +; CHECK-NEXT: vmov r0, s26 +; CHECK-NEXT: vldr s30, .LCPI49_0 +; CHECK-NEXT: vldr s28, .LCPI49_1 +; CHECK-NEXT: mov r8, r1 +; CHECK-NEXT: vcmp.f32 s24, s30 +; CHECK-NEXT: vcvtt.f32.f16 s22, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r8, #0 -; CHECK-NEXT: vcmp.f32 s24, s20 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r8, #-2147483648 +; CHECK-NEXT: movlt.w r9, #0 +; CHECK-NEXT: vcmp.f32 s24, s28 +; CHECK-NEXT: vcvtt.f32.f16 s20, s16 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: lsrl r6, r9, #28 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r8, #-1 +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r9, #-1 +; CHECK-NEXT: mvngt r8, #-2147483648 ; CHECK-NEXT: vcmp.f32 s24, s24 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: vmov r4, s22 +; CHECK-NEXT: vmov r6, s20 +; CHECK-NEXT: itt vs ; CHECK-NEXT: movvs.w r8, #0 -; CHECK-NEXT: orr.w r7, r9, r8, lsl #4 -; CHECK-NEXT: str.w r7, [r11, #95] -; CHECK-NEXT: str.w r6, [r11, #91] -; CHECK-NEXT: vcmp.f32 s24, s18 -; CHECK-NEXT: ldr r7, [sp, #8] @ 4-byte Reload +; CHECK-NEXT: movvs.w r9, #0 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: mov r10, r0 +; CHECK-NEXT: vcmp.f32 s26, s30 +; CHECK-NEXT: mov r11, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r7, #7 -; CHECK-NEXT: vcmp.f32 s24, s20 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r10, #0 +; CHECK-NEXT: movlt.w r11, #-2147483648 +; CHECK-NEXT: vcmp.f32 s26, s28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r7, #7 -; CHECK-NEXT: vcmp.f32 s24, s24 +; CHECK-NEXT: mov r0, r4 +; CHECK-NEXT: itt gt +; CHECK-NEXT: mvngt r11, #-2147483648 +; CHECK-NEXT: movgt.w r10, #-1 +; CHECK-NEXT: vcmp.f32 s26, s26 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: and r7, r7, #15 -; CHECK-NEXT: vcmp.f32 s22, s18 -; CHECK-NEXT: lsrl r8, r7, #28 +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs.w r10, #0 +; CHECK-NEXT: movvs.w r11, #0 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: mov r5, r0 +; CHECK-NEXT: vcmp.f32 s22, s30 +; CHECK-NEXT: mov r4, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: strb.w r8, [r11, #99] -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r4, #-2147483648 ; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: vcmp.f32 s22, s20 +; CHECK-NEXT: vcmp.f32 s22, s28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: mov r0, r6 +; CHECK-NEXT: itt gt ; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: mvngt r4, #-2147483648 ; CHECK-NEXT: vcmp.f32 s22, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r4, #0 ; CHECK-NEXT: movvs r5, #0 -; CHECK-NEXT: vcmp.f32 s22, s18 -; CHECK-NEXT: ldr r6, [sp, #4] @ 4-byte Reload +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcvtb.f32.f16 s16, s16 +; CHECK-NEXT: mov r7, r0 +; CHECK-NEXT: vmov r0, s16 +; CHECK-NEXT: mov r6, r1 +; CHECK-NEXT: vcmp.f32 s20, s30 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: lsrl r10, r5, #28 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r6, #-2147483648 +; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: vcmp.f32 s20, s28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: vcmp.f32 s22, s22 +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r7, #-1 +; CHECK-NEXT: mvngt r6, #-2147483648 +; CHECK-NEXT: vcmp.f32 s20, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs ; CHECK-NEXT: movvs r6, #0 -; CHECK-NEXT: orr.w r7, r5, r6, lsl #4 -; CHECK-NEXT: str.w r7, [r11, #70] -; CHECK-NEXT: str.w r10, [r11, #66] -; CHECK-NEXT: vcmp.f32 s22, s18 -; CHECK-NEXT: ldr r7, [sp, #20] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r7, #7 -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r7, #7 -; CHECK-NEXT: vcmp.f32 s22, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: and r5, r7, #15 -; CHECK-NEXT: vcmp.f32 s30, s18 -; CHECK-NEXT: lsrl r6, r5, #28 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcmp.f32 s16, s30 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r5, r4 -; CHECK-NEXT: strb.w r6, [r11, #74] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: vcmp.f32 s30, s20 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt.w r1, #-2147483648 +; CHECK-NEXT: vcmp.f32 s16, s28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r5, #-1 -; CHECK-NEXT: vcmp.f32 s30, s30 +; CHECK-NEXT: vcmp.f32 s16, s16 +; CHECK-NEXT: itt gt +; CHECK-NEXT: mvngt r1, #-2147483648 +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r5, #0 -; CHECK-NEXT: ldr r4, [sp] @ 4-byte Reload -; CHECK-NEXT: vcmp.f32 s30, s18 -; CHECK-NEXT: ldr r6, [sp, #16] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: lsrl r4, r5, #28 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movvs r0, #0 +; CHECK-NEXT: vcvtt.f32.f16 s16, s17 +; CHECK-NEXT: vmov q5[2], q5[0], r0, r7 +; CHECK-NEXT: vmov r0, s16 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r6, #0 -; CHECK-NEXT: orr.w r7, r5, r6, lsl #4 -; CHECK-NEXT: str.w r7, [r11, #45] -; CHECK-NEXT: str.w r4, [r11, #41] -; CHECK-NEXT: vcmp.f32 s30, s18 -; CHECK-NEXT: ldr r7, [sp, #28] @ 4-byte Reload +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: vmov q5[3], q5[1], r1, r6 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcmp.f32 s16, s30 +; CHECK-NEXT: mov r7, r0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r7, #7 -; CHECK-NEXT: vcmp.f32 s30, s20 +; CHECK-NEXT: mov r6, r1 +; CHECK-NEXT: vcmp.f32 s16, s28 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r6, #-2147483648 +; CHECK-NEXT: movlt r7, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r7, #7 -; CHECK-NEXT: vcmp.f32 s30, s30 +; CHECK-NEXT: vcmp.f32 s16, s16 +; CHECK-NEXT: vcvtb.f32.f16 s16, s17 +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r7, #-1 +; CHECK-NEXT: mvngt r6, #-2147483648 +; CHECK-NEXT: vmov r0, s16 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r6, #0 ; CHECK-NEXT: movvs r7, #0 -; CHECK-NEXT: and r5, r7, #15 -; CHECK-NEXT: vcmp.f32 s16, s18 -; CHECK-NEXT: lsrl r6, r5, #28 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcmp.f32 s16, s30 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: strb.w r6, [r11, #49] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt.w r1, #-2147483648 +; CHECK-NEXT: vcmp.f32 s16, s28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: itt gt +; CHECK-NEXT: mvngt r1, #-2147483648 +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s18 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: lsrl r0, r1, #28 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vcmp.f32 s16, s16 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movvs r0, #0 +; CHECK-NEXT: vcvtb.f32.f16 s16, s18 +; CHECK-NEXT: vmov q6[2], q6[0], r0, r7 +; CHECK-NEXT: vmov r0, s16 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: orr.w r1, r1, r2, lsl #4 -; CHECK-NEXT: vcmp.f32 s16, s18 +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: vmov q6[3], q6[1], r1, r6 +; CHECK-NEXT: bl __aeabi_f2lz +; CHECK-NEXT: vcmp.f32 s16, s30 +; CHECK-NEXT: vmov q3[2], q3[0], r10, r9 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: strd r0, r1, [r11, #16] -; CHECK-NEXT: it lt -; CHECK-NEXT: mvnlt r3, #7 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt.w r1, #-2147483648 +; CHECK-NEXT: vcmp.f32 s16, s28 +; CHECK-NEXT: vmov q3[3], q3[1], r11, r8 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r3, #7 +; CHECK-NEXT: itt gt +; CHECK-NEXT: mvngt r1, #-2147483648 +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: vcmp.f32 s16, s16 +; CHECK-NEXT: vmov q0, q5 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r3, #0 -; CHECK-NEXT: and r1, r3, #15 -; CHECK-NEXT: lsrl r2, r1, #28 -; CHECK-NEXT: strb.w r2, [r11, #24] -; CHECK-NEXT: add sp, #32 +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r0, #0 +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: vmov q2[2], q2[0], r0, r5 +; CHECK-NEXT: vmov q1, q6 +; CHECK-NEXT: vmov q2[3], q2[1], r1, r4 ; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15} ; CHECK-NEXT: add sp, #4 ; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} -; CHECK-NEXT: @ %bb.4: - %x = call <8 x i100> @llvm.fptosi.sat.v8f16.v8i100(<8 x half> %f) - ret <8 x i100> %x +; CHECK-NEXT: .p2align 2 +; CHECK-NEXT: @ %bb.1: +; CHECK-NEXT: .LCPI49_0: +; CHECK-NEXT: .long 0xdf000000 @ float -9.22337203E+18 +; CHECK-NEXT: .LCPI49_1: +; CHECK-NEXT: .long 0x5effffff @ float 9.22337149E+18 + %x = call <8 x i64> @llvm.fptosi.sat.v8f16.v8i64(<8 x half> %f) + ret <8 x i64> %x } -define arm_aapcs_vfpcc <8 x i128> @test_signed_v8f16_v8i128(<8 x half> %f) { -; CHECK-LABEL: test_signed_v8f16_v8i128: +define arm_aapcs_vfpcc <8 x i100> @test_signed_v8f16_v8i100(<8 x half> %f) { +; CHECK-LABEL: test_signed_v8f16_v8i100: ; CHECK: @ %bb.0: -; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, lr} -; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, lr} +; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, r11, lr} +; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} ; CHECK-NEXT: .pad #4 ; CHECK-NEXT: sub sp, #4 -; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13} +; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13} +; CHECK-NEXT: .pad #56 +; CHECK-NEXT: sub sp, #56 ; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: mov r4, r0 -; CHECK-NEXT: vcvtt.f32.f16 s28, s19 -; CHECK-NEXT: vcvtb.f32.f16 s20, s16 -; CHECK-NEXT: vmov r0, s28 +; CHECK-NEXT: mov r10, r0 ; CHECK-NEXT: vcvtt.f32.f16 s24, s16 -; CHECK-NEXT: vcvtb.f32.f16 s26, s17 -; CHECK-NEXT: vcvtb.f32.f16 s19, s19 -; CHECK-NEXT: vldr s22, .LCPI51_2 -; CHECK-NEXT: vmov r8, s20 -; CHECK-NEXT: vmov r9, s24 -; CHECK-NEXT: vcvtt.f32.f16 s30, s18 -; CHECK-NEXT: vmov r7, s26 -; CHECK-NEXT: vmov r6, s19 +; CHECK-NEXT: vmov r0, s24 ; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vldr s16, .LCPI51_3 -; CHECK-NEXT: vmov r5, s30 -; CHECK-NEXT: vcvtb.f32.f16 s18, s18 -; CHECK-NEXT: vcmp.f32 s28, s16 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r3, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r3, #-2147483648 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s16 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: vcvtt.f32.f16 s26, s17 +; CHECK-NEXT: mov r7, r0 +; CHECK-NEXT: vmov r0, s26 +; CHECK-NEXT: vldr s22, .LCPI50_0 +; CHECK-NEXT: vldr s20, .LCPI50_1 +; CHECK-NEXT: vcmp.f32 s24, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: str r3, [r4, #124] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: mvnlt r3, #7 ; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: vcmp.f32 s24, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: it gt +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r7, #-1 ; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: movgt r3, #7 +; CHECK-NEXT: vcmp.f32 s24, s24 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s16 +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: str r3, [sp, #52] @ 4-byte Spill ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: str r2, [sp, #48] @ 4-byte Spill +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r7, #0 +; CHECK-NEXT: str r7, [sp, #40] @ 4-byte Spill +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: str r1, [sp, #44] @ 4-byte Spill +; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: vcmp.f32 s26, s22 +; CHECK-NEXT: vcvtt.f32.f16 s24, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: str r2, [r4, #120] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: mvnlt r3, #7 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: vcmp.f32 s26, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: it gt +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: movgt r3, #7 +; CHECK-NEXT: vcmp.f32 s26, s26 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s16 +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: str r3, [sp, #36] @ 4-byte Spill +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: str r2, [sp, #32] @ 4-byte Spill ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: str r1, [sp, #28] @ 4-byte Spill +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r0, #0 +; CHECK-NEXT: str r0, [sp, #24] @ 4-byte Spill +; CHECK-NEXT: vmov r0, s24 +; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: vcmp.f32 s24, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #116] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: mvnlt r3, #7 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s28, s22 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: vcmp.f32 s24, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: vcmp.f32 s24, s24 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r1, #-1 ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s28, s28 +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: movgt r3, #7 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: str r3, [sp, #20] @ 4-byte Spill +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: str r2, [sp, #16] @ 4-byte Spill +; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4, #112] -; CHECK-NEXT: mov r0, r6 +; CHECK-NEXT: vcvtb.f32.f16 s24, s17 +; CHECK-NEXT: str r0, [sp, #8] @ 4-byte Spill +; CHECK-NEXT: vmov r0, s24 +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: str r1, [sp, #12] @ 4-byte Spill ; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s19, s16 -; CHECK-NEXT: vcvtt.f32.f16 s28, s17 +; CHECK-NEXT: vcvtb.f32.f16 s18, s18 +; CHECK-NEXT: mov r5, r0 +; CHECK-NEXT: vmov r0, s18 +; CHECK-NEXT: mov r8, r1 +; CHECK-NEXT: vcmp.f32 s24, s22 +; CHECK-NEXT: mov r6, r2 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: mvnlt r3, #7 +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movlt.w r8, #0 +; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: vcmp.f32 s24, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r3, #-2147483648 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r6, #-1 +; CHECK-NEXT: movgt.w r8, #-1 +; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: movgt r3, #7 +; CHECK-NEXT: vcmp.f32 s24, s24 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s16 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: str r3, [sp, #4] @ 4-byte Spill +; CHECK-NEXT: ittt vs +; CHECK-NEXT: movvs r5, #0 +; CHECK-NEXT: movvs.w r8, #0 +; CHECK-NEXT: movvs r6, #0 +; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: mov r7, r0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: str r3, [r4, #108] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: mov r11, r1 +; CHECK-NEXT: mov r4, r2 +; CHECK-NEXT: vcmp.f32 s18, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: mvnlt r3, #7 +; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: movlt.w r11, #0 +; CHECK-NEXT: movlt r4, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: vcmp.f32 s18, s18 +; CHECK-NEXT: vcvtb.f32.f16 s18, s19 +; CHECK-NEXT: vmov r0, s18 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r4, #-1 +; CHECK-NEXT: movgt.w r11, #-1 +; CHECK-NEXT: movgt.w r7, #-1 +; CHECK-NEXT: movgt r3, #7 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s16 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: str r3, [sp] @ 4-byte Spill +; CHECK-NEXT: ittt vs +; CHECK-NEXT: movvs r7, #0 +; CHECK-NEXT: movvs.w r11, #0 +; CHECK-NEXT: movvs r4, #0 +; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: mov r9, r3 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: str r2, [r4, #104] -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s18, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: mvnlt r9, #7 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s19 -; CHECK-NEXT: it gt +; CHECK-NEXT: vcmp.f32 s18, s18 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r9, #7 +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s16 +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: str.w r2, [r10, #83] ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #100] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s19, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s19, s19 +; CHECK-NEXT: str.w r1, [r10, #79] ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4, #96] -; CHECK-NEXT: mov r0, r5 -; CHECK-NEXT: vmov r6, s18 +; CHECK-NEXT: vcvtt.f32.f16 s18, s19 +; CHECK-NEXT: str.w r0, [r10, #75] +; CHECK-NEXT: vmov r0, s18 +; CHECK-NEXT: str.w r4, [r10, #58] +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: str.w r11, [r10, #54] +; CHECK-NEXT: str.w r7, [r10, #50] +; CHECK-NEXT: str.w r6, [r10, #33] +; CHECK-NEXT: str.w r8, [r10, #29] +; CHECK-NEXT: str.w r5, [r10, #25] +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs.w r9, #0 ; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s30, s16 +; CHECK-NEXT: vcvtb.f32.f16 s16, s16 +; CHECK-NEXT: mov r5, r0 +; CHECK-NEXT: vmov r0, s16 +; CHECK-NEXT: mov r7, r1 +; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: mov r6, r2 +; CHECK-NEXT: mov r4, r3 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s22 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: mvnlt r4, #7 +; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: it gt -; CHECK-NEXT: mvngt r3, #-2147483648 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: movgt.w r7, #-1 +; CHECK-NEXT: movgt.w r6, #-1 +; CHECK-NEXT: movgt r4, #7 +; CHECK-NEXT: vcmp.f32 s18, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s16 -; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: itttt vs +; CHECK-NEXT: movvs r4, #0 +; CHECK-NEXT: movvs r6, #0 +; CHECK-NEXT: movvs r7, #0 +; CHECK-NEXT: movvs r5, #0 +; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: vcmp.f32 s16, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s22 -; CHECK-NEXT: str r3, [r4, #92] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: mvnlt r3, #7 +; CHECK-NEXT: vcmp.f32 s16, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: it gt +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt r3, #7 +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: movgt.w r1, #-1 ; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: vcmp.f32 s16, s16 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s16 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s22 -; CHECK-NEXT: str r2, [r4, #88] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s16 +; CHECK-NEXT: str.w r2, [r10, #8] ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #84] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s30, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s30, s30 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: str.w r1, [r10, #4] ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4, #80] -; CHECK-NEXT: mov r0, r6 +; CHECK-NEXT: str.w r0, [r10] +; CHECK-NEXT: mov r0, r5 +; CHECK-NEXT: lsrl r0, r7, #28 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: orr.w r1, r7, r6, lsl #4 +; CHECK-NEXT: str.w r1, [r10, #95] +; CHECK-NEXT: and r1, r4, #15 +; CHECK-NEXT: str.w r0, [r10, #91] +; CHECK-NEXT: and r0, r9, #15 +; CHECK-NEXT: lsrl r6, r1, #28 +; CHECK-NEXT: strb.w r6, [r10, #99] +; CHECK-NEXT: orr.w r0, r0, r5, lsl #4 +; CHECK-NEXT: str.w r0, [r10, #87] +; CHECK-NEXT: ldr r7, [sp, #8] @ 4-byte Reload +; CHECK-NEXT: ldr r1, [sp, #12] @ 4-byte Reload +; CHECK-NEXT: ldr r2, [sp, #16] @ 4-byte Reload +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: lsrl r0, r1, #28 +; CHECK-NEXT: orr.w r1, r1, r2, lsl #4 +; CHECK-NEXT: str.w r1, [r10, #70] +; CHECK-NEXT: str.w r0, [r10, #66] +; CHECK-NEXT: ldr r0, [sp, #20] @ 4-byte Reload +; CHECK-NEXT: and r1, r0, #15 +; CHECK-NEXT: lsrl r2, r1, #28 +; CHECK-NEXT: strb.w r2, [r10, #74] +; CHECK-NEXT: ldr r0, [sp] @ 4-byte Reload +; CHECK-NEXT: and r0, r0, #15 +; CHECK-NEXT: orr.w r0, r0, r7, lsl #4 +; CHECK-NEXT: str.w r0, [r10, #62] +; CHECK-NEXT: ldr r7, [sp, #24] @ 4-byte Reload +; CHECK-NEXT: ldr r1, [sp, #28] @ 4-byte Reload +; CHECK-NEXT: ldr r2, [sp, #32] @ 4-byte Reload +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: lsrl r0, r1, #28 +; CHECK-NEXT: orr.w r1, r1, r2, lsl #4 +; CHECK-NEXT: str.w r1, [r10, #45] +; CHECK-NEXT: str.w r0, [r10, #41] +; CHECK-NEXT: ldr r0, [sp, #36] @ 4-byte Reload +; CHECK-NEXT: and r1, r0, #15 +; CHECK-NEXT: lsrl r2, r1, #28 +; CHECK-NEXT: strb.w r2, [r10, #49] +; CHECK-NEXT: ldr r0, [sp, #4] @ 4-byte Reload +; CHECK-NEXT: and r0, r0, #15 +; CHECK-NEXT: orr.w r0, r0, r7, lsl #4 +; CHECK-NEXT: str.w r0, [r10, #37] +; CHECK-NEXT: ldr r7, [sp, #40] @ 4-byte Reload +; CHECK-NEXT: ldr r1, [sp, #44] @ 4-byte Reload +; CHECK-NEXT: ldr r2, [sp, #48] @ 4-byte Reload +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: lsrl r0, r1, #28 +; CHECK-NEXT: orr.w r1, r1, r2, lsl #4 +; CHECK-NEXT: strd r0, r1, [r10, #16] +; CHECK-NEXT: ldr r0, [sp, #52] @ 4-byte Reload +; CHECK-NEXT: and r1, r0, #15 +; CHECK-NEXT: lsrl r2, r1, #28 +; CHECK-NEXT: strb.w r2, [r10, #24] +; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: and r0, r3, #15 +; CHECK-NEXT: orr.w r0, r0, r7, lsl #4 +; CHECK-NEXT: str.w r0, [r10, #12] +; CHECK-NEXT: add sp, #56 +; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13} +; CHECK-NEXT: add sp, #4 +; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} +; CHECK-NEXT: .p2align 2 +; CHECK-NEXT: @ %bb.1: +; CHECK-NEXT: .LCPI50_0: +; CHECK-NEXT: .long 0xf1000000 @ float -6.338253E+29 +; CHECK-NEXT: .LCPI50_1: +; CHECK-NEXT: .long 0x70ffffff @ float 6.33825262E+29 + %x = call <8 x i100> @llvm.fptosi.sat.v8f16.v8i100(<8 x half> %f) + ret <8 x i100> %x +} + +define arm_aapcs_vfpcc <8 x i128> @test_signed_v8f16_v8i128(<8 x half> %f) { +; CHECK-LABEL: test_signed_v8f16_v8i128: +; CHECK: @ %bb.0: +; CHECK-NEXT: .save {r4, r5, r6, r7, r8, lr} +; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, lr} +; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: vmov q4, q0 +; CHECK-NEXT: mov r4, r0 +; CHECK-NEXT: vcvtt.f32.f16 s26, s19 +; CHECK-NEXT: vcvtb.f32.f16 s28, s19 +; CHECK-NEXT: vmov r0, s26 +; CHECK-NEXT: vcvtb.f32.f16 s24, s17 +; CHECK-NEXT: vldr s20, .LCPI51_0 ; CHECK-NEXT: vmov r5, s28 +; CHECK-NEXT: vmov r8, s24 +; CHECK-NEXT: vcvtt.f32.f16 s30, s18 ; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s18, s16 +; CHECK-NEXT: vldr s22, .LCPI51_1 +; CHECK-NEXT: add.w r12, r4, #112 +; CHECK-NEXT: vmov r6, s30 +; CHECK-NEXT: vcmp.f32 s26, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s26, s20 +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt +; CHECK-NEXT: vcmp.f32 s26, s26 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: mvngt r3, #-2147483648 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s16 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: str r3, [r4, #76] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s16 -; CHECK-NEXT: it vs +; CHECK-NEXT: ittt vs ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: str r2, [r4, #72] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s16 -; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #68] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s18, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s18, s18 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4, #64] +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} ; CHECK-NEXT: mov r0, r5 +; CHECK-NEXT: vcvtb.f32.f16 s26, s18 ; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s28, s16 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: it lt +; CHECK-NEXT: add.w r12, r4, #96 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: vcmp.f32 s28, s20 +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: it gt +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: mvngt r3, #-2147483648 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s16 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: str r3, [r4, #60] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s16 -; CHECK-NEXT: it vs +; CHECK-NEXT: ittt vs ; CHECK-NEXT: movvs r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s22 -; CHECK-NEXT: str r2, [r4, #56] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s28 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s16 -; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: movvs r0, #0 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r6 +; CHECK-NEXT: vmov r7, s26 +; CHECK-NEXT: vcvtt.f32.f16 s28, s17 +; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: vcmp.f32 s30, s22 +; CHECK-NEXT: add.w r12, r4, #80 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #52] -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s30, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s28, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: vcmp.f32 s30, s30 +; CHECK-NEXT: itttt gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s28, s28 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: mvngt r3, #-2147483648 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: ittt vs +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: movvs r1, #0 ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4, #48] +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} ; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: vmov r5, s28 +; CHECK-NEXT: vcvtt.f32.f16 s18, s16 ; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s26, s16 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s26, s22 -; CHECK-NEXT: it lt +; CHECK-NEXT: add.w r12, r4, #64 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: vcmp.f32 s26, s20 +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s26, s26 -; CHECK-NEXT: it gt +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: mvngt r3, #-2147483648 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s16 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r3, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s22 -; CHECK-NEXT: str r3, [r4, #44] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s26 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s16 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s22 -; CHECK-NEXT: str r2, [r4, #40] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s26 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s16 -; CHECK-NEXT: b.w .LBB51_3 -; CHECK-NEXT: .p2align 2 -; CHECK-NEXT: @ %bb.1: -; CHECK-NEXT: .LCPI51_2: -; CHECK-NEXT: .long 0x7effffff @ float 1.70141173E+38 -; CHECK-NEXT: .p2align 2 -; CHECK-NEXT: @ %bb.2: -; CHECK-NEXT: .LCPI51_3: -; CHECK-NEXT: .long 0xff000000 @ float -1.70141183E+38 -; CHECK-NEXT: .p2align 1 -; CHECK-NEXT: .LBB51_3: -; CHECK-NEXT: it vs +; CHECK-NEXT: itt vs ; CHECK-NEXT: movvs r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #36] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s26, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s26, s26 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4, #32] -; CHECK-NEXT: mov r0, r9 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r5 +; CHECK-NEXT: vcvtb.f32.f16 s16, s16 +; CHECK-NEXT: vmov r6, s18 ; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s24, s16 +; CHECK-NEXT: vcmp.f32 s28, s22 +; CHECK-NEXT: add.w r12, r4, #48 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s22 -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s28, s20 +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s24 -; CHECK-NEXT: it gt +; CHECK-NEXT: vcmp.f32 s28, s28 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: mvngt r3, #-2147483648 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s16 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r3, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s22 -; CHECK-NEXT: str r3, [r4, #28] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s24 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s16 ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r2, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: itt vs +; CHECK-NEXT: movvs r1, #0 +; CHECK-NEXT: movvs r0, #0 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r8 +; CHECK-NEXT: vmov r7, s16 +; CHECK-NEXT: bl __fixsfti ; CHECK-NEXT: vcmp.f32 s24, s22 -; CHECK-NEXT: str r2, [r4, #24] -; CHECK-NEXT: it lt +; CHECK-NEXT: add.w r12, r4, #32 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: vcmp.f32 s24, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s24, s24 -; CHECK-NEXT: it gt +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: mvngt r3, #-2147483648 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s16 -; CHECK-NEXT: it vs +; CHECK-NEXT: ittt vs +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: movvs r2, #0 ; CHECK-NEXT: movvs r1, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #20] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s24, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s24, s24 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4, #16] -; CHECK-NEXT: mov r0, r8 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r6 ; CHECK-NEXT: bl __fixsfti -; CHECK-NEXT: vcmp.f32 s20, s16 +; CHECK-NEXT: vcmp.f32 s18, s22 +; CHECK-NEXT: add.w r12, r4, #16 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s22 -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s18, s20 +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s20 -; CHECK-NEXT: it gt +; CHECK-NEXT: vcmp.f32 s18, s18 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: mvngt r3, #-2147483648 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s16 -; CHECK-NEXT: it vs +; CHECK-NEXT: ittt vs ; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: movvs r1, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s22 -; CHECK-NEXT: str r3, [r4, #12] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s20 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s16 ; CHECK-NEXT: it vs -; CHECK-NEXT: movvs r2, #0 +; CHECK-NEXT: movvs r0, #0 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: bl __fixsfti +; CHECK-NEXT: vcmp.f32 s16, s22 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s22 -; CHECK-NEXT: str r2, [r4, #8] -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s16, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt.w r3, #-2147483648 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s20 -; CHECK-NEXT: it gt +; CHECK-NEXT: vcmp.f32 s16, s16 +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: mvngt r3, #-2147483648 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s16 -; CHECK-NEXT: it vs +; CHECK-NEXT: ittt vs +; CHECK-NEXT: movvs r3, #0 +; CHECK-NEXT: movvs r2, #0 ; CHECK-NEXT: movvs r1, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #4] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s20, s22 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s20, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it vs ; CHECK-NEXT: movvs r0, #0 -; CHECK-NEXT: str r0, [r4] +; CHECK-NEXT: stm r4!, {r0, r1, r2, r3} ; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: add sp, #4 -; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, pc} -; CHECK-NEXT: @ %bb.4: +; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, pc} +; CHECK-NEXT: .p2align 2 +; CHECK-NEXT: @ %bb.1: +; CHECK-NEXT: .LCPI51_0: +; CHECK-NEXT: .long 0x7effffff @ float 1.70141173E+38 +; CHECK-NEXT: .LCPI51_1: +; CHECK-NEXT: .long 0xff000000 @ float -1.70141183E+38 %x = call <8 x i128> @llvm.fptosi.sat.v8f16.v8i128(<8 x half> %f) ret <8 x i128> %x } diff --git a/llvm/test/CodeGen/Thumb2/mve-fptoui-sat-vector.ll b/llvm/test/CodeGen/Thumb2/mve-fptoui-sat-vector.ll index 5ab184a066e49..13609bd1903f2 100644 --- a/llvm/test/CodeGen/Thumb2/mve-fptoui-sat-vector.ll +++ b/llvm/test/CodeGen/Thumb2/mve-fptoui-sat-vector.ll @@ -39,40 +39,28 @@ define arm_aapcs_vfpcc <2 x i32> @test_unsigned_v2f32_v2i32(<2 x float> %f) { ; CHECK-NEXT: vmov r0, s16 ; CHECK-NEXT: vldr s18, .LCPI1_0 ; CHECK-NEXT: vcmp.f32 s17, #0 +; CHECK-NEXT: mov r4, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movlt r4, #0 ; CHECK-NEXT: vcmp.f32 s17, s18 -; CHECK-NEXT: mov r4, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt r4, #0 ; CHECK-NEXT: movgt.w r5, #-1 ; CHECK-NEXT: bl __aeabi_f2ulz ; CHECK-NEXT: vcmp.f32 s16, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s18 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s18 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vcmp.f32 s16, s18 -; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: movgt r1, #0 +; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 ; CHECK-NEXT: vmov q0[3], q0[1], r1, r4 ; CHECK-NEXT: vpop {d8, d9} ; CHECK-NEXT: pop {r4, r5, r7, pc} @@ -1059,38 +1047,26 @@ define arm_aapcs_vfpcc <2 x i32> @test_unsigned_v2f16_v2i32(<2 x half> %f) { ; CHECK-NEXT: vcmp.f32 s18, #0 ; CHECK-NEXT: mov r4, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movlt r4, #0 ; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt r4, #0 ; CHECK-NEXT: movgt.w r5, #-1 ; CHECK-NEXT: bl __aeabi_f2ulz ; CHECK-NEXT: vcmp.f32 s16, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: movgt r1, #0 +; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 ; CHECK-NEXT: vmov q0[3], q0[1], r1, r4 ; CHECK-NEXT: vpop {d8, d9, d10} ; CHECK-NEXT: pop {r4, r5, r7, pc} @@ -1504,112 +1480,91 @@ define arm_aapcs_vfpcc <4 x i50> @test_unsigned_v4f32_v4i50(<4 x float> %f) { ; CHECK-NEXT: .vsave {d8, d9, d10} ; CHECK-NEXT: vpush {d8, d9, d10} ; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: mov r8, r0 -; CHECK-NEXT: vmov r0, s16 -; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: mov r6, r0 +; CHECK-NEXT: mov r9, r0 ; CHECK-NEXT: vmov r0, s18 -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: mov r9, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vmov r4, s19 -; CHECK-NEXT: vldr s20, .LCPI28_0 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 ; CHECK-NEXT: bl __aeabi_f2ulz +; CHECK-NEXT: mov r8, r0 +; CHECK-NEXT: vmov r0, s19 +; CHECK-NEXT: vldr s20, .LCPI28_0 ; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r10, r0 -; CHECK-NEXT: mov r0, r4 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r5, #65535 -; CHECK-NEXT: movtgt r5, #3 -; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: mov r4, r0 -; CHECK-NEXT: vmov r0, s17 -; CHECK-NEXT: vcmp.f32 s19, #0 ; CHECK-NEXT: mov r7, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r8, #0 ; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: itt gt +; CHECK-NEXT: vmov r5, s16 +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movwgt r7, #65535 ; CHECK-NEXT: movtgt r7, #3 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movgt.w r8, #-1 +; CHECK-NEXT: bl __aeabi_f2ulz +; CHECK-NEXT: mov r4, r0 ; CHECK-NEXT: vcmp.f32 s19, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 +; CHECK-NEXT: mov r10, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: str.w r6, [r8] -; CHECK-NEXT: it lt +; CHECK-NEXT: mov r0, r5 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r10, #0 ; CHECK-NEXT: movlt r4, #0 +; CHECK-NEXT: vcmp.f32 s19, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: mov r1, r7 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movwgt r10, #65535 +; CHECK-NEXT: movtgt r10, #3 ; CHECK-NEXT: movgt.w r4, #-1 +; CHECK-NEXT: bl __aeabi_f2ulz +; CHECK-NEXT: mov r5, r1 +; CHECK-NEXT: vcmp.f32 s16, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r10, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: bfc r1, #18, #14 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: vcmp.f32 s16, s20 +; CHECK-NEXT: mov r1, r10 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r10, #-1 -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: bfc r5, #18, #14 -; CHECK-NEXT: mov r6, r10 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: str.w r0, [r9] +; CHECK-NEXT: vmov r0, s17 +; CHECK-NEXT: bfc r1, #18, #14 +; CHECK-NEXT: bfc r7, #18, #14 +; CHECK-NEXT: mov r6, r8 ; CHECK-NEXT: lsll r4, r1, #22 -; CHECK-NEXT: lsrl r6, r5, #28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r9, #0 -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: orrs r1, r5 +; CHECK-NEXT: lsrl r6, r7, #28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r9, #65535 -; CHECK-NEXT: movtgt r9, #3 -; CHECK-NEXT: str.w r1, [r8, #20] +; CHECK-NEXT: movwgt r5, #65535 +; CHECK-NEXT: movtgt r5, #3 +; CHECK-NEXT: orrs r1, r7 +; CHECK-NEXT: str.w r1, [r9, #20] ; CHECK-NEXT: bl __aeabi_f2ulz ; CHECK-NEXT: vcmp.f32 s17, #0 ; CHECK-NEXT: orr.w r2, r6, r4 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: bfc r9, #18, #14 +; CHECK-NEXT: bfc r5, #18, #14 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: itt gt ; CHECK-NEXT: movwgt r1, #65535 ; CHECK-NEXT: movtgt r1, #3 -; CHECK-NEXT: str.w r2, [r8, #16] -; CHECK-NEXT: lsrs r2, r7, #10 -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: strb.w r2, [r8, #24] -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: bfc r1, #18, #14 +; CHECK-NEXT: str.w r2, [r9, #16] +; CHECK-NEXT: lsr.w r2, r10, #10 +; CHECK-NEXT: strb.w r2, [r9, #24] ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: mov r2, r0 -; CHECK-NEXT: orr.w r0, r9, r0, lsl #18 +; CHECK-NEXT: bfc r1, #18, #14 +; CHECK-NEXT: orr.w r0, r5, r0, lsl #18 ; CHECK-NEXT: lsrl r2, r1, #14 -; CHECK-NEXT: orr.w r1, r1, r10, lsl #4 -; CHECK-NEXT: strd r2, r1, [r8, #8] -; CHECK-NEXT: str.w r0, [r8, #4] +; CHECK-NEXT: orr.w r1, r1, r8, lsl #4 +; CHECK-NEXT: strd r2, r1, [r9, #8] +; CHECK-NEXT: str.w r0, [r9, #4] ; CHECK-NEXT: vpop {d8, d9, d10} ; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, pc} ; CHECK-NEXT: .p2align 2 @@ -1636,85 +1591,61 @@ define arm_aapcs_vfpcc <4 x i64> @test_unsigned_v4f32_v4i64(<4 x float> %f) { ; CHECK-NEXT: vmov r0, s18 ; CHECK-NEXT: vldr s20, .LCPI29_0 ; CHECK-NEXT: vcmp.f32 s19, #0 +; CHECK-NEXT: mov r10, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt.w r11, #0 +; CHECK-NEXT: movlt.w r10, #0 ; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: mov r10, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vmov r9, s17 ; CHECK-NEXT: vmov r8, s16 -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r10, #-1 ; CHECK-NEXT: movgt.w r11, #-1 ; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcmp.f32 s18, #0 ; CHECK-NEXT: mov r7, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r7, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r10, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r6, r1 ; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r10, #-1 +; CHECK-NEXT: mov r6, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: mov r0, r9 -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: movlt r7, #0 ; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r7, #-1 ; CHECK-NEXT: movgt.w r6, #-1 ; CHECK-NEXT: bl __aeabi_f2ulz ; CHECK-NEXT: mov r5, r0 ; CHECK-NEXT: vcmp.f32 s17, #0 +; CHECK-NEXT: mov r4, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: mov r0, r8 -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movlt r4, #0 ; CHECK-NEXT: vcmp.f32 s17, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r4, r1 -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r4, #-1 ; CHECK-NEXT: movgt.w r5, #-1 ; CHECK-NEXT: bl __aeabi_f2ulz ; CHECK-NEXT: vcmp.f32 s16, #0 ; CHECK-NEXT: vmov q1[2], q1[0], r7, r11 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r4, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 +; CHECK-NEXT: vmov q1[3], q1[1], r6, r10 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: vmov q0[2], q0[0], r0, r5 ; CHECK-NEXT: vmov q0[3], q0[1], r1, r4 -; CHECK-NEXT: vmov q1[3], q1[1], r6, r10 ; CHECK-NEXT: vpop {d8, d9, d10} ; CHECK-NEXT: add sp, #4 ; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} @@ -1735,179 +1666,125 @@ define arm_aapcs_vfpcc <4 x i100> @test_unsigned_v4f32_v4i100(<4 x float> %f) { ; CHECK-NEXT: sub sp, #4 ; CHECK-NEXT: .vsave {d8, d9, d10} ; CHECK-NEXT: vpush {d8, d9, d10} +; CHECK-NEXT: .pad #8 +; CHECK-NEXT: sub sp, #8 ; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: mov r8, r0 -; CHECK-NEXT: vmov r0, s18 +; CHECK-NEXT: mov r4, r0 +; CHECK-NEXT: vmov r0, s17 ; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: mov r9, r3 -; CHECK-NEXT: vmov r3, s16 +; CHECK-NEXT: mov r5, r0 +; CHECK-NEXT: vmov r0, s18 ; CHECK-NEXT: vldr s20, .LCPI30_0 +; CHECK-NEXT: vcmp.f32 s17, #0 +; CHECK-NEXT: mov r7, r1 +; CHECK-NEXT: mov r6, r2 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: movlt r3, #0 +; CHECK-NEXT: vcmp.f32 s17, s20 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: it gt +; CHECK-NEXT: movgt r3, #15 +; CHECK-NEXT: str r3, [sp, #4] @ 4-byte Spill +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movgt.w r6, #-1 +; CHECK-NEXT: movgt.w r7, #-1 +; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: str r5, [sp] @ 4-byte Spill +; CHECK-NEXT: bl __fixunssfti ; CHECK-NEXT: vcmp.f32 s18, #0 +; CHECK-NEXT: mov r10, r3 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt.w r10, #0 +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: str.w r2, [r8, #33] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 +; CHECK-NEXT: str.w r2, [r4, #33] ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r8, #29] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 +; CHECK-NEXT: str.w r1, [r4, #29] +; CHECK-NEXT: vmov r1, s19 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str.w r0, [r8, #25] -; CHECK-NEXT: vmov r7, s17 -; CHECK-NEXT: vmov r4, s19 -; CHECK-NEXT: mov r0, r3 -; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: mov r10, r3 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 +; CHECK-NEXT: str.w r0, [r4, #25] ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: movgt.w r10, #15 +; CHECK-NEXT: mov r0, r1 +; CHECK-NEXT: bl __fixunssfti +; CHECK-NEXT: mov r5, r0 +; CHECK-NEXT: vmov r0, s16 +; CHECK-NEXT: vcmp.f32 s19, #0 +; CHECK-NEXT: mov r9, r1 +; CHECK-NEXT: mov r8, r2 +; CHECK-NEXT: mov r11, r3 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: str.w r2, [r8, #8] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt.w r9, #0 +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movlt.w r8, #0 +; CHECK-NEXT: movlt.w r11, #0 +; CHECK-NEXT: vcmp.f32 s19, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r11, #15 +; CHECK-NEXT: movgt.w r8, #-1 +; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: movgt.w r9, #-1 +; CHECK-NEXT: bl __fixunssfti ; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r8, #4] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: vcmp.f32 s16, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str.w r0, [r8] -; CHECK-NEXT: mov r0, r4 -; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s19, #0 -; CHECK-NEXT: mov r4, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r4, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r9, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r9, #15 -; CHECK-NEXT: and r0, r9, #15 -; CHECK-NEXT: mov r6, r2 -; CHECK-NEXT: orr.w r0, r0, r4, lsl #4 -; CHECK-NEXT: str.w r0, [r8, #37] -; CHECK-NEXT: mov r0, r7 -; CHECK-NEXT: mov r11, r3 -; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r10, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r10, #15 -; CHECK-NEXT: and r7, r10, #15 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: orr.w r7, r7, r0, lsl #4 -; CHECK-NEXT: str.w r7, [r8, #12] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r5, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: lsrl r4, r5, #28 -; CHECK-NEXT: vcmp.f32 s19, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: orr.w r7, r5, r6, lsl #4 -; CHECK-NEXT: str.w r7, [r8, #45] -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r4, [r8, #41] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r11, #0 -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r11, #15 -; CHECK-NEXT: and r5, r11, #15 -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: lsrl r6, r5, #28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: strb.w r6, [r8, #49] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: str r2, [r4, #8] ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: lsrl r0, r1, #28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: str r1, [r4, #4] ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: orr.w r1, r1, r2, lsl #4 -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: strd r0, r1, [r8, #16] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r3, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: str r0, [r4] +; CHECK-NEXT: mov r0, r5 +; CHECK-NEXT: lsrl r0, r9, #28 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: orr.w r1, r9, r8, lsl #4 +; CHECK-NEXT: str.w r1, [r4, #45] +; CHECK-NEXT: and r1, r11, #15 +; CHECK-NEXT: str.w r0, [r4, #41] +; CHECK-NEXT: and r0, r10, #15 +; CHECK-NEXT: lsrl r8, r1, #28 +; CHECK-NEXT: strb.w r8, [r4, #49] +; CHECK-NEXT: orr.w r0, r0, r5, lsl #4 +; CHECK-NEXT: str.w r0, [r4, #37] +; CHECK-NEXT: ldr r2, [sp] @ 4-byte Reload +; CHECK-NEXT: mov r0, r2 +; CHECK-NEXT: lsrl r0, r7, #28 +; CHECK-NEXT: orr.w r1, r7, r6, lsl #4 +; CHECK-NEXT: strd r0, r1, [r4, #16] +; CHECK-NEXT: ldr r0, [sp, #4] @ 4-byte Reload +; CHECK-NEXT: and r1, r0, #15 +; CHECK-NEXT: lsrl r6, r1, #28 +; CHECK-NEXT: strb r6, [r4, #24] ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt r3, #15 -; CHECK-NEXT: and r1, r3, #15 -; CHECK-NEXT: lsrl r2, r1, #28 -; CHECK-NEXT: strb.w r2, [r8, #24] +; CHECK-NEXT: and r0, r3, #15 +; CHECK-NEXT: orr.w r0, r0, r2, lsl #4 +; CHECK-NEXT: str r0, [r4, #12] +; CHECK-NEXT: add sp, #8 ; CHECK-NEXT: vpop {d8, d9, d10} ; CHECK-NEXT: add sp, #4 ; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} @@ -1932,160 +1809,87 @@ define arm_aapcs_vfpcc <4 x i128> @test_unsigned_v4f32_v4i128(<4 x float> %f) { ; CHECK-NEXT: mov r4, r0 ; CHECK-NEXT: vmov r0, s19 ; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vmov r5, s18 +; CHECK-NEXT: mov r5, r0 +; CHECK-NEXT: vmov r0, s18 ; CHECK-NEXT: vldr s20, .LCPI31_0 ; CHECK-NEXT: vcmp.f32 s19, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 +; CHECK-NEXT: vcmp.f32 s19, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, #0 ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r3, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: str r3, [r4, #60] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: str r2, [r4, #56] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s19, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #52] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s19, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4, #48] -; CHECK-NEXT: vmov r7, s16 +; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: strd r5, r1, [r4, #48] ; CHECK-NEXT: vmov r6, s17 -; CHECK-NEXT: mov r0, r5 +; CHECK-NEXT: vmov r7, s16 +; CHECK-NEXT: strd r2, r3, [r4, #56] ; CHECK-NEXT: bl __fixunssfti ; CHECK-NEXT: vcmp.f32 s18, #0 +; CHECK-NEXT: add.w r12, r4, #32 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movgt.w r3, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: str r3, [r4, #44] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: str r2, [r4, #40] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #36] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4, #32] +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} ; CHECK-NEXT: mov r0, r6 ; CHECK-NEXT: bl __fixunssfti ; CHECK-NEXT: vcmp.f32 s17, #0 +; CHECK-NEXT: add.w r12, r4, #16 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movgt.w r3, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: str r3, [r4, #28] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: str r2, [r4, #24] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #20] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s17, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4, #16] +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} ; CHECK-NEXT: mov r0, r7 ; CHECK-NEXT: bl __fixunssfti ; CHECK-NEXT: vcmp.f32 s16, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movgt.w r3, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: str r3, [r4, #12] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: str r2, [r4, #8] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #4] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4] +; CHECK-NEXT: stm r4!, {r0, r1, r2, r3} ; CHECK-NEXT: vpop {d8, d9, d10} ; CHECK-NEXT: add sp, #4 ; CHECK-NEXT: pop {r4, r5, r6, r7, pc} @@ -3762,232 +3566,198 @@ define arm_aapcs_vfpcc <8 x i50> @test_unsigned_v8f16_v8i50(<8 x half> %f) { ; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} ; CHECK-NEXT: .pad #4 ; CHECK-NEXT: sub sp, #4 -; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14} -; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14} -; CHECK-NEXT: .pad #8 -; CHECK-NEXT: sub sp, #8 +; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13} +; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13} +; CHECK-NEXT: .pad #24 +; CHECK-NEXT: sub sp, #24 ; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: mov r10, r0 +; CHECK-NEXT: mov r4, r0 +; CHECK-NEXT: vcvtb.f32.f16 s22, s17 +; CHECK-NEXT: vmov r0, s22 +; CHECK-NEXT: bl __aeabi_f2ulz ; CHECK-NEXT: vcvtb.f32.f16 s24, s18 +; CHECK-NEXT: mov r2, r0 ; CHECK-NEXT: vmov r0, s24 +; CHECK-NEXT: vcvtt.f32.f16 s20, s18 +; CHECK-NEXT: vldr s18, .LCPI48_0 +; CHECK-NEXT: vcmp.f32 s22, #0 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: vcvtt.f32.f16 s26, s17 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: vcmp.f32 s22, s18 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: itt gt +; CHECK-NEXT: movwgt r1, #65535 +; CHECK-NEXT: movtgt r1, #3 +; CHECK-NEXT: str r1, [sp, #16] @ 4-byte Spill +; CHECK-NEXT: it gt +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: vmov r5, s26 +; CHECK-NEXT: str r2, [sp, #20] @ 4-byte Spill +; CHECK-NEXT: vmov r6, s20 ; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcvtt.f32.f16 s28, s19 ; CHECK-NEXT: mov r7, r0 -; CHECK-NEXT: vmov r0, s28 -; CHECK-NEXT: vcvtb.f32.f16 s22, s16 -; CHECK-NEXT: vcvtb.f32.f16 s26, s19 ; CHECK-NEXT: vcmp.f32 s24, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r9, r1 -; CHECK-NEXT: vmov r5, s22 -; CHECK-NEXT: vldr s20, .LCPI48_0 -; CHECK-NEXT: vmov r11, s26 -; CHECK-NEXT: it lt +; CHECK-NEXT: mov r0, r5 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcmp.f32 s28, #0 -; CHECK-NEXT: mov r4, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: mov r6, r0 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r4, #0 +; CHECK-NEXT: vcmp.f32 s24, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r0, r5 -; CHECK-NEXT: vcmp.f32 s24, s20 ; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r4, #65535 -; CHECK-NEXT: movtgt r4, #3 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movwgt r1, #65535 +; CHECK-NEXT: movtgt r1, #3 +; CHECK-NEXT: str r1, [sp, #12] @ 4-byte Spill ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r7, #-1 -; CHECK-NEXT: str.w r7, [r10, #25] ; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: str r1, [sp, #4] @ 4-byte Spill +; CHECK-NEXT: vcmp.f32 s26, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: vcmp.f32 s26, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, #0 +; CHECK-NEXT: itt gt +; CHECK-NEXT: movwgt r1, #65535 +; CHECK-NEXT: movtgt r1, #3 +; CHECK-NEXT: str r1, [sp, #8] @ 4-byte Spill +; CHECK-NEXT: str.w r7, [r4, #25] ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: mov r7, r4 -; CHECK-NEXT: str.w r0, [r10] +; CHECK-NEXT: str r0, [sp, #4] @ 4-byte Spill +; CHECK-NEXT: mov r0, r6 +; CHECK-NEXT: bl __aeabi_f2ulz +; CHECK-NEXT: vcmp.f32 s20, #0 +; CHECK-NEXT: mov r6, r0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s20, s18 +; CHECK-NEXT: vcvtb.f32.f16 s20, s19 +; CHECK-NEXT: mov r7, r1 +; CHECK-NEXT: vmov r0, s20 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r7, #0 ; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: mov r0, r11 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: bfc r7, #18, #14 -; CHECK-NEXT: lsll r6, r7, #22 +; CHECK-NEXT: movwgt r7, #65535 +; CHECK-NEXT: movtgt r7, #3 ; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: mov r5, r0 +; CHECK-NEXT: vcmp.f32 s20, #0 +; CHECK-NEXT: mov r9, r0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: vcmp.f32 s20, s18 +; CHECK-NEXT: vcvtt.f32.f16 s20, s19 +; CHECK-NEXT: mov r11, r1 +; CHECK-NEXT: vmov r0, s20 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r11, #0 +; CHECK-NEXT: movlt.w r9, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movgt.w r9, #-1 +; CHECK-NEXT: movwgt r11, #65535 +; CHECK-NEXT: movtgt r11, #3 +; CHECK-NEXT: bl __aeabi_f2ulz +; CHECK-NEXT: vcmp.f32 s20, #0 +; CHECK-NEXT: mov r10, r0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r1, #65535 -; CHECK-NEXT: movtgt r1, #3 -; CHECK-NEXT: mov r2, r5 -; CHECK-NEXT: bfc r1, #18, #14 -; CHECK-NEXT: vcvtt.f32.f16 s26, s18 -; CHECK-NEXT: lsrl r2, r1, #28 -; CHECK-NEXT: orr.w r0, r1, r7 -; CHECK-NEXT: str.w r0, [r10, #45] -; CHECK-NEXT: vmov r0, s26 -; CHECK-NEXT: orrs r6, r2 -; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: mov r7, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: vcvtb.f32.f16 s18, s17 -; CHECK-NEXT: lsrs r0, r4, #10 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r1, #65535 -; CHECK-NEXT: movtgt r1, #3 -; CHECK-NEXT: str.w r6, [r10, #41] -; CHECK-NEXT: strb.w r0, [r10, #49] -; CHECK-NEXT: vmov r0, s18 -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: bfc r1, #18, #14 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r7, #-1 -; CHECK-NEXT: mov r4, r7 -; CHECK-NEXT: lsrl r4, r1, #14 -; CHECK-NEXT: orr.w r6, r1, r5, lsl #4 -; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcvtt.f32.f16 s26, s17 -; CHECK-NEXT: mov r11, r0 -; CHECK-NEXT: vmov r0, s26 +; CHECK-NEXT: vcmp.f32 s20, s18 +; CHECK-NEXT: vcvtb.f32.f16 s20, s16 ; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: vmov r0, s20 +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 +; CHECK-NEXT: movlt.w r10, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movgt.w r10, #-1 ; CHECK-NEXT: movwgt r5, #65535 ; CHECK-NEXT: movtgt r5, #3 -; CHECK-NEXT: str.w r6, [r10, #37] -; CHECK-NEXT: str.w r4, [r10, #33] ; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: mov r6, r1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, #0 -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r6, #65535 -; CHECK-NEXT: movtgt r6, #3 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r9, #0 -; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: mov r4, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r9, #65535 -; CHECK-NEXT: movtgt r9, #3 -; CHECK-NEXT: bfc r9, #18, #14 -; CHECK-NEXT: vcvtt.f32.f16 s16, s16 -; CHECK-NEXT: orr.w r0, r9, r7, lsl #18 -; CHECK-NEXT: str.w r0, [r10, #29] -; CHECK-NEXT: vmov r0, s16 -; CHECK-NEXT: mov r1, r6 -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: bfc r1, #18, #14 +; CHECK-NEXT: vcmp.f32 s20, #0 +; CHECK-NEXT: mov r8, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r4, #0 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt.w r8, #0 +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: vcmp.f32 s20, s18 +; CHECK-NEXT: bfc r11, #18, #14 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r4, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r11, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: mov r2, r9 +; CHECK-NEXT: str r0, [r4] +; CHECK-NEXT: lsrs r0, r5, #10 ; CHECK-NEXT: bfc r5, #18, #14 +; CHECK-NEXT: lsll r10, r5, #22 +; CHECK-NEXT: lsrl r2, r11, #28 +; CHECK-NEXT: orr.w r1, r11, r5 +; CHECK-NEXT: str.w r1, [r4, #45] +; CHECK-NEXT: orr.w r1, r2, r10 +; CHECK-NEXT: str.w r1, [r4, #41] +; CHECK-NEXT: strb.w r0, [r4, #49] +; CHECK-NEXT: bfc r7, #18, #14 +; CHECK-NEXT: mov r0, r6 +; CHECK-NEXT: vcvtt.f32.f16 s16, s16 +; CHECK-NEXT: lsrl r0, r7, #14 +; CHECK-NEXT: mov r5, r4 +; CHECK-NEXT: orr.w r1, r7, r9, lsl #4 +; CHECK-NEXT: str.w r1, [r4, #37] +; CHECK-NEXT: str.w r0, [r4, #33] ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r11, #-1 -; CHECK-NEXT: mov r8, r11 -; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: ldr r7, [sp, #4] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: ldr r0, [sp, #12] @ 4-byte Reload +; CHECK-NEXT: bfc r0, #18, #14 +; CHECK-NEXT: orr.w r0, r0, r6, lsl #18 +; CHECK-NEXT: str.w r0, [r4, #29] +; CHECK-NEXT: vmov r0, s16 +; CHECK-NEXT: ldr r7, [sp, #8] @ 4-byte Reload +; CHECK-NEXT: ldr r3, [sp, #16] @ 4-byte Reload +; CHECK-NEXT: ldr.w r9, [sp, #20] @ 4-byte Reload +; CHECK-NEXT: mov r1, r7 +; CHECK-NEXT: ldr r4, [sp, #4] @ 4-byte Reload +; CHECK-NEXT: bfc r1, #18, #14 +; CHECK-NEXT: bfc r3, #18, #14 +; CHECK-NEXT: mov r6, r9 ; CHECK-NEXT: lsll r4, r1, #22 -; CHECK-NEXT: lsrl r8, r5, #28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: lsrl r6, r3, #28 ; CHECK-NEXT: itt gt -; CHECK-NEXT: movwgt r7, #65535 -; CHECK-NEXT: movtgt r7, #3 -; CHECK-NEXT: orrs r1, r5 -; CHECK-NEXT: str.w r1, [r10, #20] +; CHECK-NEXT: movwgt r8, #65535 +; CHECK-NEXT: movtgt r8, #3 +; CHECK-NEXT: orrs r1, r3 +; CHECK-NEXT: str r1, [r5, #20] ; CHECK-NEXT: bl __aeabi_f2ulz ; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: orr.w r2, r8, r4 +; CHECK-NEXT: orr.w r2, r6, r4 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: bfc r7, #18, #14 +; CHECK-NEXT: vcmp.f32 s16, s18 +; CHECK-NEXT: bfc r8, #18, #14 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: itt gt ; CHECK-NEXT: movwgt r1, #65535 ; CHECK-NEXT: movtgt r1, #3 -; CHECK-NEXT: str.w r2, [r10, #16] -; CHECK-NEXT: lsrs r2, r6, #10 -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: strb.w r2, [r10, #24] -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: bfc r1, #18, #14 +; CHECK-NEXT: str r2, [r5, #16] +; CHECK-NEXT: lsrs r2, r7, #10 +; CHECK-NEXT: strb r2, [r5, #24] ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: mov r2, r0 -; CHECK-NEXT: orr.w r0, r7, r0, lsl #18 +; CHECK-NEXT: bfc r1, #18, #14 +; CHECK-NEXT: orr.w r0, r8, r0, lsl #18 ; CHECK-NEXT: lsrl r2, r1, #14 -; CHECK-NEXT: orr.w r1, r1, r11, lsl #4 -; CHECK-NEXT: strd r2, r1, [r10, #8] -; CHECK-NEXT: str.w r0, [r10, #4] -; CHECK-NEXT: add sp, #8 -; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14} +; CHECK-NEXT: orr.w r1, r1, r9, lsl #4 +; CHECK-NEXT: strd r2, r1, [r5, #8] +; CHECK-NEXT: str r0, [r5, #4] +; CHECK-NEXT: add sp, #24 +; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13} ; CHECK-NEXT: add sp, #4 ; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} ; CHECK-NEXT: .p2align 2 @@ -4005,8 +3775,8 @@ define arm_aapcs_vfpcc <8 x i64> @test_unsigned_v8f16_v8i64(<8 x half> %f) { ; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} ; CHECK-NEXT: .pad #4 ; CHECK-NEXT: sub sp, #4 -; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14} +; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14} ; CHECK-NEXT: vmov q4, q0 ; CHECK-NEXT: vcvtt.f32.f16 s20, s19 ; CHECK-NEXT: vmov r0, s20 @@ -4016,178 +3786,128 @@ define arm_aapcs_vfpcc <8 x i64> @test_unsigned_v8f16_v8i64(<8 x half> %f) { ; CHECK-NEXT: vmov r0, s22 ; CHECK-NEXT: vldr s28, .LCPI49_0 ; CHECK-NEXT: vcmp.f32 s20, #0 -; CHECK-NEXT: vcvtt.f32.f16 s24, s16 +; CHECK-NEXT: mov r8, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcvtb.f32.f16 s16, s16 -; CHECK-NEXT: it lt +; CHECK-NEXT: vcvtt.f32.f16 s24, s18 +; CHECK-NEXT: vcvtt.f32.f16 s26, s16 +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt.w r9, #0 +; CHECK-NEXT: movlt.w r8, #0 ; CHECK-NEXT: vcmp.f32 s20, s28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r8, r1 -; CHECK-NEXT: vmov r5, s24 -; CHECK-NEXT: it gt +; CHECK-NEXT: vmov r4, s24 +; CHECK-NEXT: vmov r6, s26 +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r8, #-1 ; CHECK-NEXT: movgt.w r9, #-1 -; CHECK-NEXT: vmov r4, s16 ; CHECK-NEXT: bl __aeabi_f2ulz +; CHECK-NEXT: mov r10, r0 ; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: mov r11, r0 +; CHECK-NEXT: mov r11, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, s28 -; CHECK-NEXT: it lt +; CHECK-NEXT: mov r0, r4 +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt.w r11, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r11, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s20, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r8, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r10, r1 -; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r8, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r0, r5 -; CHECK-NEXT: it lt ; CHECK-NEXT: movlt.w r10, #0 ; CHECK-NEXT: vcmp.f32 s22, s28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt ; CHECK-NEXT: movgt.w r10, #-1 +; CHECK-NEXT: movgt.w r11, #-1 ; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: mov r6, r0 +; CHECK-NEXT: mov r5, r0 ; CHECK-NEXT: vcmp.f32 s24, #0 +; CHECK-NEXT: mov r4, r1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r0, r4 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: mov r0, r6 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movlt r4, #0 ; CHECK-NEXT: vcmp.f32 s24, s28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcvtt.f32.f16 s30, s17 -; CHECK-NEXT: mov r7, r1 -; CHECK-NEXT: vmov r1, s30 -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s16, s28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vmov q5[2], q5[0], r0, r6 -; CHECK-NEXT: mov r0, r1 +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r4, #-1 +; CHECK-NEXT: movgt.w r5, #-1 ; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcmp.f32 s30, #0 -; CHECK-NEXT: mov r6, r0 +; CHECK-NEXT: vcvtb.f32.f16 s16, s16 +; CHECK-NEXT: mov r7, r0 +; CHECK-NEXT: vmov r0, s16 +; CHECK-NEXT: mov r6, r1 +; CHECK-NEXT: vcmp.f32 s26, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s28 -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r7, #0 ; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: vcmp.f32 s26, s28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt ; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r5, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s28 -; CHECK-NEXT: vcvtb.f32.f16 s16, s17 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: vmov r0, s16 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r7, #-1 -; CHECK-NEXT: mov r4, r1 -; CHECK-NEXT: vmov q5[3], q5[1], r7, r5 ; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcvtt.f32.f16 s17, s18 -; CHECK-NEXT: mov r7, r1 -; CHECK-NEXT: vmov r1, s17 ; CHECK-NEXT: vcmp.f32 s16, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vcmp.f32 s16, s28 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vmov q6[2], q6[0], r0, r6 -; CHECK-NEXT: mov r0, r1 -; CHECK-NEXT: bl __aeabi_f2ulz -; CHECK-NEXT: vcmp.f32 s17, #0 -; CHECK-NEXT: mov r6, r0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, #0 +; CHECK-NEXT: vcvtt.f32.f16 s16, s17 +; CHECK-NEXT: vmov q5[2], q5[0], r0, r7 +; CHECK-NEXT: vmov r0, s16 ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r4, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: vmov q5[3], q5[1], r1, r6 +; CHECK-NEXT: bl __aeabi_f2ulz ; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r4, #-1 +; CHECK-NEXT: mov r7, r0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s16, s28 -; CHECK-NEXT: vcvtb.f32.f16 s16, s18 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: vcvtb.f32.f16 s16, s17 +; CHECK-NEXT: mov r6, r1 ; CHECK-NEXT: vmov r0, s16 +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: movlt r6, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r6, #-1 ; CHECK-NEXT: movgt.w r7, #-1 -; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: vmov q6[3], q6[1], r7, r4 ; CHECK-NEXT: bl __aeabi_f2ulz ; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: vmov q3[2], q3[0], r11, r9 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s16, s28 -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt +; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, #0 ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s17, s28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 +; CHECK-NEXT: vcvtb.f32.f16 s16, s18 +; CHECK-NEXT: vmov q6[2], q6[0], r0, r7 +; CHECK-NEXT: vmov r0, s16 ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: vmov q6[3], q6[1], r1, r6 +; CHECK-NEXT: bl __aeabi_f2ulz +; CHECK-NEXT: vcmp.f32 s16, #0 +; CHECK-NEXT: vmov q3[2], q3[0], r10, r9 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt +; CHECK-NEXT: itt lt ; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: vcmp.f32 s16, s28 -; CHECK-NEXT: vmov q2[2], q2[0], r0, r6 +; CHECK-NEXT: vmov q3[3], q3[1], r11, r8 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt +; CHECK-NEXT: movgt.w r0, #-1 ; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmov q2[3], q2[1], r1, r5 -; CHECK-NEXT: vmov q3[3], q3[1], r10, r8 +; CHECK-NEXT: vmov q2[2], q2[0], r0, r5 ; CHECK-NEXT: vmov q0, q5 +; CHECK-NEXT: vmov q2[3], q2[1], r1, r4 ; CHECK-NEXT: vmov q1, q6 -; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14} ; CHECK-NEXT: add sp, #4 ; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} ; CHECK-NEXT: .p2align 2 @@ -4205,385 +3925,272 @@ define arm_aapcs_vfpcc <8 x i100> @test_unsigned_v8f16_v8i100(<8 x half> %f) { ; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} ; CHECK-NEXT: .pad #4 ; CHECK-NEXT: sub sp, #4 -; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: .pad #32 -; CHECK-NEXT: sub sp, #32 +; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12} +; CHECK-NEXT: vpush {d8, d9, d10, d11, d12} +; CHECK-NEXT: .pad #56 +; CHECK-NEXT: sub sp, #56 ; CHECK-NEXT: vmov q4, q0 -; CHECK-NEXT: mov r8, r0 -; CHECK-NEXT: vcvtb.f32.f16 s30, s19 -; CHECK-NEXT: vcvtb.f32.f16 s28, s18 -; CHECK-NEXT: vmov r0, s30 -; CHECK-NEXT: vcvtt.f32.f16 s22, s19 -; CHECK-NEXT: vcvtb.f32.f16 s24, s16 -; CHECK-NEXT: vcvtb.f32.f16 s26, s17 -; CHECK-NEXT: vldr s20, .LCPI50_1 -; CHECK-NEXT: vmov r4, s22 -; CHECK-NEXT: vmov r7, s28 -; CHECK-NEXT: vcvtt.f32.f16 s18, s18 -; CHECK-NEXT: vmov r9, s24 -; CHECK-NEXT: vmov r6, s26 +; CHECK-NEXT: mov r4, r0 +; CHECK-NEXT: vcvtt.f32.f16 s22, s16 +; CHECK-NEXT: vmov r0, s22 ; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s30, #0 -; CHECK-NEXT: mov r5, r3 +; CHECK-NEXT: vcvtt.f32.f16 s24, s17 +; CHECK-NEXT: mov r7, r0 +; CHECK-NEXT: vmov r0, s24 +; CHECK-NEXT: vldr s20, .LCPI50_0 +; CHECK-NEXT: vcmp.f32 s22, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r3, #0 +; CHECK-NEXT: vcmp.f32 s22, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, #0 +; CHECK-NEXT: it gt +; CHECK-NEXT: movgt r3, #15 +; CHECK-NEXT: str r3, [sp, #52] @ 4-byte Spill ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: str.w r2, [r8, #83] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, #0 +; CHECK-NEXT: str r2, [sp, #48] @ 4-byte Spill ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r8, #79] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: str r1, [sp, #44] @ 4-byte Spill ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str.w r0, [r8, #75] -; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: movgt.w r7, #-1 +; CHECK-NEXT: str r7, [sp, #40] @ 4-byte Spill ; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s28, #0 -; CHECK-NEXT: mov r7, r3 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: vcmp.f32 s24, #0 +; CHECK-NEXT: vcvtt.f32.f16 s22, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: str.w r2, [r8, #58] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r8, #54] -; CHECK-NEXT: it lt ; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str.w r0, [r8, #50] -; CHECK-NEXT: mov r0, r6 -; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: str r3, [sp, #24] @ 4-byte Spill -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: it lt ; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r3, #0 +; CHECK-NEXT: vcmp.f32 s24, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, #0 ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: str.w r2, [r8, #33] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, #0 +; CHECK-NEXT: movgt r3, #15 +; CHECK-NEXT: str r3, [sp, #36] @ 4-byte Spill ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r8, #29] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: str r2, [sp, #32] @ 4-byte Spill ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str.w r0, [r8, #25] -; CHECK-NEXT: mov r0, r9 +; CHECK-NEXT: str r0, [sp, #24] @ 4-byte Spill +; CHECK-NEXT: vmov r0, s22 +; CHECK-NEXT: it gt +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: str r1, [sp, #28] @ 4-byte Spill ; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s24, #0 -; CHECK-NEXT: str r3, [sp, #12] @ 4-byte Spill +; CHECK-NEXT: vcmp.f32 s22, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s22, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 ; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, #0 +; CHECK-NEXT: it gt +; CHECK-NEXT: movgt r3, #15 +; CHECK-NEXT: str r3, [sp, #20] @ 4-byte Spill ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: str.w r2, [r8, #8] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, #0 +; CHECK-NEXT: str r2, [sp, #16] @ 4-byte Spill ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r1, [r8, #4] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: str r1, [sp, #12] @ 4-byte Spill ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str.w r0, [r8] -; CHECK-NEXT: mov r0, r4 +; CHECK-NEXT: vcvtb.f32.f16 s22, s17 +; CHECK-NEXT: str r0, [sp, #4] @ 4-byte Spill +; CHECK-NEXT: vmov r0, s22 ; CHECK-NEXT: bl __fixunssfti +; CHECK-NEXT: vcvtb.f32.f16 s18, s18 +; CHECK-NEXT: mov r9, r0 +; CHECK-NEXT: vmov r0, s18 +; CHECK-NEXT: mov r8, r1 ; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: mov r6, r0 +; CHECK-NEXT: mov r6, r2 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: str r3, [sp, #8] @ 4-byte Spill -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: movlt.w r8, #0 +; CHECK-NEXT: movlt.w r9, #0 +; CHECK-NEXT: movlt r3, #0 +; CHECK-NEXT: vcmp.f32 s22, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, #0 ; CHECK-NEXT: it gt +; CHECK-NEXT: movgt r3, #15 +; CHECK-NEXT: str r3, [sp, #8] @ 4-byte Spill +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movgt.w r9, #-1 +; CHECK-NEXT: movgt.w r8, #-1 ; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r5, #15 -; CHECK-NEXT: and r0, r5, #15 -; CHECK-NEXT: mov r9, r1 -; CHECK-NEXT: orr.w r1, r0, r6, lsl #4 -; CHECK-NEXT: vmov r0, s18 -; CHECK-NEXT: mov r4, r2 -; CHECK-NEXT: str.w r1, [r8, #87] ; CHECK-NEXT: bl __fixunssfti ; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: mov r10, r0 +; CHECK-NEXT: mov r5, r0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: str r2, [sp, #4] @ 4-byte Spill -; CHECK-NEXT: mov r5, r1 -; CHECK-NEXT: str r3, [sp, #20] @ 4-byte Spill -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r10, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r10, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: vcvtb.f32.f16 s18, s19 +; CHECK-NEXT: mov r11, r1 +; CHECK-NEXT: vmov r0, s18 +; CHECK-NEXT: mov r7, r2 +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt r7, #0 +; CHECK-NEXT: movlt.w r11, #0 +; CHECK-NEXT: movlt r5, #0 +; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r7, #15 -; CHECK-NEXT: and r0, r7, #15 -; CHECK-NEXT: vcvtt.f32.f16 s28, s17 -; CHECK-NEXT: orr.w r0, r0, r10, lsl #4 -; CHECK-NEXT: str.w r0, [r8, #62] -; CHECK-NEXT: vmov r0, s28 +; CHECK-NEXT: movgt r3, #15 +; CHECK-NEXT: str r3, [sp] @ 4-byte Spill +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: movgt.w r11, #-1 +; CHECK-NEXT: movgt.w r7, #-1 ; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s28, #0 -; CHECK-NEXT: str r2, [sp, #16] @ 4-byte Spill +; CHECK-NEXT: vcmp.f32 s18, #0 +; CHECK-NEXT: mov r10, r3 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r3, [sp, #28] @ 4-byte Spill -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s18, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt.w r10, #0 ; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s28, s20 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: mov r11, r1 -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: mov r1, r0 -; CHECK-NEXT: str r0, [sp] @ 4-byte Spill -; CHECK-NEXT: ldr r0, [sp, #24] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: str.w r2, [r4, #83] ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r0, #15 -; CHECK-NEXT: and r0, r0, #15 -; CHECK-NEXT: vcvtt.f32.f16 s16, s16 -; CHECK-NEXT: orr.w r0, r0, r1, lsl #4 -; CHECK-NEXT: str.w r0, [r8, #37] -; CHECK-NEXT: vmov r0, s16 -; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s16, s20 +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: str.w r1, [r4, #79] ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: vcmp.f32 s24, #0 -; CHECK-NEXT: ldr r7, [sp, #12] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r7, #15 -; CHECK-NEXT: and r7, r7, #15 -; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: orr.w r7, r7, r0, lsl #4 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str.w r7, [r8, #12] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r9, #0 -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r9, #-1 -; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: lsrl r6, r9, #28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r4, #0 -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r4, #-1 -; CHECK-NEXT: orr.w r7, r9, r4, lsl #4 -; CHECK-NEXT: str.w r7, [r8, #95] -; CHECK-NEXT: str.w r6, [r8, #91] -; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: ldr r7, [sp, #8] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: vcmp.f32 s22, s20 +; CHECK-NEXT: vcvtt.f32.f16 s18, s19 +; CHECK-NEXT: str.w r0, [r4, #75] +; CHECK-NEXT: vmov r0, s18 +; CHECK-NEXT: str.w r7, [r4, #58] +; CHECK-NEXT: str.w r11, [r4, #54] ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: str.w r5, [r4, #50] +; CHECK-NEXT: str.w r6, [r4, #33] +; CHECK-NEXT: str.w r8, [r4, #29] +; CHECK-NEXT: str.w r9, [r4, #25] ; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r7, #15 -; CHECK-NEXT: and r7, r7, #15 +; CHECK-NEXT: movgt.w r10, #15 +; CHECK-NEXT: bl __fixunssfti +; CHECK-NEXT: vcvtb.f32.f16 s16, s16 +; CHECK-NEXT: mov r5, r0 +; CHECK-NEXT: vmov r0, s16 +; CHECK-NEXT: mov r7, r1 ; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: lsrl r4, r7, #28 +; CHECK-NEXT: mov r6, r2 +; CHECK-NEXT: mov r8, r3 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: strb.w r4, [r8, #99] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r7, #0 ; CHECK-NEXT: movlt r5, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r5, #-1 -; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: ldr r6, [sp, #4] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: lsrl r10, r5, #28 -; CHECK-NEXT: it lt ; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: movlt.w r8, #0 ; CHECK-NEXT: vcmp.f32 s18, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: orr.w r7, r5, r6, lsl #4 -; CHECK-NEXT: str.w r7, [r8, #70] -; CHECK-NEXT: str.w r10, [r8, #66] -; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: ldr r7, [sp, #20] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r7, #15 -; CHECK-NEXT: and r5, r7, #15 -; CHECK-NEXT: vcmp.f32 s28, #0 -; CHECK-NEXT: lsrl r6, r5, #28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: strb.w r6, [r8, #74] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt.w r11, #0 -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r11, #-1 -; CHECK-NEXT: ldr r4, [sp] @ 4-byte Reload -; CHECK-NEXT: vcmp.f32 s28, #0 -; CHECK-NEXT: ldr r6, [sp, #16] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: lsrl r4, r11, #28 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r6, #0 -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: b.w .LBB50_2 -; CHECK-NEXT: .p2align 2 -; CHECK-NEXT: @ %bb.1: -; CHECK-NEXT: .LCPI50_1: -; CHECK-NEXT: .long 0x717fffff @ float 1.26765052E+30 -; CHECK-NEXT: .p2align 1 -; CHECK-NEXT: .LBB50_2: -; CHECK-NEXT: it gt +; CHECK-NEXT: itttt gt +; CHECK-NEXT: movgt.w r8, #15 ; CHECK-NEXT: movgt.w r6, #-1 -; CHECK-NEXT: orr.w r7, r11, r6, lsl #4 -; CHECK-NEXT: str.w r7, [r8, #45] -; CHECK-NEXT: str.w r4, [r8, #41] -; CHECK-NEXT: vcmp.f32 s28, #0 -; CHECK-NEXT: ldr r7, [sp, #28] @ 4-byte Reload -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r7, #0 -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt r7, #15 -; CHECK-NEXT: and r5, r7, #15 +; CHECK-NEXT: movgt.w r5, #-1 +; CHECK-NEXT: movgt.w r7, #-1 +; CHECK-NEXT: bl __fixunssfti ; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: lsrl r6, r5, #28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: strb.w r6, [r8, #49] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r3, #0 +; CHECK-NEXT: movlt r0, #0 ; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: lsrl r0, r1, #28 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it lt ; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: vcmp.f32 s16, s20 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: orr.w r1, r1, r2, lsl #4 -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: strd r0, r1, [r8, #16] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r3, #0 +; CHECK-NEXT: str r2, [r4, #8] +; CHECK-NEXT: it gt +; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: str r1, [r4, #4] +; CHECK-NEXT: it gt +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: str r0, [r4] +; CHECK-NEXT: mov r0, r5 +; CHECK-NEXT: lsrl r0, r7, #28 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: orr.w r1, r7, r6, lsl #4 +; CHECK-NEXT: str.w r1, [r4, #95] +; CHECK-NEXT: and r1, r8, #15 +; CHECK-NEXT: str.w r0, [r4, #91] +; CHECK-NEXT: and r0, r10, #15 +; CHECK-NEXT: lsrl r6, r1, #28 +; CHECK-NEXT: strb.w r6, [r4, #99] +; CHECK-NEXT: orr.w r0, r0, r5, lsl #4 +; CHECK-NEXT: str.w r0, [r4, #87] +; CHECK-NEXT: ldr r7, [sp, #4] @ 4-byte Reload +; CHECK-NEXT: ldr r1, [sp, #12] @ 4-byte Reload +; CHECK-NEXT: ldr r2, [sp, #16] @ 4-byte Reload +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: lsrl r0, r1, #28 +; CHECK-NEXT: orr.w r1, r1, r2, lsl #4 +; CHECK-NEXT: str.w r1, [r4, #70] +; CHECK-NEXT: str.w r0, [r4, #66] +; CHECK-NEXT: ldr r0, [sp, #20] @ 4-byte Reload +; CHECK-NEXT: and r1, r0, #15 +; CHECK-NEXT: lsrl r2, r1, #28 +; CHECK-NEXT: strb.w r2, [r4, #74] +; CHECK-NEXT: ldr r0, [sp] @ 4-byte Reload +; CHECK-NEXT: and r0, r0, #15 +; CHECK-NEXT: orr.w r0, r0, r7, lsl #4 +; CHECK-NEXT: str.w r0, [r4, #62] +; CHECK-NEXT: ldr r7, [sp, #24] @ 4-byte Reload +; CHECK-NEXT: ldr r1, [sp, #28] @ 4-byte Reload +; CHECK-NEXT: ldr r2, [sp, #32] @ 4-byte Reload +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: lsrl r0, r1, #28 +; CHECK-NEXT: orr.w r1, r1, r2, lsl #4 +; CHECK-NEXT: str.w r1, [r4, #45] +; CHECK-NEXT: str.w r0, [r4, #41] +; CHECK-NEXT: ldr r0, [sp, #36] @ 4-byte Reload +; CHECK-NEXT: and r1, r0, #15 +; CHECK-NEXT: lsrl r2, r1, #28 +; CHECK-NEXT: strb.w r2, [r4, #49] +; CHECK-NEXT: ldr r0, [sp, #8] @ 4-byte Reload +; CHECK-NEXT: and r0, r0, #15 +; CHECK-NEXT: orr.w r0, r0, r7, lsl #4 +; CHECK-NEXT: str.w r0, [r4, #37] +; CHECK-NEXT: ldr r7, [sp, #40] @ 4-byte Reload +; CHECK-NEXT: ldr r1, [sp, #44] @ 4-byte Reload +; CHECK-NEXT: ldr r2, [sp, #48] @ 4-byte Reload +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: lsrl r0, r1, #28 +; CHECK-NEXT: orr.w r1, r1, r2, lsl #4 +; CHECK-NEXT: strd r0, r1, [r4, #16] +; CHECK-NEXT: ldr r0, [sp, #52] @ 4-byte Reload +; CHECK-NEXT: and r1, r0, #15 +; CHECK-NEXT: lsrl r2, r1, #28 +; CHECK-NEXT: strb r2, [r4, #24] ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt r3, #15 -; CHECK-NEXT: and r1, r3, #15 -; CHECK-NEXT: lsrl r2, r1, #28 -; CHECK-NEXT: strb.w r2, [r8, #24] -; CHECK-NEXT: add sp, #32 -; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: and r0, r3, #15 +; CHECK-NEXT: orr.w r0, r0, r7, lsl #4 +; CHECK-NEXT: str r0, [r4, #12] +; CHECK-NEXT: add sp, #56 +; CHECK-NEXT: vpop {d8, d9, d10, d11, d12} ; CHECK-NEXT: add sp, #4 ; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} -; CHECK-NEXT: @ %bb.3: +; CHECK-NEXT: .p2align 2 +; CHECK-NEXT: @ %bb.1: +; CHECK-NEXT: .LCPI50_0: +; CHECK-NEXT: .long 0x717fffff @ float 1.26765052E+30 %x = call <8 x i100> @llvm.fptoui.sat.v8f16.v8i100(<8 x half> %f) ret <8 x i100> %x } @@ -4591,337 +4198,195 @@ define arm_aapcs_vfpcc <8 x i100> @test_unsigned_v8f16_v8i100(<8 x half> %f) { define arm_aapcs_vfpcc <8 x i128> @test_unsigned_v8f16_v8i128(<8 x half> %f) { ; CHECK-LABEL: test_unsigned_v8f16_v8i128: ; CHECK: @ %bb.0: -; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, lr} -; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, lr} +; CHECK-NEXT: .save {r4, r5, r6, r7, lr} +; CHECK-NEXT: push {r4, r5, r6, r7, lr} ; CHECK-NEXT: .pad #4 ; CHECK-NEXT: sub sp, #4 -; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15} -; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14} +; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14} ; CHECK-NEXT: vmov q4, q0 ; CHECK-NEXT: mov r4, r0 -; CHECK-NEXT: vcvtt.f32.f16 s26, s19 -; CHECK-NEXT: vcvtb.f32.f16 s22, s16 -; CHECK-NEXT: vmov r0, s26 -; CHECK-NEXT: vcvtt.f32.f16 s16, s16 -; CHECK-NEXT: vcvtb.f32.f16 s24, s17 -; CHECK-NEXT: vcvtb.f32.f16 s30, s19 -; CHECK-NEXT: vldr s20, .LCPI51_0 -; CHECK-NEXT: vmov r8, s22 -; CHECK-NEXT: vmov r9, s16 -; CHECK-NEXT: vcvtt.f32.f16 s28, s18 -; CHECK-NEXT: vmov r7, s24 -; CHECK-NEXT: vmov r6, s30 +; CHECK-NEXT: vcvtt.f32.f16 s22, s19 +; CHECK-NEXT: vmov r0, s22 ; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: vcvtb.f32.f16 s18, s18 +; CHECK-NEXT: vcvtb.f32.f16 s28, s19 +; CHECK-NEXT: mov r6, r0 +; CHECK-NEXT: vmov r0, s28 +; CHECK-NEXT: vldr s20, .LCPI51_0 +; CHECK-NEXT: vcmp.f32 s22, #0 +; CHECK-NEXT: vcvtt.f32.f16 s24, s18 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s22, s20 +; CHECK-NEXT: vcvtb.f32.f16 s26, s18 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r6, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, #0 ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r3, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: str r3, [r4, #124] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: str r2, [r4, #120] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #116] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4, #112] -; CHECK-NEXT: mov r0, r6 -; CHECK-NEXT: vmov r5, s28 +; CHECK-NEXT: movgt.w r6, #-1 +; CHECK-NEXT: strd r6, r1, [r4, #112] +; CHECK-NEXT: vmov r7, s24 +; CHECK-NEXT: vmov r5, s26 +; CHECK-NEXT: vcvtt.f32.f16 s18, s17 +; CHECK-NEXT: strd r2, r3, [r4, #120] ; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s30, #0 -; CHECK-NEXT: vcvtt.f32.f16 s26, s17 +; CHECK-NEXT: vcmp.f32 s28, #0 +; CHECK-NEXT: add.w r12, r4, #96 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s28, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, #0 ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r3, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: str r3, [r4, #108] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, #0 ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, s20 -; CHECK-NEXT: str r2, [r4, #104] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s30, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt ; CHECK-NEXT: movgt.w r1, #-1 +; CHECK-NEXT: movgt.w r0, #-1 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: vmov r6, s18 +; CHECK-NEXT: vcvtb.f32.f16 s22, s17 +; CHECK-NEXT: bl __fixunssfti +; CHECK-NEXT: vcmp.f32 s24, #0 +; CHECK-NEXT: add.w r12, r4, #80 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #100] -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s24, s20 +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s30, s20 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 +; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt +; CHECK-NEXT: movgt.w r3, #-1 +; CHECK-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-NEXT: ittt gt +; CHECK-NEXT: movgt.w r2, #-1 +; CHECK-NEXT: movgt.w r1, #-1 ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4, #96] +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} ; CHECK-NEXT: mov r0, r5 -; CHECK-NEXT: vmov r6, s18 +; CHECK-NEXT: vcvtt.f32.f16 s24, s16 +; CHECK-NEXT: vmov r7, s22 ; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s28, #0 +; CHECK-NEXT: vcmp.f32 s26, #0 +; CHECK-NEXT: add.w r12, r4, #64 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s26, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, #0 ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r3, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: str r3, [r4, #92] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, #0 ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: str r2, [r4, #88] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s28, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt ; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #84] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s28, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4, #80] +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} ; CHECK-NEXT: mov r0, r6 -; CHECK-NEXT: vmov r5, s26 +; CHECK-NEXT: vmov r5, s24 +; CHECK-NEXT: vcvtb.f32.f16 s16, s16 ; CHECK-NEXT: bl __fixunssfti ; CHECK-NEXT: vcmp.f32 s18, #0 +; CHECK-NEXT: add.w r12, r4, #48 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r3, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: str r3, [r4, #76] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: str r2, [r4, #72] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s18, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: itt gt ; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #68] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s18, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4, #64] -; CHECK-NEXT: mov r0, r5 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: vmov r6, s16 ; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s26, #0 +; CHECK-NEXT: vcmp.f32 s22, #0 +; CHECK-NEXT: add.w r12, r4, #32 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: vcmp.f32 s22, s20 +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movgt.w r3, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: str r3, [r4, #60] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: str r2, [r4, #56] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s26, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #52] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s26, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4, #48] -; CHECK-NEXT: mov r0, r7 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r5 ; CHECK-NEXT: bl __fixunssfti ; CHECK-NEXT: vcmp.f32 s24, #0 +; CHECK-NEXT: add.w r12, r4, #16 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt +; CHECK-NEXT: movlt r0, #0 +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movgt.w r3, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: str r3, [r4, #44] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: str r2, [r4, #40] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s24, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #36] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s24, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4, #32] -; CHECK-NEXT: mov r0, r9 +; CHECK-NEXT: stm.w r12, {r0, r1, r2, r3} +; CHECK-NEXT: mov r0, r6 ; CHECK-NEXT: bl __fixunssfti ; CHECK-NEXT: vcmp.f32 s16, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r3, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r3, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: str r3, [r4, #28] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: str r2, [r4, #24] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s16, #0 -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r1, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #20] -; CHECK-NEXT: it lt +; CHECK-NEXT: itttt lt ; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s16, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: it gt -; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4, #16] -; CHECK-NEXT: mov r0, r8 -; CHECK-NEXT: bl __fixunssfti -; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: it lt +; CHECK-NEXT: movlt r1, #0 +; CHECK-NEXT: movlt r2, #0 ; CHECK-NEXT: movlt r3, #0 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: it gt +; CHECK-NEXT: ittt gt ; CHECK-NEXT: movgt.w r3, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: str r3, [r4, #12] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r2, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r2, #-1 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: str r2, [r4, #8] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r1, #0 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: vcmp.f32 s22, #0 -; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r1, #-1 ; CHECK-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-NEXT: str r1, [r4, #4] -; CHECK-NEXT: it lt -; CHECK-NEXT: movlt r0, #0 -; CHECK-NEXT: vcmp.f32 s22, s20 -; CHECK-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-NEXT: it gt ; CHECK-NEXT: movgt.w r0, #-1 -; CHECK-NEXT: str r0, [r4] -; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: stm r4!, {r0, r1, r2, r3} +; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14} ; CHECK-NEXT: add sp, #4 -; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, pc} +; CHECK-NEXT: pop {r4, r5, r6, r7, pc} ; CHECK-NEXT: .p2align 2 ; CHECK-NEXT: @ %bb.1: ; CHECK-NEXT: .LCPI51_0: diff --git a/llvm/test/CodeGen/Thumb2/mve-pred-ext.ll b/llvm/test/CodeGen/Thumb2/mve-pred-ext.ll index 117469f3bd788..101b49fea488a 100644 --- a/llvm/test/CodeGen/Thumb2/mve-pred-ext.ll +++ b/llvm/test/CodeGen/Thumb2/mve-pred-ext.ll @@ -300,27 +300,27 @@ define arm_aapcs_vfpcc <8 x i16> @zext_v8i1_v8f32(<8 x half> %src1, <8 x half> % ; CHECK-MVE-NEXT: vcmp.f16 s10, s8 ; CHECK-MVE-NEXT: vmovx.f16 s8, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vcmp.f16 s3, s7 ; CHECK-MVE-NEXT: vmovx.f16 s10, s2 -; CHECK-MVE-NEXT: vcmp.f16 s10, s8 -; CHECK-MVE-NEXT: vmovx.f16 s8, s5 -; CHECK-MVE-NEXT: vmovx.f16 s10, s1 ; CHECK-MVE-NEXT: csetm r12, ne ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f16 s3, s7 +; CHECK-MVE-NEXT: vcmp.f16 s10, s8 ; CHECK-MVE-NEXT: csetm lr, ne ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f16 s10, s8 +; CHECK-MVE-NEXT: vcmp.f16 s2, s6 +; CHECK-MVE-NEXT: vmovx.f16 s2, s5 +; CHECK-MVE-NEXT: vmovx.f16 s6, s1 ; CHECK-MVE-NEXT: csetm r2, ne ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f16 s2, s6 +; CHECK-MVE-NEXT: vcmp.f16 s6, s2 ; CHECK-MVE-NEXT: vmovx.f16 s2, s4 ; CHECK-MVE-NEXT: vmovx.f16 s6, s0 ; CHECK-MVE-NEXT: csetm r3, ne ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f16 s6, s2 +; CHECK-MVE-NEXT: vcmp.f16 s1, s5 ; CHECK-MVE-NEXT: csetm r0, ne ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f16 s1, s5 +; CHECK-MVE-NEXT: vcmp.f16 s6, s2 ; CHECK-MVE-NEXT: csetm r1, ne ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s0, s4 @@ -329,12 +329,12 @@ define arm_aapcs_vfpcc <8 x i16> @zext_v8i1_v8f32(<8 x half> %src1, <8 x half> % ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: csetm r5, ne ; CHECK-MVE-NEXT: vmov.16 q1[0], r5 -; CHECK-MVE-NEXT: vmov.16 q1[1], r1 -; CHECK-MVE-NEXT: vmov.16 q1[2], r4 -; CHECK-MVE-NEXT: vmov.16 q1[3], r3 -; CHECK-MVE-NEXT: vmov.16 q1[4], r0 -; CHECK-MVE-NEXT: vmov.16 q1[5], lr -; CHECK-MVE-NEXT: vmov.16 q1[6], r2 +; CHECK-MVE-NEXT: vmov.16 q1[1], r4 +; CHECK-MVE-NEXT: vmov.16 q1[2], r1 +; CHECK-MVE-NEXT: vmov.16 q1[3], r0 +; CHECK-MVE-NEXT: vmov.16 q1[4], r3 +; CHECK-MVE-NEXT: vmov.16 q1[5], r2 +; CHECK-MVE-NEXT: vmov.16 q1[6], lr ; CHECK-MVE-NEXT: vmov.16 q1[7], r12 ; CHECK-MVE-NEXT: vand q0, q1, q0 ; CHECK-MVE-NEXT: pop {r4, r5, r7, pc} diff --git a/llvm/test/CodeGen/Thumb2/mve-vcmpf.ll b/llvm/test/CodeGen/Thumb2/mve-vcmpf.ll index baf0076277e50..6f2539e3cad9a 100644 --- a/llvm/test/CodeGen/Thumb2/mve-vcmpf.ll +++ b/llvm/test/CodeGen/Thumb2/mve-vcmpf.ll @@ -43,10 +43,8 @@ define arm_aapcs_vfpcc <4 x float> @vcmp_one_v4f32(<4 x float> %src, <4 x float> ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vcmp.f32 s1, s5 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s1, s5 -; CHECK-MVE-NEXT: cset r0, mi -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s0, s4 +; CHECK-MVE-NEXT: cset r0, mi ; CHECK-MVE-NEXT: csinc r0, r0, zr, le ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s3, s7 @@ -230,10 +228,8 @@ define arm_aapcs_vfpcc <4 x float> @vcmp_ueq_v4f32(<4 x float> %src, <4 x float> ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vcmp.f32 s1, s5 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s1, s5 -; CHECK-MVE-NEXT: cset r0, eq -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s0, s4 +; CHECK-MVE-NEXT: cset r0, eq ; CHECK-MVE-NEXT: csinc r0, r0, zr, vc ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s3, s7 @@ -271,17 +267,17 @@ entry: define arm_aapcs_vfpcc <4 x float> @vcmp_une_v4f32(<4 x float> %src, <4 x float> %src2, <4 x float> %a, <4 x float> %b) { ; CHECK-MVE-LABEL: vcmp_une_v4f32: ; CHECK-MVE: @ %bb.0: @ %entry -; CHECK-MVE-NEXT: vcmp.f32 s3, s7 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s2, s6 -; CHECK-MVE-NEXT: vseleq.f32 s3, s15, s11 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s1, s5 ; CHECK-MVE-NEXT: vseleq.f32 s2, s14, s10 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s0, s4 +; CHECK-MVE-NEXT: vcmp.f32 s3, s7 ; CHECK-MVE-NEXT: vseleq.f32 s1, s13, s9 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vcmp.f32 s0, s4 +; CHECK-MVE-NEXT: vseleq.f32 s3, s15, s11 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f32 s0, s12, s8 ; CHECK-MVE-NEXT: bx lr ; @@ -1090,53 +1086,53 @@ entry: define arm_aapcs_vfpcc <8 x half> @vcmp_une_v8f16(<8 x half> %src, <8 x half> %src2, <8 x half> %a, <8 x half> %b) { ; CHECK-MVE-LABEL: vcmp_une_v8f16: ; CHECK-MVE: @ %bb.0: @ %entry -; CHECK-MVE-NEXT: .vsave {d8, d9, d10, d11} -; CHECK-MVE-NEXT: vpush {d8, d9, d10, d11} +; CHECK-MVE-NEXT: .vsave {d8, d9} +; CHECK-MVE-NEXT: vpush {d8, d9} ; CHECK-MVE-NEXT: vmovx.f16 s16, s4 ; CHECK-MVE-NEXT: vmovx.f16 s18, s0 ; CHECK-MVE-NEXT: vcmp.f16 s18, s16 -; CHECK-MVE-NEXT: vmovx.f16 s20, s8 +; CHECK-MVE-NEXT: vmovx.f16 s16, s8 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s22, s12 +; CHECK-MVE-NEXT: vmovx.f16 s18, s12 ; CHECK-MVE-NEXT: vcmp.f16 s0, s4 ; CHECK-MVE-NEXT: vmovx.f16 s4, s5 -; CHECK-MVE-NEXT: vseleq.f16 s16, s22, s20 +; CHECK-MVE-NEXT: vseleq.f16 s16, s18, s16 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s0, s12, s8 ; CHECK-MVE-NEXT: vmovx.f16 s8, s1 ; CHECK-MVE-NEXT: vcmp.f16 s8, s4 -; CHECK-MVE-NEXT: vins.f16 s0, s16 +; CHECK-MVE-NEXT: vmovx.f16 s4, s9 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vmovx.f16 s8, s13 ; CHECK-MVE-NEXT: vcmp.f16 s1, s5 -; CHECK-MVE-NEXT: vmovx.f16 s12, s9 -; CHECK-MVE-NEXT: vmovx.f16 s16, s13 -; CHECK-MVE-NEXT: vmovx.f16 s8, s2 -; CHECK-MVE-NEXT: vmovx.f16 s5, s14 -; CHECK-MVE-NEXT: vseleq.f16 s4, s16, s12 +; CHECK-MVE-NEXT: vins.f16 s0, s16 +; CHECK-MVE-NEXT: vseleq.f16 s4, s8, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s12, s10 +; CHECK-MVE-NEXT: vmovx.f16 s8, s2 ; CHECK-MVE-NEXT: vseleq.f16 s1, s13, s9 ; CHECK-MVE-NEXT: vins.f16 s1, s4 ; CHECK-MVE-NEXT: vmovx.f16 s4, s6 ; CHECK-MVE-NEXT: vcmp.f16 s8, s4 -; CHECK-MVE-NEXT: vmovx.f16 s8, s11 +; CHECK-MVE-NEXT: vmovx.f16 s4, s10 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vmovx.f16 s8, s14 ; CHECK-MVE-NEXT: vcmp.f16 s2, s6 ; CHECK-MVE-NEXT: vmovx.f16 s6, s3 -; CHECK-MVE-NEXT: vseleq.f16 s4, s5, s12 +; CHECK-MVE-NEXT: vseleq.f16 s4, s8, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s2, s14, s10 -; CHECK-MVE-NEXT: vmovx.f16 s10, s15 ; CHECK-MVE-NEXT: vins.f16 s2, s4 ; CHECK-MVE-NEXT: vmovx.f16 s4, s7 ; CHECK-MVE-NEXT: vcmp.f16 s6, s4 +; CHECK-MVE-NEXT: vmovx.f16 s4, s11 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vmovx.f16 s6, s15 ; CHECK-MVE-NEXT: vcmp.f16 s3, s7 -; CHECK-MVE-NEXT: vseleq.f16 s4, s10, s8 +; CHECK-MVE-NEXT: vseleq.f16 s4, s6, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s3, s15, s11 ; CHECK-MVE-NEXT: vins.f16 s3, s4 -; CHECK-MVE-NEXT: vpop {d8, d9, d10, d11} +; CHECK-MVE-NEXT: vpop {d8, d9} ; CHECK-MVE-NEXT: bx lr ; ; CHECK-MVEFP-LABEL: vcmp_une_v8f16: diff --git a/llvm/test/CodeGen/Thumb2/mve-vcmpfr.ll b/llvm/test/CodeGen/Thumb2/mve-vcmpfr.ll index fe82255bff6c8..d42c393743f4f 100644 --- a/llvm/test/CodeGen/Thumb2/mve-vcmpfr.ll +++ b/llvm/test/CodeGen/Thumb2/mve-vcmpfr.ll @@ -46,10 +46,8 @@ define arm_aapcs_vfpcc <4 x float> @vcmp_one_v4f32(<4 x float> %src, float %src2 ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vcmp.f32 s1, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s1, s4 -; CHECK-MVE-NEXT: cset r0, mi -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s0, s4 +; CHECK-MVE-NEXT: cset r0, mi ; CHECK-MVE-NEXT: csinc r0, r0, zr, le ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s3, s4 @@ -248,10 +246,8 @@ define arm_aapcs_vfpcc <4 x float> @vcmp_ueq_v4f32(<4 x float> %src, float %src2 ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vcmp.f32 s1, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s1, s4 -; CHECK-MVE-NEXT: cset r0, eq -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s0, s4 +; CHECK-MVE-NEXT: cset r0, eq ; CHECK-MVE-NEXT: csinc r0, r0, zr, vc ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s3, s4 @@ -292,17 +288,17 @@ entry: define arm_aapcs_vfpcc <4 x float> @vcmp_une_v4f32(<4 x float> %src, float %src2, <4 x float> %a, <4 x float> %b) { ; CHECK-MVE-LABEL: vcmp_une_v4f32: ; CHECK-MVE: @ %bb.0: @ %entry -; CHECK-MVE-NEXT: vcmp.f32 s3, s4 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s2, s4 -; CHECK-MVE-NEXT: vseleq.f32 s3, s15, s11 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s1, s4 ; CHECK-MVE-NEXT: vseleq.f32 s2, s14, s10 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s0, s4 +; CHECK-MVE-NEXT: vcmp.f32 s3, s4 ; CHECK-MVE-NEXT: vseleq.f32 s1, s13, s9 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vcmp.f32 s0, s4 +; CHECK-MVE-NEXT: vseleq.f32 s3, s15, s11 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f32 s0, s12, s8 ; CHECK-MVE-NEXT: bx lr ; @@ -1105,42 +1101,42 @@ define arm_aapcs_vfpcc <8 x half> @vcmp_une_v8f16(<8 x half> %src, half %src2, < ; CHECK-MVE-LABEL: vcmp_une_v8f16: ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vmovx.f16 s6, s0 -; CHECK-MVE-NEXT: vmovx.f16 s5, s8 +; CHECK-MVE-NEXT: vmovx.f16 s5, s12 ; CHECK-MVE-NEXT: vcmp.f16 s6, s4 -; CHECK-MVE-NEXT: vmovx.f16 s7, s12 +; CHECK-MVE-NEXT: vmovx.f16 s6, s8 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s0, s4 -; CHECK-MVE-NEXT: vseleq.f16 s6, s7, s5 +; CHECK-MVE-NEXT: vseleq.f16 s6, s5, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s0, s12, s8 -; CHECK-MVE-NEXT: vmovx.f16 s8, s9 +; CHECK-MVE-NEXT: vmovx.f16 s8, s13 ; CHECK-MVE-NEXT: vins.f16 s0, s6 ; CHECK-MVE-NEXT: vmovx.f16 s6, s1 ; CHECK-MVE-NEXT: vcmp.f16 s6, s4 -; CHECK-MVE-NEXT: vmovx.f16 s12, s13 +; CHECK-MVE-NEXT: vmovx.f16 s6, s9 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s1, s4 -; CHECK-MVE-NEXT: vseleq.f16 s6, s12, s8 +; CHECK-MVE-NEXT: vseleq.f16 s6, s8, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s8, s10 -; CHECK-MVE-NEXT: vmovx.f16 s12, s14 +; CHECK-MVE-NEXT: vmovx.f16 s8, s14 ; CHECK-MVE-NEXT: vseleq.f16 s1, s13, s9 ; CHECK-MVE-NEXT: vins.f16 s1, s6 ; CHECK-MVE-NEXT: vmovx.f16 s6, s2 ; CHECK-MVE-NEXT: vcmp.f16 s6, s4 +; CHECK-MVE-NEXT: vmovx.f16 s6, s10 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s2, s4 -; CHECK-MVE-NEXT: vseleq.f16 s6, s12, s8 +; CHECK-MVE-NEXT: vseleq.f16 s6, s8, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s8, s11 +; CHECK-MVE-NEXT: vmovx.f16 s8, s15 ; CHECK-MVE-NEXT: vseleq.f16 s2, s14, s10 -; CHECK-MVE-NEXT: vmovx.f16 s10, s15 ; CHECK-MVE-NEXT: vins.f16 s2, s6 ; CHECK-MVE-NEXT: vmovx.f16 s6, s3 ; CHECK-MVE-NEXT: vcmp.f16 s6, s4 +; CHECK-MVE-NEXT: vmovx.f16 s6, s11 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s3, s4 -; CHECK-MVE-NEXT: vseleq.f16 s6, s10, s8 +; CHECK-MVE-NEXT: vseleq.f16 s6, s8, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s3, s15, s11 ; CHECK-MVE-NEXT: vins.f16 s3, s6 @@ -1659,10 +1655,8 @@ define arm_aapcs_vfpcc <4 x float> @vcmp_r_one_v4f32(<4 x float> %src, float %sr ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vcmp.f32 s4, s1 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s4, s1 -; CHECK-MVE-NEXT: cset r0, mi -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, s0 +; CHECK-MVE-NEXT: cset r0, mi ; CHECK-MVE-NEXT: csinc r0, r0, zr, le ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, s3 @@ -1861,10 +1855,8 @@ define arm_aapcs_vfpcc <4 x float> @vcmp_r_ueq_v4f32(<4 x float> %src, float %sr ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vcmp.f32 s4, s1 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s4, s1 -; CHECK-MVE-NEXT: cset r0, eq -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, s0 +; CHECK-MVE-NEXT: cset r0, eq ; CHECK-MVE-NEXT: csinc r0, r0, zr, vc ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, s3 @@ -1905,17 +1897,17 @@ entry: define arm_aapcs_vfpcc <4 x float> @vcmp_r_une_v4f32(<4 x float> %src, float %src2, <4 x float> %a, <4 x float> %b) { ; CHECK-MVE-LABEL: vcmp_r_une_v4f32: ; CHECK-MVE: @ %bb.0: @ %entry -; CHECK-MVE-NEXT: vcmp.f32 s4, s3 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, s2 -; CHECK-MVE-NEXT: vseleq.f32 s3, s15, s11 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s4, s1 ; CHECK-MVE-NEXT: vseleq.f32 s2, s14, s10 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s4, s0 +; CHECK-MVE-NEXT: vcmp.f32 s4, s3 ; CHECK-MVE-NEXT: vseleq.f32 s1, s13, s9 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vcmp.f32 s4, s0 +; CHECK-MVE-NEXT: vseleq.f32 s3, s15, s11 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f32 s0, s12, s8 ; CHECK-MVE-NEXT: bx lr ; @@ -2718,42 +2710,42 @@ define arm_aapcs_vfpcc <8 x half> @vcmp_r_une_v8f16(<8 x half> %src, half %src2, ; CHECK-MVE-LABEL: vcmp_r_une_v8f16: ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vmovx.f16 s6, s0 -; CHECK-MVE-NEXT: vmovx.f16 s5, s8 +; CHECK-MVE-NEXT: vmovx.f16 s5, s12 ; CHECK-MVE-NEXT: vcmp.f16 s4, s6 -; CHECK-MVE-NEXT: vmovx.f16 s7, s12 +; CHECK-MVE-NEXT: vmovx.f16 s6, s8 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s4, s0 -; CHECK-MVE-NEXT: vseleq.f16 s6, s7, s5 +; CHECK-MVE-NEXT: vseleq.f16 s6, s5, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s0, s12, s8 -; CHECK-MVE-NEXT: vmovx.f16 s8, s9 +; CHECK-MVE-NEXT: vmovx.f16 s8, s13 ; CHECK-MVE-NEXT: vins.f16 s0, s6 ; CHECK-MVE-NEXT: vmovx.f16 s6, s1 ; CHECK-MVE-NEXT: vcmp.f16 s4, s6 -; CHECK-MVE-NEXT: vmovx.f16 s12, s13 +; CHECK-MVE-NEXT: vmovx.f16 s6, s9 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s4, s1 -; CHECK-MVE-NEXT: vseleq.f16 s6, s12, s8 +; CHECK-MVE-NEXT: vseleq.f16 s6, s8, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s8, s10 -; CHECK-MVE-NEXT: vmovx.f16 s12, s14 +; CHECK-MVE-NEXT: vmovx.f16 s8, s14 ; CHECK-MVE-NEXT: vseleq.f16 s1, s13, s9 ; CHECK-MVE-NEXT: vins.f16 s1, s6 ; CHECK-MVE-NEXT: vmovx.f16 s6, s2 ; CHECK-MVE-NEXT: vcmp.f16 s4, s6 +; CHECK-MVE-NEXT: vmovx.f16 s6, s10 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s4, s2 -; CHECK-MVE-NEXT: vseleq.f16 s6, s12, s8 +; CHECK-MVE-NEXT: vseleq.f16 s6, s8, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s8, s11 +; CHECK-MVE-NEXT: vmovx.f16 s8, s15 ; CHECK-MVE-NEXT: vseleq.f16 s2, s14, s10 -; CHECK-MVE-NEXT: vmovx.f16 s10, s15 ; CHECK-MVE-NEXT: vins.f16 s2, s6 ; CHECK-MVE-NEXT: vmovx.f16 s6, s3 ; CHECK-MVE-NEXT: vcmp.f16 s4, s6 +; CHECK-MVE-NEXT: vmovx.f16 s6, s11 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s4, s3 -; CHECK-MVE-NEXT: vseleq.f16 s6, s10, s8 +; CHECK-MVE-NEXT: vseleq.f16 s6, s8, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s3, s15, s11 ; CHECK-MVE-NEXT: vins.f16 s3, s6 diff --git a/llvm/test/CodeGen/Thumb2/mve-vcmpfz.ll b/llvm/test/CodeGen/Thumb2/mve-vcmpfz.ll index 16689f1e7ecd1..718657839d38d 100644 --- a/llvm/test/CodeGen/Thumb2/mve-vcmpfz.ll +++ b/llvm/test/CodeGen/Thumb2/mve-vcmpfz.ll @@ -43,10 +43,8 @@ define arm_aapcs_vfpcc <4 x float> @vcmp_one_v4f32(<4 x float> %src, <4 x float> ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vcmp.f32 s1, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s1, #0 -; CHECK-MVE-NEXT: cset r0, mi -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s0, #0 +; CHECK-MVE-NEXT: cset r0, mi ; CHECK-MVE-NEXT: csinc r0, r0, zr, le ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s3, #0 @@ -230,10 +228,8 @@ define arm_aapcs_vfpcc <4 x float> @vcmp_ueq_v4f32(<4 x float> %src, <4 x float> ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vcmp.f32 s1, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s1, #0 -; CHECK-MVE-NEXT: cset r0, eq -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s0, #0 +; CHECK-MVE-NEXT: cset r0, eq ; CHECK-MVE-NEXT: csinc r0, r0, zr, vc ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s3, #0 @@ -271,17 +267,17 @@ entry: define arm_aapcs_vfpcc <4 x float> @vcmp_une_v4f32(<4 x float> %src, <4 x float> %a, <4 x float> %b) { ; CHECK-MVE-LABEL: vcmp_une_v4f32: ; CHECK-MVE: @ %bb.0: @ %entry -; CHECK-MVE-NEXT: vcmp.f32 s3, #0 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s2, #0 -; CHECK-MVE-NEXT: vseleq.f32 s3, s11, s7 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s1, #0 ; CHECK-MVE-NEXT: vseleq.f32 s2, s10, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s0, #0 +; CHECK-MVE-NEXT: vcmp.f32 s3, #0 ; CHECK-MVE-NEXT: vseleq.f32 s1, s9, s5 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vcmp.f32 s0, #0 +; CHECK-MVE-NEXT: vseleq.f32 s3, s11, s7 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f32 s0, s8, s4 ; CHECK-MVE-NEXT: bx lr ; @@ -1042,42 +1038,42 @@ define arm_aapcs_vfpcc <8 x half> @vcmp_une_v8f16(<8 x half> %src, <8 x half> %a ; CHECK-MVE-LABEL: vcmp_une_v8f16: ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vmovx.f16 s12, s0 -; CHECK-MVE-NEXT: vmovx.f16 s14, s4 +; CHECK-MVE-NEXT: vmovx.f16 s14, s8 ; CHECK-MVE-NEXT: vcmp.f16 s12, #0 -; CHECK-MVE-NEXT: vmovx.f16 s13, s8 +; CHECK-MVE-NEXT: vmovx.f16 s12, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s0, #0 -; CHECK-MVE-NEXT: vseleq.f16 s12, s13, s14 +; CHECK-MVE-NEXT: vseleq.f16 s12, s14, s12 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s0, s8, s4 ; CHECK-MVE-NEXT: vmovx.f16 s4, s1 ; CHECK-MVE-NEXT: vcmp.f16 s4, #0 -; CHECK-MVE-NEXT: vins.f16 s0, s12 +; CHECK-MVE-NEXT: vmovx.f16 s4, s5 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s8, s5 -; CHECK-MVE-NEXT: vmovx.f16 s12, s9 +; CHECK-MVE-NEXT: vmovx.f16 s8, s9 ; CHECK-MVE-NEXT: vcmp.f16 s1, #0 -; CHECK-MVE-NEXT: vseleq.f16 s4, s12, s8 +; CHECK-MVE-NEXT: vins.f16 s0, s12 +; CHECK-MVE-NEXT: vseleq.f16 s4, s8, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s8, s6 -; CHECK-MVE-NEXT: vmovx.f16 s12, s10 +; CHECK-MVE-NEXT: vmovx.f16 s8, s10 ; CHECK-MVE-NEXT: vseleq.f16 s1, s9, s5 ; CHECK-MVE-NEXT: vins.f16 s1, s4 ; CHECK-MVE-NEXT: vmovx.f16 s4, s2 ; CHECK-MVE-NEXT: vcmp.f16 s4, #0 +; CHECK-MVE-NEXT: vmovx.f16 s4, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s2, #0 -; CHECK-MVE-NEXT: vseleq.f16 s4, s12, s8 +; CHECK-MVE-NEXT: vseleq.f16 s4, s8, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s8, s11 ; CHECK-MVE-NEXT: vseleq.f16 s2, s10, s6 -; CHECK-MVE-NEXT: vmovx.f16 s6, s7 +; CHECK-MVE-NEXT: vmovx.f16 s6, s11 ; CHECK-MVE-NEXT: vins.f16 s2, s4 ; CHECK-MVE-NEXT: vmovx.f16 s4, s3 ; CHECK-MVE-NEXT: vcmp.f16 s4, #0 +; CHECK-MVE-NEXT: vmovx.f16 s4, s7 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s3, #0 -; CHECK-MVE-NEXT: vseleq.f16 s4, s8, s6 +; CHECK-MVE-NEXT: vseleq.f16 s4, s6, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s3, s11, s7 ; CHECK-MVE-NEXT: vins.f16 s3, s4 @@ -1572,10 +1568,8 @@ define arm_aapcs_vfpcc <4 x float> @vcmp_r_one_v4f32(<4 x float> %src, <4 x floa ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vcmp.f32 s1, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s1, #0 -; CHECK-MVE-NEXT: cset r0, mi -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s0, #0 +; CHECK-MVE-NEXT: cset r0, mi ; CHECK-MVE-NEXT: csinc r0, r0, zr, le ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s3, #0 @@ -1759,10 +1753,8 @@ define arm_aapcs_vfpcc <4 x float> @vcmp_r_ueq_v4f32(<4 x float> %src, <4 x floa ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vcmp.f32 s1, #0 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s1, #0 -; CHECK-MVE-NEXT: cset r0, eq -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s0, #0 +; CHECK-MVE-NEXT: cset r0, eq ; CHECK-MVE-NEXT: csinc r0, r0, zr, vc ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s3, #0 @@ -1800,17 +1792,17 @@ entry: define arm_aapcs_vfpcc <4 x float> @vcmp_r_une_v4f32(<4 x float> %src, <4 x float> %a, <4 x float> %b) { ; CHECK-MVE-LABEL: vcmp_r_une_v4f32: ; CHECK-MVE: @ %bb.0: @ %entry -; CHECK-MVE-NEXT: vcmp.f32 s3, #0 -; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s2, #0 -; CHECK-MVE-NEXT: vseleq.f32 s3, s11, s7 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f32 s1, #0 ; CHECK-MVE-NEXT: vseleq.f32 s2, s10, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vcmp.f32 s0, #0 +; CHECK-MVE-NEXT: vcmp.f32 s3, #0 ; CHECK-MVE-NEXT: vseleq.f32 s1, s9, s5 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr +; CHECK-MVE-NEXT: vcmp.f32 s0, #0 +; CHECK-MVE-NEXT: vseleq.f32 s3, s11, s7 +; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f32 s0, s8, s4 ; CHECK-MVE-NEXT: bx lr ; @@ -2571,42 +2563,42 @@ define arm_aapcs_vfpcc <8 x half> @vcmp_r_une_v8f16(<8 x half> %src, <8 x half> ; CHECK-MVE-LABEL: vcmp_r_une_v8f16: ; CHECK-MVE: @ %bb.0: @ %entry ; CHECK-MVE-NEXT: vmovx.f16 s12, s0 -; CHECK-MVE-NEXT: vmovx.f16 s14, s4 +; CHECK-MVE-NEXT: vmovx.f16 s14, s8 ; CHECK-MVE-NEXT: vcmp.f16 s12, #0 -; CHECK-MVE-NEXT: vmovx.f16 s13, s8 +; CHECK-MVE-NEXT: vmovx.f16 s12, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s0, #0 -; CHECK-MVE-NEXT: vseleq.f16 s12, s13, s14 +; CHECK-MVE-NEXT: vseleq.f16 s12, s14, s12 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s0, s8, s4 ; CHECK-MVE-NEXT: vmovx.f16 s4, s1 ; CHECK-MVE-NEXT: vcmp.f16 s4, #0 -; CHECK-MVE-NEXT: vins.f16 s0, s12 +; CHECK-MVE-NEXT: vmovx.f16 s4, s5 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s8, s5 -; CHECK-MVE-NEXT: vmovx.f16 s12, s9 +; CHECK-MVE-NEXT: vmovx.f16 s8, s9 ; CHECK-MVE-NEXT: vcmp.f16 s1, #0 -; CHECK-MVE-NEXT: vseleq.f16 s4, s12, s8 +; CHECK-MVE-NEXT: vins.f16 s0, s12 +; CHECK-MVE-NEXT: vseleq.f16 s4, s8, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s8, s6 -; CHECK-MVE-NEXT: vmovx.f16 s12, s10 +; CHECK-MVE-NEXT: vmovx.f16 s8, s10 ; CHECK-MVE-NEXT: vseleq.f16 s1, s9, s5 ; CHECK-MVE-NEXT: vins.f16 s1, s4 ; CHECK-MVE-NEXT: vmovx.f16 s4, s2 ; CHECK-MVE-NEXT: vcmp.f16 s4, #0 +; CHECK-MVE-NEXT: vmovx.f16 s4, s6 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s2, #0 -; CHECK-MVE-NEXT: vseleq.f16 s4, s12, s8 +; CHECK-MVE-NEXT: vseleq.f16 s4, s8, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr -; CHECK-MVE-NEXT: vmovx.f16 s8, s11 ; CHECK-MVE-NEXT: vseleq.f16 s2, s10, s6 -; CHECK-MVE-NEXT: vmovx.f16 s6, s7 +; CHECK-MVE-NEXT: vmovx.f16 s6, s11 ; CHECK-MVE-NEXT: vins.f16 s2, s4 ; CHECK-MVE-NEXT: vmovx.f16 s4, s3 ; CHECK-MVE-NEXT: vcmp.f16 s4, #0 +; CHECK-MVE-NEXT: vmovx.f16 s4, s7 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vcmp.f16 s3, #0 -; CHECK-MVE-NEXT: vseleq.f16 s4, s8, s6 +; CHECK-MVE-NEXT: vseleq.f16 s4, s6, s4 ; CHECK-MVE-NEXT: vmrs APSR_nzcv, fpscr ; CHECK-MVE-NEXT: vseleq.f16 s3, s11, s7 ; CHECK-MVE-NEXT: vins.f16 s3, s4