[LLVM][SVE] Add isel for bfloat based (de)interleave operations.

Author: paulwalker-arm

llvm/lib/Target/AArch64/AArch64ISelLowering.cpp

@@ -1702,6 +1702,8 @@ AArch64TargetLowering::AArch64TargetLowering(const TargetMachine &TM,
       setOperationAction(ISD::MLOAD, VT, Custom);
       setOperationAction(ISD::INSERT_SUBVECTOR, VT, Custom);
       setOperationAction(ISD::SPLAT_VECTOR, VT, Legal);
+      setOperationAction(ISD::VECTOR_DEINTERLEAVE, VT, Custom);
+      setOperationAction(ISD::VECTOR_INTERLEAVE, VT, Custom);
       setOperationAction(ISD::VECTOR_SPLICE, VT, Custom);
 
       if (Subtarget->hasSVEB16B16()) {

llvm/lib/Target/AArch64/SVEInstrFormats.td

@@ -3079,6 +3079,8 @@ multiclass sve_int_perm_bin_perm_zz<bits<3> opc, string asm,
   def : SVE_2_Op_Pat<nxv2f64, op, nxv2f64, nxv2f64, !cast<Instruction>(NAME # _D)>;
 
   def : SVE_2_Op_Pat<nxv8bf16, op, nxv8bf16, nxv8bf16, !cast<Instruction>(NAME # _H)>;
+  def : SVE_2_Op_Pat<nxv4bf16, op, nxv4bf16, nxv4bf16, !cast<Instruction>(NAME # _S)>;
+  def : SVE_2_Op_Pat<nxv2bf16, op, nxv2bf16, nxv2bf16, !cast<Instruction>(NAME # _D)>;
 }
 
 //===----------------------------------------------------------------------===//

llvm/test/CodeGen/AArch64/sve-vector-deinterleave.ll

@@ -70,6 +70,41 @@ define {<vscale x 2 x double>, <vscale x 2 x double>} @vector_deinterleave_nxv2f
   ret {<vscale x 2 x double>, <vscale x 2 x double>}   %retval
 }
 
+define {<vscale x 2 x bfloat>, <vscale x 2 x bfloat>} @vector_deinterleave_nxv2bf16_nxv4bf16(<vscale x 4 x bfloat> %vec) {
+; CHECK-LABEL: vector_deinterleave_nxv2bf16_nxv4bf16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    uzp1 z1.s, z0.s, z0.s
+; CHECK-NEXT:    uzp2 z2.s, z0.s, z0.s
+; CHECK-NEXT:    uunpklo z0.d, z1.s
+; CHECK-NEXT:    uunpklo z1.d, z2.s
+; CHECK-NEXT:    ret
+  %retval = call {<vscale x 2 x bfloat>, <vscale x 2 x bfloat>} @llvm.vector.deinterleave2.nxv4bf16(<vscale x 4 x bfloat> %vec)
+  ret {<vscale x 2 x bfloat>, <vscale x 2 x bfloat>}   %retval
+}
+
+define {<vscale x 4 x bfloat>, <vscale x 4 x bfloat>} @vector_deinterleave_nxv4bf16_nxv8bf16(<vscale x 8 x bfloat> %vec) {
+; CHECK-LABEL: vector_deinterleave_nxv4bf16_nxv8bf16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    uzp1 z1.h, z0.h, z0.h
+; CHECK-NEXT:    uzp2 z2.h, z0.h, z0.h
+; CHECK-NEXT:    uunpklo z0.s, z1.h
+; CHECK-NEXT:    uunpklo z1.s, z2.h
+; CHECK-NEXT:    ret
+  %retval = call {<vscale x 4 x bfloat>, <vscale x 4 x bfloat>} @llvm.vector.deinterleave2.nxv8bf16(<vscale x 8 x bfloat> %vec)
+  ret {<vscale x 4 x bfloat>, <vscale x 4 x bfloat>}   %retval
+}
+
+define {<vscale x 8 x bfloat>, <vscale x 8 x bfloat>} @vector_deinterleave_nxv8bf16_nxv16bf16(<vscale x 16 x bfloat> %vec) {
+; CHECK-LABEL: vector_deinterleave_nxv8bf16_nxv16bf16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    uzp1 z2.h, z0.h, z1.h
+; CHECK-NEXT:    uzp2 z1.h, z0.h, z1.h
+; CHECK-NEXT:    mov z0.d, z2.d
+; CHECK-NEXT:    ret
+  %retval = call {<vscale x 8 x bfloat>, <vscale x 8 x bfloat>} @llvm.vector.deinterleave2.nxv16bf16(<vscale x 16 x bfloat> %vec)
+  ret {<vscale x 8 x bfloat>, <vscale x 8 x bfloat>}   %retval
+}
+
 ; Integers
 
 define {<vscale x 16 x i8>, <vscale x 16 x i8>} @vector_deinterleave_nxv16i8_nxv32i8(<vscale x 32 x i8> %vec) {

llvm/test/CodeGen/AArch64/sve-vector-interleave.ll

@@ -67,6 +67,39 @@ define <vscale x 4 x double> @interleave2_nxv4f64(<vscale x 2 x double> %vec0, <
   ret  <vscale x 4 x double>   %retval
 }
 
+define <vscale x 4 x bfloat> @interleave2_nxv4bf16(<vscale x 2 x bfloat> %vec0, <vscale x 2 x bfloat> %vec1) {
+; CHECK-LABEL: interleave2_nxv4bf16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    zip2 z2.d, z0.d, z1.d
+; CHECK-NEXT:    zip1 z0.d, z0.d, z1.d
+; CHECK-NEXT:    uzp1 z0.s, z0.s, z2.s
+; CHECK-NEXT:    ret
+  %retval = call <vscale x 4 x bfloat> @llvm.vector.interleave2.nxv4bf16(<vscale x 2 x bfloat> %vec0, <vscale x 2 x bfloat> %vec1)
+  ret  <vscale x 4 x bfloat>   %retval
+}
+
+define <vscale x 8 x bfloat> @interleave2_nxv8bf16(<vscale x 4 x bfloat> %vec0, <vscale x 4 x bfloat> %vec1) {
+; CHECK-LABEL: interleave2_nxv8bf16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    zip2 z2.s, z0.s, z1.s
+; CHECK-NEXT:    zip1 z0.s, z0.s, z1.s
+; CHECK-NEXT:    uzp1 z0.h, z0.h, z2.h
+; CHECK-NEXT:    ret
+  %retval = call <vscale x 8 x bfloat> @llvm.vector.interleave2.nxv8bf16(<vscale x 4 x bfloat> %vec0, <vscale x 4 x bfloat> %vec1)
+  ret  <vscale x 8 x bfloat>   %retval
+}
+
+define <vscale x 16 x bfloat> @interleave2_nxv16bf16(<vscale x 8 x bfloat> %vec0, <vscale x 8 x bfloat> %vec1) {
+; CHECK-LABEL: interleave2_nxv16bf16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    zip1 z2.h, z0.h, z1.h
+; CHECK-NEXT:    zip2 z1.h, z0.h, z1.h
+; CHECK-NEXT:    mov z0.d, z2.d
+; CHECK-NEXT:    ret
+  %retval = call <vscale x 16 x bfloat> @llvm.vector.interleave2.nxv16bf16(<vscale x 8 x bfloat> %vec0, <vscale x 8 x bfloat> %vec1)
+  ret  <vscale x 16 x bfloat>   %retval
+}
+
 ; Integers
 
 define <vscale x 32 x i8> @interleave2_nxv32i8(<vscale x 16 x i8> %vec0, <vscale x 16 x i8> %vec1) {