Skip to content

[Hexagon] Support lowering of setuo & seto for vector types in Hexagon #158740

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 6 commits into from
Oct 3, 2025
Merged

[Hexagon] Support lowering of setuo & seto for vector types in Hexagon #158740

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
6588982
[Hexagon] Support lowering of setuo & seto for vector types in Hexagon
kaushik-quicinc May 22, 2025
deeed94
Add check for setO
fhossein-quic Sep 16, 2025
6cd424a
add checks for second var as an operand
fhossein-quic Oct 2, 2025
5957719
clang-format
fhossein-quic Oct 2, 2025
381fb2f
Merge branch 'main' into upstream_setuo_seto
fhossein-quic Oct 3, 2025
1298609
Merge branch 'main' into upstream_setuo_seto
fhossein-quic Oct 3, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions llvm/lib/Target/Hexagon/HexagonISelLoweringHVX.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -355,6 +355,8 @@ HexagonTargetLowering::initializeHVXLowering() {
setCondCodeAction(ISD::SETULE, MVT::v64f16, Expand);
setCondCodeAction(ISD::SETUGE, MVT::v64f16, Expand);
setCondCodeAction(ISD::SETULT, MVT::v64f16, Expand);
setCondCodeAction(ISD::SETUO, MVT::v64f16, Expand);
setCondCodeAction(ISD::SETO, MVT::v64f16, Expand);

setCondCodeAction(ISD::SETNE, MVT::v32f32, Expand);
setCondCodeAction(ISD::SETLE, MVT::v32f32, Expand);
Expand All @@ -368,6 +370,8 @@ HexagonTargetLowering::initializeHVXLowering() {
setCondCodeAction(ISD::SETULE, MVT::v32f32, Expand);
setCondCodeAction(ISD::SETUGE, MVT::v32f32, Expand);
setCondCodeAction(ISD::SETULT, MVT::v32f32, Expand);
setCondCodeAction(ISD::SETUO, MVT::v32f32, Expand);
setCondCodeAction(ISD::SETO, MVT::v32f32, Expand);

// Boolean vectors.

Expand Down
93 changes: 93 additions & 0 deletions llvm/test/CodeGen/Hexagon/inst_setcc_uno_uo.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,93 @@
;; RUN: llc --mtriple=hexagon -mattr=+hvxv79,+hvx-length128b %s -o - | FileCheck %s

define dso_local void @store_isnan_f32(ptr %a, ptr %b, ptr %isnan_cmp) local_unnamed_addr {
entry:
%arrayidx_a = getelementptr inbounds nuw float, ptr %a, i32 0
%arrayidx_b = getelementptr inbounds nuw float, ptr %b, i32 0
%0 = load <32 x float>, ptr %arrayidx_a, align 4
%1 = load <32 x float>, ptr %arrayidx_b, align 4
%.vectorized = fcmp uno <32 x float> %0, %1
%.LS.instance = zext <32 x i1> %.vectorized to <32 x i32>
%arrayidx1 = getelementptr inbounds nuw i32, ptr %isnan_cmp, i32 0
store <32 x i32> %.LS.instance, ptr %arrayidx1, align 4
ret void
}

; CHECK: store_isnan_f32
; CHECK: [[RONE32:r[0-9]+]] = #1
; CHECK: [[VOP2_F32:v[0-9]+]] = vxor([[VOP2_F32]],[[VOP2_F32]])
; CHECK: [[VOP1_F32:v[0-9]+]] = vmemu(r0+#0)
; CHECK: [[VONES32:v[0-9]+]] = vsplat([[RONE32]])
; CHECK: [[Q1_F32:q[0-9]+]] = vcmp.eq([[VOP1_F32]].w,[[VOP1_F32]].w)
; CHECK: [[VOP3_F32:v[0-9]+]] = vmemu(r1+#0)
; CHECK: [[Q1_F32]] &= vcmp.eq([[VOP3_F32]].w,[[VOP3_F32]].w)
; CHECK: [[VOUT_F32:v[0-9]+]] = vmux([[Q1_F32]],[[VOP2_F32]],[[VONES32]])
; CHECK: vmemu(r2+#0) = [[VOUT_F32]]

define dso_local void @store_isnan_f16(ptr %a, ptr %b, ptr %isnan_cmp) local_unnamed_addr {
entry:
%arrayidx_a = getelementptr inbounds nuw half, ptr %a, i32 0
%arrayidx_b = getelementptr inbounds nuw half, ptr %b, i32 0
%0 = load <64 x half>, ptr %arrayidx_a, align 2
%1 = load <64 x half>, ptr %arrayidx_b, align 2
%.vectorized = fcmp uno <64 x half> %0, %1
%conv.LS.instance = zext <64 x i1> %.vectorized to <64 x i16>
%arrayidx1 = getelementptr inbounds nuw i16, ptr %isnan_cmp, i32 0
store <64 x i16> %conv.LS.instance, ptr %arrayidx1, align 2
ret void
}
; CHECK-LABEL: store_isnan_f16
; CHECK: [[RONE16:r[0-9]+]] = #1
; CHECK: [[VOP2_F16:v[0-9]+]] = vxor([[VOP2_F16]],[[VOP2_F16]])
; CHECK: [[VOP1_F16:v[0-9]+]] = vmemu(r0+#0)
; CHECK: [[VONES16:v[0-9]+]].h = vsplat([[RONE16]])
; CHECK: [[Q1_F16:q[0-9]+]] = vcmp.eq([[VOP1_F16]].h,[[VOP1_F16]].h)
; CHECK: [[VOP3_F16:v[0-9]+]] = vmemu(r1+#0)
; CHECK: [[Q1_F16]] &= vcmp.eq([[VOP3_F16]].h,[[VOP3_F16]].h)
; CHECK: [[VOUT_F16:v[0-9]+]] = vmux([[Q1_F16]],[[VOP2_F16]],[[VONES16]])
; CHECK: vmemu(r2+#0) = [[VOUT_F32]]

define dso_local void @store_isordered_f32(ptr %a, ptr %b, ptr %isordered_cmp) local_unnamed_addr {
entry:
%arrayidx_a = getelementptr inbounds nuw float, ptr %a, i32 0
%arrayidx_b = getelementptr inbounds nuw float, ptr %b, i32 0
%0 = load <32 x float>, ptr %arrayidx_a, align 4
%1 = load <32 x float>, ptr %arrayidx_b, align 4
%.vectorized = fcmp ord <32 x float> %0, %1
%.LS.instance = zext <32 x i1> %.vectorized to <32 x i32>
%arrayidx1 = getelementptr inbounds nuw i32, ptr %isordered_cmp, i32 0
store <32 x i32> %.LS.instance, ptr %arrayidx1, align 4
ret void
}
; CHECK-LABEL: store_isordered_f32
; CHECK: [[VOP2_ORD_F32:v[0-9]+]] = vxor([[VOP2_ORD_F32]],[[VOP2_ORD_F32]])
; CHECK: [[VOP1_ORD_F32:v[0-9]+]] = vmemu(r0+#0)
; CHECK: [[VONES_ORD_F32:v[0-9]+]] = vsplat([[RONE32]])
; CHECK: [[Q1_ORD_F32:q[0-9]+]] = vcmp.eq([[VOP1_ORD_F32]].w,[[VOP1_ORD_F32]].w)
; CHECK: [[VOP3_ORD_F32:v[0-9]+]] = vmemu(r1+#0)
; CHECK: [[Q1_ORD_F32]] &= vcmp.eq([[VOP3_ORD_F32]].w,[[VOP3_ORD_F32]].w)
; CHECK: [[VOUT_ORD_F32:v[0-9]+]] = vmux([[Q1_ORD_F32]],[[VONES_ORD_F32]],[[VOP2_ORD_F32]])
; CHECK: vmemu(r2+#0) = [[VOUT_ORD_F32]]


define dso_local void @store_isordered_f16(ptr %a, ptr %b, ptr %isordered_cmp) local_unnamed_addr {
entry:
%arrayidx_a = getelementptr inbounds nuw half, ptr %a, i32 0
%arrayidx_b = getelementptr inbounds nuw half, ptr %b, i32 0
%0 = load <64 x half>, ptr %arrayidx_a, align 2
%1 = load <64 x half>, ptr %arrayidx_b, align 2
%.vectorized = fcmp ord <64 x half> %0, %1
%conv.LS.instance = zext <64 x i1> %.vectorized to <64 x i16>
%arrayidx1 = getelementptr inbounds nuw i16, ptr %isordered_cmp, i32 0
store <64 x i16> %conv.LS.instance, ptr %arrayidx1, align 2
ret void
}
; CHECK-LABEL: store_isordered_f16
; CHECK: [[VOP2_ORD_F16:v[0-9]+]] = vxor([[VOP2_ORD_F16]],[[VOP2_ORD_F16]])
; CHECK: [[VOP1_ORD_F16:v[0-9]+]] = vmemu(r0+#0)
; CHECK: [[VONES_ORD_F16:v[0-9]+]].h = vsplat([[RONE16]])
; CHECK: [[Q1_ORD_F16:q[0-9]+]] = vcmp.eq([[VOP1_ORD_F16]].h,[[VOP1_ORD_F16]].h)
; CHECK: [[VOP3_ORD_F16:v[0-9]+]] = vmemu(r1+#0)
; CHECK: [[Q1_ORD_F16]] &= vcmp.eq([[VOP3_ORD_F16]].h,[[VOP3_ORD_F16]].h)
; CHECK: [[VOUT_ORD_F16:v[0-9]+]] = vmux([[Q1_ORD_F16]],[[VONES_ORD_F16]],[[VOP2_ORD_F16]])
; CHECK: vmemu(r2+#0) = [[VOUT_ORD_F16]]
Loading