Skip to content

[DirectX] Make firstbithigh intrinsic use first arg as overload type #145401

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 2 commits into from
Jun 24, 2025
Merged

[DirectX] Make firstbithigh intrinsic use first arg as overload type #145401

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
19552c5
[DirectX] Make firstbithigh intrinsic use first arg as overload type
V-FEXrt Jun 23, 2025
6114727
Address comments
V-FEXrt Jun 23, 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
2 changes: 2 additions & 0 deletions llvm/lib/Target/DirectX/DirectXTargetTransformInfo.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,6 +31,8 @@ bool DirectXTTIImpl::isTargetIntrinsicWithOverloadTypeAtArg(Intrinsic::ID ID,
case Intrinsic::dx_asdouble:
case Intrinsic::dx_isinf:
case Intrinsic::dx_firstbitlow:
case Intrinsic::dx_firstbituhigh:
case Intrinsic::dx_firstbitshigh:
return OpdIdx == 0;
default:
return OpdIdx == -1;
Expand Down
92 changes: 60 additions & 32 deletions llvm/test/CodeGen/DirectX/firstbithigh.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,90 +4,118 @@

define noundef i32 @test_firstbithigh_ushort(i16 noundef %a) {
entry:
; CHECK: call i32 @dx.op.unaryBits.i16(i32 33, i16 %{{.*}}) #[[#ATTR:]]
; CHECK: call i32 @dx.op.unaryBits.i16(i32 33, i16 %{{.*}})
%elt.firstbithigh = call i32 @llvm.dx.firstbituhigh.i16(i16 %a)
ret i32 %elt.firstbithigh
}

define noundef i32 @test_firstbithigh_short(i16 noundef %a) {
entry:
; CHECK: call i32 @dx.op.unaryBits.i16(i32 34, i16 %{{.*}}) #[[#ATTR]]
; CHECK: call i32 @dx.op.unaryBits.i16(i32 34, i16 %{{.*}})
%elt.firstbithigh = call i32 @llvm.dx.firstbitshigh.i16(i16 %a)
ret i32 %elt.firstbithigh
}

define noundef i32 @test_firstbithigh_uint(i32 noundef %a) {
entry:
; CHECK: call i32 @dx.op.unaryBits.i32(i32 33, i32 %{{.*}}) #[[#ATTR]]
; CHECK: call i32 @dx.op.unaryBits.i32(i32 33, i32 %{{.*}})
%elt.firstbithigh = call i32 @llvm.dx.firstbituhigh.i32(i32 %a)
ret i32 %elt.firstbithigh
}

define noundef i32 @test_firstbithigh_int(i32 noundef %a) {
entry:
; CHECK: call i32 @dx.op.unaryBits.i32(i32 34, i32 %{{.*}}) #[[#ATTR]]
; CHECK: call i32 @dx.op.unaryBits.i32(i32 34, i32 %{{.*}})
%elt.firstbithigh = call i32 @llvm.dx.firstbitshigh.i32(i32 %a)
ret i32 %elt.firstbithigh
}

define noundef i32 @test_firstbithigh_ulong(i64 noundef %a) {
entry:
; CHECK: call i32 @dx.op.unaryBits.i64(i32 33, i64 %{{.*}}) #[[#ATTR]]
; CHECK: call i32 @dx.op.unaryBits.i64(i32 33, i64 %{{.*}})
%elt.firstbithigh = call i32 @llvm.dx.firstbituhigh.i64(i64 %a)
ret i32 %elt.firstbithigh
}

define noundef i32 @test_firstbithigh_long(i64 noundef %a) {
entry:
; CHECK: call i32 @dx.op.unaryBits.i64(i32 34, i64 %{{.*}}) #[[#ATTR]]
; CHECK: call i32 @dx.op.unaryBits.i64(i32 34, i64 %{{.*}})
%elt.firstbithigh = call i32 @llvm.dx.firstbitshigh.i64(i64 %a)
ret i32 %elt.firstbithigh
}

define noundef <4 x i32> @test_firstbituhigh_vec4_i32(<4 x i32> noundef %a) {
entry:
; CHECK: [[ee0:%.*]] = extractelement <4 x i32> %a, i64 0
; CHECK: [[ie0:%.*]] = call i32 @dx.op.unaryBits.i32(i32 33, i32 [[ee0]]) #[[#ATTR]]
; CHECK: [[ie0:%.*]] = call i32 @dx.op.unaryBits.i32(i32 33, i32 [[ee0]])
; CHECK: [[ee1:%.*]] = extractelement <4 x i32> %a, i64 1
; CHECK: [[ie1:%.*]] = call i32 @dx.op.unaryBits.i32(i32 33, i32 [[ee1]]) #[[#ATTR]]
; CHECK: [[ie1:%.*]] = call i32 @dx.op.unaryBits.i32(i32 33, i32 [[ee1]])
; CHECK: [[ee2:%.*]] = extractelement <4 x i32> %a, i64 2
; CHECK: [[ie2:%.*]] = call i32 @dx.op.unaryBits.i32(i32 33, i32 [[ee2]]) #[[#ATTR]]
; CHECK: [[ie2:%.*]] = call i32 @dx.op.unaryBits.i32(i32 33, i32 [[ee2]])
; CHECK: [[ee3:%.*]] = extractelement <4 x i32> %a, i64 3
; CHECK: [[ie3:%.*]] = call i32 @dx.op.unaryBits.i32(i32 33, i32 [[ee3]]) #[[#ATTR]]
; CHECK: insertelement <4 x i32> poison, i32 [[ie0]], i64 0
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie1]], i64 1
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie2]], i64 2
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie3]], i64 3
; CHECK: [[ie3:%.*]] = call i32 @dx.op.unaryBits.i32(i32 33, i32 [[ee3]])
; CHECK: [[rt0:%.*]] = insertelement <4 x i32> poison, i32 [[ie0]], i64 0
; CHECK: [[rt1:%.*]] = insertelement <4 x i32> [[rt0]], i32 [[ie1]], i64 1
; CHECK: [[rt2:%.*]] = insertelement <4 x i32> [[rt1]], i32 [[ie2]], i64 2
; CHECK: [[rt3:%.*]] = insertelement <4 x i32> [[rt2]], i32 [[ie3]], i64 3
; CHECK: ret <4 x i32> [[rt3]]
%2 = call <4 x i32> @llvm.dx.firstbituhigh.v4i32(<4 x i32> %a)
ret <4 x i32> %2
}

define noundef <4 x i32> @test_firstbitshigh_vec4_i32(<4 x i32> noundef %a) {
entry:
; CHECK: [[ee0:%.*]] = extractelement <4 x i32> %a, i64 0
; CHECK: [[ie0:%.*]] = call i32 @dx.op.unaryBits.i32(i32 34, i32 [[ee0]]) #[[#ATTR]]
; CHECK: [[ie0:%.*]] = call i32 @dx.op.unaryBits.i32(i32 34, i32 [[ee0]])
; CHECK: [[ee1:%.*]] = extractelement <4 x i32> %a, i64 1
; CHECK: [[ie1:%.*]] = call i32 @dx.op.unaryBits.i32(i32 34, i32 [[ee1]]) #[[#ATTR]]
; CHECK: [[ie1:%.*]] = call i32 @dx.op.unaryBits.i32(i32 34, i32 [[ee1]])
; CHECK: [[ee2:%.*]] = extractelement <4 x i32> %a, i64 2
; CHECK: [[ie2:%.*]] = call i32 @dx.op.unaryBits.i32(i32 34, i32 [[ee2]]) #[[#ATTR]]
; CHECK: [[ie2:%.*]] = call i32 @dx.op.unaryBits.i32(i32 34, i32 [[ee2]])
; CHECK: [[ee3:%.*]] = extractelement <4 x i32> %a, i64 3
; CHECK: [[ie3:%.*]] = call i32 @dx.op.unaryBits.i32(i32 34, i32 [[ee3]]) #[[#ATTR]]
; CHECK: insertelement <4 x i32> poison, i32 [[ie0]], i64 0
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie1]], i64 1
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie2]], i64 2
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie3]], i64 3
; CHECK: [[ie3:%.*]] = call i32 @dx.op.unaryBits.i32(i32 34, i32 [[ee3]])
; CHECK: [[rt0:%.*]] = insertelement <4 x i32> poison, i32 [[ie0]], i64 0
; CHECK: [[rt1:%.*]] = insertelement <4 x i32> [[rt0]], i32 [[ie1]], i64 1
; CHECK: [[rt2:%.*]] = insertelement <4 x i32> [[rt1]], i32 [[ie2]], i64 2
; CHECK: [[rt3:%.*]] = insertelement <4 x i32> [[rt2]], i32 [[ie3]], i64 3
; CHECK: ret <4 x i32> [[rt3]]
%2 = call <4 x i32> @llvm.dx.firstbitshigh.v4i32(<4 x i32> %a)
ret <4 x i32> %2
}

; CHECK: attributes #[[#ATTR]] = {{{.*}} memory(none) {{.*}}}

declare i32 @llvm.dx.firstbituhigh.i16(i16)
declare i32 @llvm.dx.firstbituhigh.i32(i32)
declare i32 @llvm.dx.firstbituhigh.i64(i64)
declare <4 x i32> @llvm.dx.firstbituhigh.v4i32(<4 x i32>)
define noundef <4 x i32> @test_firstbituhigh_vec4_i16(<4 x i16> noundef %a) {
entry:
; CHECK: [[ee0:%.*]] = extractelement <4 x i16> %a, i64 0
; CHECK: [[ie0:%.*]] = call i32 @dx.op.unaryBits.i16(i32 33, i16 [[ee0]])
; CHECK: [[ee1:%.*]] = extractelement <4 x i16> %a, i64 1
; CHECK: [[ie1:%.*]] = call i32 @dx.op.unaryBits.i16(i32 33, i16 [[ee1]])
; CHECK: [[ee2:%.*]] = extractelement <4 x i16> %a, i64 2
; CHECK: [[ie2:%.*]] = call i32 @dx.op.unaryBits.i16(i32 33, i16 [[ee2]])
; CHECK: [[ee3:%.*]] = extractelement <4 x i16> %a, i64 3
; CHECK: [[ie3:%.*]] = call i32 @dx.op.unaryBits.i16(i32 33, i16 [[ee3]])
; CHECK: [[rt0:%.*]] = insertelement <4 x i32> poison, i32 [[ie0]], i64 0
; CHECK: [[rt1:%.*]] = insertelement <4 x i32> [[rt0]], i32 [[ie1]], i64 1
; CHECK: [[rt2:%.*]] = insertelement <4 x i32> [[rt1]], i32 [[ie2]], i64 2
; CHECK: [[rt3:%.*]] = insertelement <4 x i32> [[rt2]], i32 [[ie3]], i64 3
; CHECK: ret <4 x i32> [[rt3]]
%2 = call <4 x i32> @llvm.dx.firstbituhigh.v4i16(<4 x i16> %a)
ret <4 x i32> %2
}

declare i32 @llvm.dx.firstbitshigh.i16(i16)
declare i32 @llvm.dx.firstbitshigh.i32(i32)
declare i32 @llvm.dx.firstbitshigh.i64(i64)
declare <4 x i32> @llvm.dx.firstbitshigh.v4i32(<4 x i32>)
define noundef <4 x i32> @test_firstbitshigh_vec4_i16(<4 x i16> noundef %a) {
entry:
; CHECK: [[ee0:%.*]] = extractelement <4 x i16> %a, i64 0
; CHECK: [[ie0:%.*]] = call i32 @dx.op.unaryBits.i16(i32 34, i16 [[ee0]])
; CHECK: [[ee1:%.*]] = extractelement <4 x i16> %a, i64 1
; CHECK: [[ie1:%.*]] = call i32 @dx.op.unaryBits.i16(i32 34, i16 [[ee1]])
; CHECK: [[ee2:%.*]] = extractelement <4 x i16> %a, i64 2
; CHECK: [[ie2:%.*]] = call i32 @dx.op.unaryBits.i16(i32 34, i16 [[ee2]])
; CHECK: [[ee3:%.*]] = extractelement <4 x i16> %a, i64 3
; CHECK: [[ie3:%.*]] = call i32 @dx.op.unaryBits.i16(i32 34, i16 [[ee3]])
; CHECK: [[rt0:%.*]] = insertelement <4 x i32> poison, i32 [[ie0]], i64 0
; CHECK: [[rt1:%.*]] = insertelement <4 x i32> [[rt0]], i32 [[ie1]], i64 1
; CHECK: [[rt2:%.*]] = insertelement <4 x i32> [[rt1]], i32 [[ie2]], i64 2
; CHECK: [[rt3:%.*]] = insertelement <4 x i32> [[rt2]], i32 [[ie3]], i64 3
; CHECK: ret <4 x i32> [[rt3]]
%2 = call <4 x i32> @llvm.dx.firstbitshigh.v4i16(<4 x i16> %a)
ret <4 x i32> %2
}
18 changes: 10 additions & 8 deletions llvm/test/CodeGen/DirectX/firstbitlow.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,10 +33,11 @@ entry:
; CHECK: [[ie2:%.*]] = call i32 @dx.op.unaryBits.i32(i32 32, i32 [[ee2]])
; CHECK: [[ee3:%.*]] = extractelement <4 x i32> %a, i64 3
; CHECK: [[ie3:%.*]] = call i32 @dx.op.unaryBits.i32(i32 32, i32 [[ee3]])
; CHECK: insertelement <4 x i32> poison, i32 [[ie0]], i64 0
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie1]], i64 1
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie2]], i64 2
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie3]], i64 3
; CHECK: [[rt0:%.*]] = insertelement <4 x i32> poison, i32 [[ie0]], i64 0
; CHECK: [[rt1:%.*]] = insertelement <4 x i32> [[rt0]], i32 [[ie1]], i64 1
; CHECK: [[rt2:%.*]] = insertelement <4 x i32> [[rt1]], i32 [[ie2]], i64 2
; CHECK: [[rt3:%.*]] = insertelement <4 x i32> [[rt2]], i32 [[ie3]], i64 3
; CHECK: ret <4 x i32> [[rt3]]
%2 = call <4 x i32> @llvm.dx.firstbitlow.v4i32(<4 x i32> %a)
ret <4 x i32> %2
}
Expand All @@ -51,10 +52,11 @@ entry:
; CHECK: [[ie2:%.*]] = call i32 @dx.op.unaryBits.i16(i32 32, i16 [[ee2]])
; CHECK: [[ee3:%.*]] = extractelement <4 x i16> %a, i64 3
; CHECK: [[ie3:%.*]] = call i32 @dx.op.unaryBits.i16(i32 32, i16 [[ee3]])
; CHECK: insertelement <4 x i32> poison, i32 [[ie0]], i64 0
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie1]], i64 1
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie2]], i64 2
; CHECK: insertelement <4 x i32> %{{.*}}, i32 [[ie3]], i64 3
; CHECK: [[rt0:%.*]] = insertelement <4 x i32> poison, i32 [[ie0]], i64 0
; CHECK: [[rt1:%.*]] = insertelement <4 x i32> [[rt0]], i32 [[ie1]], i64 1
; CHECK: [[rt2:%.*]] = insertelement <4 x i32> [[rt1]], i32 [[ie2]], i64 2
; CHECK: [[rt3:%.*]] = insertelement <4 x i32> [[rt2]], i32 [[ie3]], i64 3
; CHECK: ret <4 x i32> [[rt3]]
%2 = call <4 x i32> @llvm.dx.firstbitlow.v4i16(<4 x i16> %a)
ret <4 x i32> %2
}