[AMDGPU] Set Convergent property for image.(getlod/sample*) intrinsics which uses WQM #122908
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…s which uses WQM This change adds IntrConvergent property to image.getlod intrinsic and to several image.sample intrinsics. All image.sample intrinsics apart from LOD(_L), Level 0(_LZ), Derivative(_D) will be marked as Convergent.
|
@llvm/pr-subscribers-backend-amdgpu Author: Mariusz Sikora (mariusz-sikora-at-amd) ChangesThis change adds IntrConvergent property to image.getlod intrinsic and to several image.sample intrinsics. All image.sample intrinsics apart from LOD(_L), Level 0(_LZ), Derivative(_D) will be marked as Convergent. Full diff: https://github.com/llvm/llvm-project/pull/122908.diff 2 Files Affected:
diff --git a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
index b930d6983e2251..b529642a558710 100644
--- a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
+++ b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
@@ -876,6 +876,8 @@ class AMDGPUSampleVariant<string ucmod, string lcmod, list<AMDGPUArg> extra_addr
// Name of the {lod} or {clamp} argument that is appended to the coordinates,
// if any.
string LodOrClamp = "";
+
+ bit UsesWQM = false;
}
// AMDGPUSampleVariants: all variants supported by IMAGE_SAMPLE
@@ -905,8 +907,9 @@ defset list<AMDGPUSampleVariant> AMDGPUSampleVariants = {
}
defset list<AMDGPUSampleVariant> AMDGPUSampleVariantsNoGradients = {
+ let UsesWQM = true in
defm AMDGPUSample : AMDGPUSampleHelper_Clamp<"", "", []>;
- let Bias = true in
+ let Bias = true, UsesWQM = true in
defm AMDGPUSample : AMDGPUSampleHelper_Clamp<
"_B", "_b", [AMDGPUArg<llvm_anyfloat_ty, "bias">]>;
let LodOrClamp = "lod" in
@@ -1172,7 +1175,8 @@ defset list<AMDGPUImageDimIntrinsic> AMDGPUImageDimIntrinsics = {
foreach dim = AMDGPUDims.NoMsaa in {
def !strconcat(NAME, "_", dim.Name) : AMDGPUImageDimIntrinsic<
AMDGPUDimSampleProfile<opmod, dim, sample>,
- !if(NoMem, [IntrNoMem], [IntrReadMem]),
+ !listconcat(!if(NoMem, [IntrNoMem], [IntrReadMem]),
+ !if(sample.UsesWQM, [IntrConvergent], [])),
!if(NoMem, [], [SDNPMemOperand])>;
}
}
@@ -1188,7 +1192,8 @@ defset list<AMDGPUImageDimIntrinsic> AMDGPUImageDimIntrinsics = {
foreach dim = AMDGPUDims.NoMsaa in {
def !strconcat(NAME, "_", dim.Name, "_nortn") : AMDGPUImageDimIntrinsic<
AMDGPUDimSampleNoReturnProfile<opmod, dim, sample>,
- [IntrWillReturn], [SDNPMemOperand]>;
+ !listconcat([IntrWillReturn], !if(sample.UsesWQM, [IntrConvergent], [])),
+ [SDNPMemOperand]>;
}
}
foreach sample = AMDGPUSampleVariants in {
diff --git a/llvm/test/CodeGen/AMDGPU/sink-image-sample.ll b/llvm/test/CodeGen/AMDGPU/sink-image-sample.ll
index e1273e1a4bcd08..8a3b544ac2b80f 100644
--- a/llvm/test/CodeGen/AMDGPU/sink-image-sample.ll
+++ b/llvm/test/CodeGen/AMDGPU/sink-image-sample.ll
@@ -11,7 +11,7 @@
define amdgpu_ps float @sinking_img_sample() {
main_body:
- %i = call <3 x float> @llvm.amdgcn.image.sample.2d.v3f32.f32(i32 7, float undef, float undef, <8 x i32> undef, <4 x i32> undef, i1 false, i32 0, i32 0)
+ %i = call <3 x float> @llvm.amdgcn.image.sample.l.2d.v3f32.f32(i32 7, float undef, float undef, float undef, <8 x i32> undef, <4 x i32> undef, i1 false, i32 0, i32 0)
br i1 undef, label %endif1, label %if1
if1: ; preds = %main_body
@@ -28,7 +28,7 @@ exit: ; preds = %endif1, %if1
ret float %i24
}
; Function Attrs: nounwind readonly willreturn
-declare <3 x float> @llvm.amdgcn.image.sample.2d.v3f32.f32(i32 immarg, float, float, <8 x i32>, <4 x i32>, i1 immarg, i32 immarg, i32 immarg) #3
+declare <3 x float> @llvm.amdgcn.image.sample.l.2d.v3f32.f32(i32 immarg, float, float, float, <8 x i32>, <4 x i32>, i1 immarg, i32 immarg, i32 immarg) #3
; Function Attrs: nofree nosync nounwind readnone speculatable willreturn
declare float @llvm.fma.f32(float, float, float) #2
|
|
@llvm/pr-subscribers-llvm-ir Author: Mariusz Sikora (mariusz-sikora-at-amd) ChangesThis change adds IntrConvergent property to image.getlod intrinsic and to several image.sample intrinsics. All image.sample intrinsics apart from LOD(_L), Level 0(_LZ), Derivative(_D) will be marked as Convergent. Full diff: https://github.com/llvm/llvm-project/pull/122908.diff 2 Files Affected:
diff --git a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
index b930d6983e2251..b529642a558710 100644
--- a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
+++ b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
@@ -876,6 +876,8 @@ class AMDGPUSampleVariant<string ucmod, string lcmod, list<AMDGPUArg> extra_addr
// Name of the {lod} or {clamp} argument that is appended to the coordinates,
// if any.
string LodOrClamp = "";
+
+ bit UsesWQM = false;
}
// AMDGPUSampleVariants: all variants supported by IMAGE_SAMPLE
@@ -905,8 +907,9 @@ defset list<AMDGPUSampleVariant> AMDGPUSampleVariants = {
}
defset list<AMDGPUSampleVariant> AMDGPUSampleVariantsNoGradients = {
+ let UsesWQM = true in
defm AMDGPUSample : AMDGPUSampleHelper_Clamp<"", "", []>;
- let Bias = true in
+ let Bias = true, UsesWQM = true in
defm AMDGPUSample : AMDGPUSampleHelper_Clamp<
"_B", "_b", [AMDGPUArg<llvm_anyfloat_ty, "bias">]>;
let LodOrClamp = "lod" in
@@ -1172,7 +1175,8 @@ defset list<AMDGPUImageDimIntrinsic> AMDGPUImageDimIntrinsics = {
foreach dim = AMDGPUDims.NoMsaa in {
def !strconcat(NAME, "_", dim.Name) : AMDGPUImageDimIntrinsic<
AMDGPUDimSampleProfile<opmod, dim, sample>,
- !if(NoMem, [IntrNoMem], [IntrReadMem]),
+ !listconcat(!if(NoMem, [IntrNoMem], [IntrReadMem]),
+ !if(sample.UsesWQM, [IntrConvergent], [])),
!if(NoMem, [], [SDNPMemOperand])>;
}
}
@@ -1188,7 +1192,8 @@ defset list<AMDGPUImageDimIntrinsic> AMDGPUImageDimIntrinsics = {
foreach dim = AMDGPUDims.NoMsaa in {
def !strconcat(NAME, "_", dim.Name, "_nortn") : AMDGPUImageDimIntrinsic<
AMDGPUDimSampleNoReturnProfile<opmod, dim, sample>,
- [IntrWillReturn], [SDNPMemOperand]>;
+ !listconcat([IntrWillReturn], !if(sample.UsesWQM, [IntrConvergent], [])),
+ [SDNPMemOperand]>;
}
}
foreach sample = AMDGPUSampleVariants in {
diff --git a/llvm/test/CodeGen/AMDGPU/sink-image-sample.ll b/llvm/test/CodeGen/AMDGPU/sink-image-sample.ll
index e1273e1a4bcd08..8a3b544ac2b80f 100644
--- a/llvm/test/CodeGen/AMDGPU/sink-image-sample.ll
+++ b/llvm/test/CodeGen/AMDGPU/sink-image-sample.ll
@@ -11,7 +11,7 @@
define amdgpu_ps float @sinking_img_sample() {
main_body:
- %i = call <3 x float> @llvm.amdgcn.image.sample.2d.v3f32.f32(i32 7, float undef, float undef, <8 x i32> undef, <4 x i32> undef, i1 false, i32 0, i32 0)
+ %i = call <3 x float> @llvm.amdgcn.image.sample.l.2d.v3f32.f32(i32 7, float undef, float undef, float undef, <8 x i32> undef, <4 x i32> undef, i1 false, i32 0, i32 0)
br i1 undef, label %endif1, label %if1
if1: ; preds = %main_body
@@ -28,7 +28,7 @@ exit: ; preds = %endif1, %if1
ret float %i24
}
; Function Attrs: nounwind readonly willreturn
-declare <3 x float> @llvm.amdgcn.image.sample.2d.v3f32.f32(i32 immarg, float, float, <8 x i32>, <4 x i32>, i1 immarg, i32 immarg, i32 immarg) #3
+declare <3 x float> @llvm.amdgcn.image.sample.l.2d.v3f32.f32(i32 immarg, float, float, float, <8 x i32>, <4 x i32>, i1 immarg, i32 immarg, i32 immarg) #3
; Function Attrs: nofree nosync nounwind readnone speculatable willreturn
declare float @llvm.fma.f32(float, float, float) #2
|
|
✅ With the latest revision this PR passed the undef deprecator. |
arsenm
reviewed
Jan 14, 2025
| @@ -11,7 +11,7 @@ | |||
|
|
|||
| define amdgpu_ps float @sinking_img_sample() { | |||
| main_body: | |||
| %i = call <3 x float> @llvm.amdgcn.image.sample.2d.v3f32.f32(i32 7, float undef, float undef, <8 x i32> undef, <4 x i32> undef, i1 false, i32 0, i32 0) | |||
| %i = call <3 x float> @llvm.amdgcn.image.sample.l.2d.v3f32.f32(i32 7, float undef, float undef, float undef, <8 x i32> undef, <4 x i32> undef, i1 false, i32 0, i32 0) | |||
There was a problem hiding this comment.
Add a new test instead of changing the operation in an old one? Presumably we should have tests for all the cases
There was a problem hiding this comment.
I added more tests to cover other cases of image.sample
arsenm
approved these changes
Jan 15, 2025
perlfu
reviewed
Jan 15, 2025
I will prepare new PR for image_gather4 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This change adds IntrConvergent property to image.getlod intrinsic and to several image.sample intrinsics. All image.sample intrinsics apart from LOD(_L), Level 0(_LZ), Derivative(_D) will be marked as Convergent.