Skip to content

[NFC][AMDGPU] Re-enable two tests previously disabled due to missing upstream features#149568

Merged
shiltian merged 1 commit into
mainfrom
users/shiltian/enable-disabled-gfx1250-codegen-tests
Jul 18, 2025
Merged

[NFC][AMDGPU] Re-enable two tests previously disabled due to missing upstream features#149568
shiltian merged 1 commit into
mainfrom
users/shiltian/enable-disabled-gfx1250-codegen-tests

Conversation

@shiltian

@shiltian shiltian commented Jul 18, 2025

Copy link
Copy Markdown
Contributor

This PR re-enables two tests that were previously disabled because they depended on features not yet upstreamed.

…g upstream features

This PR re-enables three tests that were previously disabled because they depended on features not yet upstreamed.
@shiltian shiltian requested review from changpeng and rampitec July 18, 2025 19:23
@shiltian shiltian changed the title [NFC][AMDGPU] Re-enable three tests previously disabled due to missing upstream features [NFC][AMDGPU] Re-enable two tests previously disabled due to missing upstream features Jul 18, 2025

shiltian commented Jul 18, 2025

Copy link
Copy Markdown
Contributor Author

This stack of pull requests is managed by Graphite. Learn more about stacking.

@llvmbot

llvmbot commented Jul 18, 2025

Copy link
Copy Markdown
Member

@llvm/pr-subscribers-backend-amdgpu

Author: Shilei Tian (shiltian)

Changes

This PR re-enables two tests that were previously disabled because they depended on features not yet upstreamed.


Full diff: https://github.com/llvm/llvm-project/pull/149568.diff

3 Files Affected:

  • (modified) llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll (+1-2)
  • (modified) llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rsq.bf16.ll (+1-2)
  • (modified) llvm/test/CodeGen/AMDGPU/llvm.sqrt.bf16.ll (-2)
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll
index 3c49d0b9c01b1..199494d1c3473 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll
@@ -1,10 +1,9 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
-; xUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=+real-true16 < %s | FileCheck -check-prefix=SDAG-TRUE16 %s
+; RUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=+real-true16 < %s | FileCheck -check-prefix=SDAG-TRUE16 %s
 ; RUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=-real-true16 < %s | FileCheck -check-prefix=SDAG-FAKE16 %s
 ; xUN: llc -global-isel=1 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=+real-true16 < %s | FileCheck -check-prefix=GI-TRUE16 %s
 ; xUN: llc -global-isel=1 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=-real-true16 < %s | FileCheck -check-prefix=GI-FAKE16 %s
 
-; FIXME: t16 doesn't work at the moment because the store of s16 under t16 mode fails to select.
 ; FIXME: GlobalISel does not work with bf16
 
 declare bfloat @llvm.amdgcn.rcp.bf16(bfloat) #0
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rsq.bf16.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rsq.bf16.ll
index 0a8a90422d1f2..42d12fd0fb3c9 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rsq.bf16.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rsq.bf16.ll
@@ -1,10 +1,9 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
-; xUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=+real-true16 %s -o - | FileCheck -check-prefix=SDAG-REAL16 %s
+; RUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=+real-true16 %s -o - | FileCheck -check-prefix=SDAG-REAL16 %s
 ; RUN: llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=-real-true16 %s -o - | FileCheck -check-prefix=SDAG-FAKE16 %s
 ; xUN: llc -global-isel=1 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=+real-true16 %s -o - | FileCheck -check-prefix=GISEL-REAL16 %s
 ; xUN: llc -global-isel=1 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=-real-true16 %s -o - | FileCheck -check-prefix=GISEL-FAKE16 %s
 
-; FIXME: t16 doesn't work at the moment because the store of s16 under t16 mode fails to select.
 ; FIXME: GlobalISel does not work with bf16
 
 declare bfloat @llvm.amdgcn.rsq.bf16(bfloat) #0
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.sqrt.bf16.ll b/llvm/test/CodeGen/AMDGPU/llvm.sqrt.bf16.ll
index 47b2b68f05abc..dcf01f744945f 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.sqrt.bf16.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.sqrt.bf16.ll
@@ -2,8 +2,6 @@
 ; RUN: llc -amdgpu-scalarize-global-loads=false -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=+real-true16 -mattr=-flat-for-global < %s | FileCheck -check-prefixes=GFX12-TRUE16 %s
 ; RUN: llc -amdgpu-scalarize-global-loads=false -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=-real-true16 -mattr=-flat-for-global < %s | FileCheck -check-prefixes=GFX12-FAKE16 %s
 
-; FIXME: t16 doesn't work at the moment because the store of s16 under t16 mode fails to select.
-
 declare bfloat @llvm.sqrt.bf16(bfloat %a)
 declare <2 x bfloat> @llvm.sqrt.v2bf16(<2 x bfloat> %a)
 

@shiltian shiltian merged commit d46de86 into main Jul 18, 2025
11 checks passed
@shiltian shiltian deleted the users/shiltian/enable-disabled-gfx1250-codegen-tests branch July 18, 2025 21:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants