[AMDGPU] Tests for unnecessary S_WAIT_XCNT insertion #145688
Merged
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
Member
|
@llvm/pr-subscribers-backend-amdgpu Author: Jay Foad (jayfoad) ChangesFull diff: https://github.com/llvm/llvm-project/pull/145688.diff 1 Files Affected:
diff --git a/llvm/test/CodeGen/AMDGPU/wait-xcnt.mir b/llvm/test/CodeGen/AMDGPU/wait-xcnt.mir
index 73b994ab2ab8c..884f14094695d 100644
--- a/llvm/test/CodeGen/AMDGPU/wait-xcnt.mir
+++ b/llvm/test/CodeGen/AMDGPU/wait-xcnt.mir
@@ -920,3 +920,45 @@ body: |
GLOBAL_STORE_DWORD $vgpr0_vgpr1, $vgpr3, 0, 0, implicit $exec
$vgpr2 = V_MOV_B32_e32 1, implicit $exec
...
+
+# TODO: Unnecessary wait before overwriting vgpr0.
+---
+name: overwrite_vgpr_after_smem
+tracksRegLiveness: true
+machineFunctionInfo:
+ isEntryFunction: true
+body: |
+ bb.0:
+ liveins: $vgpr0_vgpr1, $sgpr0_sgpr1
+ ; GCN-LABEL: name: overwrite_vgpr_after_smem
+ ; GCN: liveins: $vgpr0_vgpr1, $sgpr0_sgpr1
+ ; GCN-NEXT: {{ $}}
+ ; GCN-NEXT: $vgpr2 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec
+ ; GCN-NEXT: $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0
+ ; GCN-NEXT: S_WAIT_XCNT 0
+ ; GCN-NEXT: $vgpr0 = V_MOV_B32_e32 0, implicit $exec
+ $vgpr2 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec
+ $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0
+ $vgpr0 = V_MOV_B32_e32 0, implicit $exec
+...
+
+# TODO: Unnecessary wait before overwriting sgpr0.
+---
+name: overwrite_sgpr_after_vmem
+tracksRegLiveness: true
+machineFunctionInfo:
+ isEntryFunction: true
+body: |
+ bb.0:
+ liveins: $vgpr0_vgpr1, $sgpr0_sgpr1
+ ; GCN-LABEL: name: overwrite_sgpr_after_vmem
+ ; GCN: liveins: $vgpr0_vgpr1, $sgpr0_sgpr1
+ ; GCN-NEXT: {{ $}}
+ ; GCN-NEXT: $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0
+ ; GCN-NEXT: $vgpr2 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec
+ ; GCN-NEXT: S_WAIT_XCNT 0
+ ; GCN-NEXT: $sgpr0 = S_MOV_B32 0
+ $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0
+ $vgpr2 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec
+ $sgpr0 = S_MOV_B32 0
+...
|
shiltian
approved these changes
Jun 25, 2025
cdevadas
reviewed
Jun 25, 2025
| ; GCN-NEXT: {{ $}} | ||
| ; GCN-NEXT: $vgpr2 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec | ||
| ; GCN-NEXT: $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 | ||
| ; GCN-NEXT: S_WAIT_XCNT 0 |
Collaborator
There was a problem hiding this comment.
Why is it unnecessary? Because of the S_LOAD?
Contributor
There was a problem hiding this comment.
There is no dependence afterwards?
Contributor
Author
There was a problem hiding this comment.
Why is it unnecessary?
Yeah, I should have explained: hardware does an implicit "S_WAIT_XCNT 0" between SMEM and VMEM instructions, so you never have outstanding address translations for both SMEM and VMEM at the same time.
| ; GCN-NEXT: {{ $}} | ||
| ; GCN-NEXT: $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 | ||
| ; GCN-NEXT: $vgpr2 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec | ||
| ; GCN-NEXT: S_WAIT_XCNT 0 |
Collaborator
There was a problem hiding this comment.
Again, why is it not needed?
Collaborator
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/138/builds/16485 Here is the relevant piece of the build log for the reference |
mahesh-attarde
pushed a commit
to mahesh-attarde/llvm-project
that referenced
this pull request
Jul 28, 2025
Hardware does an implicit "S_WAIT_XCNT 0" between SMEM and VMEM instructions, so there will never be outstanding address translations for both SMEM and VMEM at the same time.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
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.
Hardware does an implicit "S_WAIT_XCNT 0" between SMEM and VMEM instructions, so there will never be outstanding address translations for both SMEM and VMEM at the same time.