[RISCV][EVL] Drop EVLIndVarSimplifyPass from the pipeline #151483
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-risc-v Author: Shih-Po Hung (arcbbb) ChangesWith the VPlan-based canonical induction variable replacement landed in #147222, it appears to cover the functionality previously provided by the EVLIndVarSimplify pass introduced in #131005. As a next step, this patch proposes disabling EVLIndVarSimplify by default to begin a gradual phase-out. Full diff: https://github.com/llvm/llvm-project/pull/151483.diff 1 Files Affected:
diff --git a/llvm/lib/Target/RISCV/RISCVTargetMachine.cpp b/llvm/lib/Target/RISCV/RISCVTargetMachine.cpp
index da6ac2f6f31e9..9d648237743b4 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetMachine.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetMachine.cpp
@@ -114,6 +114,11 @@ static cl::opt<bool>
cl::desc("Enable Machine Pipeliner for RISC-V"),
cl::init(false), cl::Hidden);
+static cl::opt<bool> EnableEVLIndVarSimplify(
+ "riscv-simplify-evl-iv",
+ cl::desc("Enable the EVLIndVarSimplify pass."), cl::init(false),
+ cl::Hidden);
+
extern "C" LLVM_ABI LLVM_EXTERNAL_VISIBILITY void LLVMInitializeRISCVTarget() {
RegisterTargetMachine<RISCVTargetMachine> X(getTheRISCV32Target());
RegisterTargetMachine<RISCVTargetMachine> Y(getTheRISCV64Target());
@@ -645,7 +650,7 @@ void RISCVTargetMachine::registerPassBuilderCallbacks(PassBuilder &PB) {
PB.registerVectorizerEndEPCallback(
[](FunctionPassManager &FPM, OptimizationLevel Level) {
- if (Level.isOptimizingForSpeed())
+ if (Level.isOptimizingForSpeed() && EnableEVLIndVarSimplify)
FPM.addPass(createFunctionToLoopPassAdaptor(EVLIndVarSimplifyPass()));
});
}
|
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
lukel97
reviewed
Jul 31, 2025
Contributor
There was a problem hiding this comment.
I think we're quite early in the LLVM release cycle, should we just go ahead and stop emitting the metadata from the LoopVectorizer and remove the pass?
I don't think EVLIndVarSimplfiyPass is actually kicking in today anymore, it always seems to bail with Cannot retrieve IV from loop vector.body becausecannot recognize induction variable after #147222. So I'm not sure if adding a flag really changes anything.
mshockwave
reviewed
Jul 31, 2025
Member
mshockwave
left a comment
mshockwave
left a comment
There was a problem hiding this comment.
should we just go ahead and stop emitting the metadata
Personally I'll say we can stop emitting the metadata and remove it from RISCV's pipeline, but keeping the Pass just in case.
Contributor
I'm ok with that too |
arcbbb
changed the title
Contributor
Author
Let's handle the metadata in a separate PR and keep this one focused on removing the pass from the pipeline. @mshockwave I do have a question regarding the metadata. Does "stop emitting" imply a complete removal, or just adding a switch to disable it? I'm not sure it is consistent to remove the metadata entirely while there's still a pass that queries it. |
Collaborator
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/123/builds/24704 Here is the relevant piece of the build log for the reference |
Collaborator
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/137/builds/22971 Here is the relevant piece of the build log for the reference |
Member
Sorry this message slipped through my radar: yes, I think we could just add a switch to disable it for now. |
lukel97
added a commit
to lukel97/llvm-project
that referenced
this pull request
Sep 24, 2025
Initially this was needed to replace the fixed-step canonical IV with the variable-step EVL IV, but this was eventually superseded by the loop vectorizer doing this transform itself in llvm#147222. The pass was then removed from the RISC-V pipeline in llvm#151483 and the loop vectorizer stopped emitting the metadata used by the pass in llvm#155760, so now there's no users of it.
lukel97
added a commit
that referenced
this pull request
Sep 25, 2025
Initially this was needed to replace the fixed-step canonical IV with the variable-step EVL IV, but this was eventually superseded by the loop vectorizer doing this transform itself in #147222. The pass was then removed from the RISC-V pipeline in #151483 and the loop vectorizer stopped emitting the metadata used by the pass in #155760, so now there's no users of it.
mahesh-attarde
pushed a commit
to mahesh-attarde/llvm-project
that referenced
this pull request
Oct 3, 2025
Initially this was needed to replace the fixed-step canonical IV with the variable-step EVL IV, but this was eventually superseded by the loop vectorizer doing this transform itself in llvm#147222. The pass was then removed from the RISC-V pipeline in llvm#151483 and the loop vectorizer stopped emitting the metadata used by the pass in llvm#155760, so now there's no users of it.
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.
With the VPlan-based canonical induction variable replacement landed in #147222, it appears to cover the functionality previously provided by the EVLIndVarSimplify pass introduced in #131005.
This patch suggests removing EVLIndVarSimplify from the RISC-V pipeline as a follow-up step. Feedback is very welcome!