[DAG] Fold rem(rem(A, BCst), Op1Cst) -> rem(A, Op1Cst) #159517
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
|
@llvm/pr-subscribers-backend-risc-v @llvm/pr-subscribers-llvm-selectiondag Author: None (kper) ChangesFixes 157370 General proof: https://alive2.llvm.org/ce/z/b_GQJX I have added it as rv32i and rv64i tests because they are the only architectures where I could verify that it works. Full diff: https://github.com/llvm/llvm-project/pull/159517.diff 2 Files Affected:
diff --git a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index 77bc47f28fc80..8e23282aad0fd 100644
--- a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -54,6 +54,7 @@
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/Metadata.h"
+#include "llvm/IR/PatternMatch.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/CodeGen.h"
#include "llvm/Support/CommandLine.h"
@@ -5405,6 +5406,14 @@ SDValue DAGCombiner::visitREM(SDNode *N) {
if (SDValue DivRem = useDivRem(N))
return DivRem.getValue(1);
+ // fold urem(urem(A, BCst), Op1Cst) -> urem(A, Op1Cst)
+ SDValue A;
+ APInt Op1Cst, BCCst;
+ if (sd_match(N0, m_URem(m_Value(A), m_ConstInt(BCCst))) &&
+ sd_match(N1, m_ConstInt(Op1Cst)) && BCCst.urem(Op1Cst).isZero()) {
+ return DAG.getNode(ISD::UREM, DL, VT, A, DAG.getConstant(Op1Cst, DL, VT));
+ }
+
return SDValue();
}
diff --git a/llvm/test/CodeGen/RISCV/urem.ll b/llvm/test/CodeGen/RISCV/urem.ll
new file mode 100644
index 0000000000000..aeec9abd0ed9f
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/urem.ll
@@ -0,0 +1,71 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv32 -verify-machineinstrs < %s \
+; RUN: | FileCheck -check-prefixes=CHECK,RV32I %s
+; RUN: llc -mtriple=riscv64 -verify-machineinstrs < %s \
+; RUN: | FileCheck -check-prefixes=CHECK,RV64I %s
+
+define i32 @fold_urem_constants(i32 %v0) {
+; RV32I-LABEL: fold_urem_constants:
+; RV32I: # %bb.0:
+; RV32I-NEXT: li a1, 5
+; RV32I-NEXT: tail __umodsi3
+;
+; RV64I-LABEL: fold_urem_constants:
+; RV64I: # %bb.0:
+; RV64I-NEXT: addi sp, sp, -16
+; RV64I-NEXT: .cfi_def_cfa_offset 16
+; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
+; RV64I-NEXT: .cfi_offset ra, -8
+; RV64I-NEXT: slli a0, a0, 32
+; RV64I-NEXT: srli a0, a0, 32
+; RV64I-NEXT: li a1, 5
+; RV64I-NEXT: call __umoddi3
+; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
+; RV64I-NEXT: .cfi_restore ra
+; RV64I-NEXT: addi sp, sp, 16
+; RV64I-NEXT: .cfi_def_cfa_offset 0
+; RV64I-NEXT: ret
+ %v1 = urem i32 %v0, 25
+ %v2 = urem i32 %v1, 5
+ ret i32 %v2
+}
+
+define i32 @dont_test_fold_urem_constants(i32 %v0) {
+; RV32I-LABEL: dont_test_fold_urem_constants:
+; RV32I: # %bb.0:
+; RV32I-NEXT: addi sp, sp, -16
+; RV32I-NEXT: .cfi_def_cfa_offset 16
+; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
+; RV32I-NEXT: .cfi_offset ra, -4
+; RV32I-NEXT: li a1, 25
+; RV32I-NEXT: call __umodsi3
+; RV32I-NEXT: li a1, 3
+; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
+; RV32I-NEXT: .cfi_restore ra
+; RV32I-NEXT: addi sp, sp, 16
+; RV32I-NEXT: .cfi_def_cfa_offset 0
+; RV32I-NEXT: tail __umodsi3
+;
+; RV64I-LABEL: dont_test_fold_urem_constants:
+; RV64I: # %bb.0:
+; RV64I-NEXT: addi sp, sp, -16
+; RV64I-NEXT: .cfi_def_cfa_offset 16
+; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
+; RV64I-NEXT: .cfi_offset ra, -8
+; RV64I-NEXT: slli a0, a0, 32
+; RV64I-NEXT: srli a0, a0, 32
+; RV64I-NEXT: li a1, 25
+; RV64I-NEXT: call __umoddi3
+; RV64I-NEXT: li a1, 3
+; RV64I-NEXT: call __umoddi3
+; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
+; RV64I-NEXT: .cfi_restore ra
+; RV64I-NEXT: addi sp, sp, 16
+; RV64I-NEXT: .cfi_def_cfa_offset 0
+; RV64I-NEXT: ret
+ %v1 = urem i32 %v0, 25
+ %v2 = urem i32 %v1, 3
+ ret i32 %v2
+}
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; CHECK: {{.*}}
|
|
cc @RKSimon |
jayfoad
reviewed
Sep 18, 2025
|
|
||
| // fold urem(urem(A, BCst), Op1Cst) -> urem(A, Op1Cst) | ||
| SDValue A; | ||
| APInt Op1Cst, BCCst; |
There was a problem hiding this comment.
Nit: call it BCst to match the comment?
There was a problem hiding this comment.
Thanks, it was a typo :)
I wasn't really sure. We only added it because it showed that we're breaking up the div+rem patterns. Beside that there wasn't a functional reason. So should I add it again? |
RKSimon
reviewed
Sep 18, 2025
RKSimon
reviewed
Sep 18, 2025
llvm/test/CodeGen/RISCV/srem.ll
Outdated
| ; RUN: llc -mtriple=riscv64 -verify-machineinstrs < %s \ | ||
| ; RUN: | FileCheck -check-prefixes=CHECK,RV64I %s | ||
|
|
||
| define i32 @fold_srem_constants(i32 %v0) { |
There was a problem hiding this comment.
add nounwind to get rid of all the cfi noise
RKSimon
reviewed
Sep 18, 2025
| #include "llvm/IR/DerivedTypes.h" | ||
| #include "llvm/IR/Function.h" | ||
| #include "llvm/IR/Metadata.h" | ||
| #include "llvm/IR/Type.h" |
RKSimon
reviewed
Sep 18, 2025
| if (sd_match(N, m_SRem(m_SRem(m_Value(A), m_ConstInt(BCst)), | ||
| m_ConstInt(Op1Cst))) && | ||
| BCst.srem(Op1Cst).isZero() && | ||
| Op1Cst.ne(APInt::getAllOnes(Op1Cst.getBitWidth()))) { |
|
Gentle ping |
RKSimon
changed the title
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/27/builds/16404 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/65/builds/22972 Here is the relevant piece of the build log for the reference |
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.
Fixes 157370
UREM General proof: https://alive2.llvm.org/ce/z/b_GQJX
SREM General proof: https://alive2.llvm.org/ce/z/Whkaxh
I have added it as rv32i and rv64i tests because they are the only architectures where I could verify that it works.