Skip to content

[ExpandVectorPredication] Add vp.select to foldEVLIntoMask. #157720

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Sep 9, 2025
Merged

[ExpandVectorPredication] Add vp.select to foldEVLIntoMask. #157720

merged 1 commit into from
Sep 9, 2025

Conversation

@topperc
Copy link
Collaborator

@topperc topperc commented Sep 9, 2025

No description provided.

@topperc topperc requested review from lukel97 and preames September 9, 2025 17:29
@llvmbot
Copy link
Member

llvmbot commented Sep 9, 2025

@llvm/pr-subscribers-llvm-transforms

Author: Craig Topper (topperc)

Changes

Full diff: https://github.com/llvm/llvm-project/pull/157720.diff

2 Files Affected:

  • (modified) llvm/lib/CodeGen/ExpandVectorPredication.cpp (+4-2)
  • (modified) llvm/test/Transforms/PreISelIntrinsicLowering/expand-vp.ll (+8)
diff --git a/llvm/lib/CodeGen/ExpandVectorPredication.cpp b/llvm/lib/CodeGen/ExpandVectorPredication.cpp
index 673c4f7434d6c..75e8fe4965025 100644
--- a/llvm/lib/CodeGen/ExpandVectorPredication.cpp
+++ b/llvm/lib/CodeGen/ExpandVectorPredication.cpp
@@ -521,7 +521,8 @@ bool CachingVPExpander::foldEVLIntoMask(VPIntrinsic &VPI) {
   // Only VP intrinsics can have an %evl parameter.
   Value *OldMaskParam = VPI.getMaskParam();
   if (!OldMaskParam) {
-    assert(VPI.getIntrinsicID() == Intrinsic::vp_merge &&
+    assert((VPI.getIntrinsicID() == Intrinsic::vp_merge ||
+            VPI.getIntrinsicID() == Intrinsic::vp_select) &&
            "Unexpected VP intrinsic without mask operand");
     OldMaskParam = VPI.getArgOperand(0);
   }
@@ -537,7 +538,8 @@ bool CachingVPExpander::foldEVLIntoMask(VPIntrinsic &VPI) {
   ElementCount ElemCount = VPI.getStaticVectorLength();
   Value *VLMask = convertEVLToMask(Builder, OldEVLParam, ElemCount);
   Value *NewMaskParam = Builder.CreateAnd(VLMask, OldMaskParam);
-  if (VPI.getIntrinsicID() == Intrinsic::vp_merge)
+  if (VPI.getIntrinsicID() == Intrinsic::vp_merge ||
+      VPI.getIntrinsicID() == Intrinsic::vp_select)
     VPI.setArgOperand(0, NewMaskParam);
   else
     VPI.setMaskParam(NewMaskParam);
diff --git a/llvm/test/Transforms/PreISelIntrinsicLowering/expand-vp.ll b/llvm/test/Transforms/PreISelIntrinsicLowering/expand-vp.ll
index fe7d725439060..0c3a7c681c4d0 100644
--- a/llvm/test/Transforms/PreISelIntrinsicLowering/expand-vp.ll
+++ b/llvm/test/Transforms/PreISelIntrinsicLowering/expand-vp.ll
@@ -69,6 +69,7 @@ define void @test_vp_int_v8(<8 x i32> %i0, <8 x i32> %i1, <8 x i32> %i2, <8 x i3
   %rF = call <8 x i32> @llvm.vp.lshr.v8i32(<8 x i32> %i0, <8 x i32> %i1, <8 x i1> %m, i32 %n)
   %r10 = call <8 x i32> @llvm.vp.shl.v8i32(<8 x i32> %i0, <8 x i32> %i1, <8 x i1> %m, i32 %n)
   %r11 = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %i0, <8 x i32> %i1, i32 %n)
+  %r12 = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %m, <8 x i32> %i0, <8 x i32> %i1, i32 %n)
   ret void
 }
 
@@ -113,6 +114,7 @@ define void @test_vp_int_vscale(<vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1,
   %rF = call <vscale x 4 x i32> @llvm.vp.lshr.nxv4i32(<vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, <vscale x 4 x i1> %m, i32 %n)
   %r10 = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, <vscale x 4 x i1> %m, i32 %n)
   %r11 = call <vscale x 4 x i32> @llvm.vp.merge.nxv4i32(<vscale x 4 x i1> %m, <vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, i32 %n)
+  %r12 = call <vscale x 4 x i32> @llvm.vp.select.nxv4i32(<vscale x 4 x i1> %m, <vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, i32 %n)
   ret void
 }
 
@@ -325,6 +327,7 @@ define void @test_vp_cmp_v8(<8 x i32> %i0, <8 x i32> %i1, <8 x float> %f0, <8 x
 ; LEGAL_LEGAL-NEXT:   %rF = call <8 x i32> @llvm.vp.lshr.v8i32(<8 x i32> %i0, <8 x i32> %i1, <8 x i1> %m, i32 %n)
 ; LEGAL_LEGAL-NEXT:   %r10 = call <8 x i32> @llvm.vp.shl.v8i32(<8 x i32> %i0, <8 x i32> %i1, <8 x i1> %m, i32 %n)
 ; LEGAL_LEGAL-NEXT:   %r11 = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %i0, <8 x i32> %i1, i32 %n)
+; LEGAL_LEGAL-NEXT:   %r12 = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %m, <8 x i32> %i0, <8 x i32> %i1, i32 %n)
 ; LEGAL_LEGAL-NEXT:   ret void
 
 ; LEGAL_LEGAL:define void @test_vp_int_vscale(<vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, <vscale x 4 x i32> %i2, <vscale x 4 x i32> %f3, <vscale x 4 x i1> %m, i32 %n) {
@@ -346,6 +349,7 @@ define void @test_vp_cmp_v8(<8 x i32> %i0, <8 x i32> %i1, <8 x float> %f0, <8 x
 ; LEGAL_LEGAL-NEXT:  %rF = call <vscale x 4 x i32> @llvm.vp.lshr.nxv4i32(<vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, <vscale x 4 x i1> %m, i32 %n)
 ; LEGAL_LEGAL-NEXT:  %r10 = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, <vscale x 4 x i1> %m, i32 %n)
 ; LEGAL_LEGAL-NEXT:  %r11 = call <vscale x 4 x i32> @llvm.vp.merge.nxv4i32(<vscale x 4 x i1> %m, <vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, i32 %n)
+; LEGAL_LEGAL-NEXT:  %r12 = call <vscale x 4 x i32> @llvm.vp.select.nxv4i32(<vscale x 4 x i1> %m, <vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, i32 %n)
 ; LEGAL_LEGAL-NEXT:  ret void
 
 ; LEGAL_LEGAL: define void @test_vp_reduce_int_v4(i32 %start, <4 x i32> %vi, <4 x i1> %m, i32 %n) {
@@ -424,6 +428,7 @@ define void @test_vp_cmp_v8(<8 x i32> %i0, <8 x i32> %i1, <8 x float> %f0, <8 x
 ; DISCARD_LEGAL-NEXT:   [[EVLMASK2:%.+]] = icmp ult <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>, [[NSPLAT2]]
 ; DISCARD_LEGAL-NEXT:   [[NEWMASK2:%.+]] = and <8 x i1> [[EVLMASK2]], %m
 ; DISCARD_LEGAL-NEXT:   %r11 = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> [[NEWMASK2]], <8 x i32> %i0, <8 x i32> %i1, i32 8)
+; DISCARD_LEGAL-NEXT:   %r12 = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %m, <8 x i32> %i0, <8 x i32> %i1, i32 8)
 ; DISCARD_LEGAL-NEXT:   ret void
 
 ; TODO compute vscale only once and use caching.
@@ -441,6 +446,7 @@ define void @test_vp_cmp_v8(<8 x i32> %i0, <8 x i32> %i1, <8 x float> %f0, <8 x
 ; DISCARD_LEGAL:      %r3 = call <vscale x 4 x i32> @llvm.vp.sdiv.nxv4i32(<vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, <vscale x 4 x i1> [[NEWM]], i32 %scalable_size{{.*}})
 ; DISCARD_LEGAL-NOT:  %{{.+}} = call <vscale x 4 x i32> @llvm.vp.{{.*}}, i32 %n)
 ; DISCARD_LEGAL:      %r11 = call <vscale x 4 x i32> @llvm.vp.merge.nxv4i32(<vscale x 4 x i1> %{{.*}}, <vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, i32 %scalable_size{{.*}})
+; DISCARD_LEGAL:      %r12 = call <vscale x 4 x i32> @llvm.vp.select.nxv4i32(<vscale x 4 x i1> %m, <vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, i32 %scalable_size{{.*}})
 ; DISCARD_LEGAL-NEXT: ret void
 
 ; DISCARD_LEGAL: define void @test_vp_reduce_int_v4(i32 %start, <4 x i32> %vi, <4 x i1> %m, i32 %n) {
@@ -514,6 +520,7 @@ define void @test_vp_cmp_v8(<8 x i32> %i0, <8 x i32> %i1, <8 x float> %f0, <8 x
 ; CONVERT_LEGAL-NOT:   %{{.+}} = call <8 x i32> @llvm.vp.lshr.v8i32(<8 x i32> %i0, <8 x i32> %i1, <8 x i1> %m, i32 8)
 ; CONVERT_LEGAL-NOT:   %{{.+}} = call <8 x i32> @llvm.vp.shl.v8i32(<8 x i32> %i0, <8 x i32> %i1, <8 x i1> %m, i32 8)
 ; CONVERT_LEGAL:        %r11 = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %{{.*}}, <8 x i32> %i0, <8 x i32> %i1, i32 8)
+; CONVERT_LEGAL:        %r12 = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %{{.*}}, <8 x i32> %i0, <8 x i32> %i1, i32 8)
 ; CONVERT_LEGAL:       ret void
 
 ; Similar to %evl discard, %mask legal but make sure the first VP intrinsic has a legal expansion
@@ -525,6 +532,7 @@ define void @test_vp_cmp_v8(<8 x i32> %i0, <8 x i32> %i1, <8 x float> %f0, <8 x
 ; CONVERT_LEGAL-NEXT:   %r0 = call <vscale x 4 x i32> @llvm.vp.add.nxv4i32(<vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, <vscale x 4 x i1> [[NEWM]], i32 %scalable_size)
 ; CONVERT_LEGAL-NOT:    %{{.*}} = call <vscale x 4 x i32> @llvm.vp.{{.*}}, i32 %n)
 ; CONVERT_LEGAL:        %r11 = call <vscale x 4 x i32> @llvm.vp.merge.nxv4i32(<vscale x 4 x i1> %{{.*}}, <vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, i32 %scalable_size{{.*}})
+; CONVERT_LEGAL:        %r12 = call <vscale x 4 x i32> @llvm.vp.select.nxv4i32(<vscale x 4 x i1> %{{.*}}, <vscale x 4 x i32> %i0, <vscale x 4 x i32> %i1, i32 %scalable_size{{.*}})
 ; CONVERT_LEGAL:        ret void
 
 ; CONVERT_LEGAL: define void @test_vp_reduce_int_v4(i32 %start, <4 x i32> %vi, <4 x i1> %m, i32 %n) {

@topperc topperc requested a review from wangpc-pp September 9, 2025 18:43
artagnon
artagnon reviewed Sep 9, 2025
View reviewed changes
llvm/test/Transforms/PreISelIntrinsicLowering/expand-vp.ll
%rF = call <8 x i32> @llvm.vp.lshr.v8i32(<8 x i32> %i0, <8 x i32> %i1, <8 x i1> %m, i32 %n)
%r10 = call <8 x i32> @llvm.vp.shl.v8i32(<8 x i32> %i0, <8 x i32> %i1, <8 x i1> %m, i32 %n)
%r11 = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %i0, <8 x i32> %i1, i32 %n)
%r12 = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %m, <8 x i32> %i0, <8 x i32> %i1, i32 %n)
Copy link
Contributor

@artagnon artagnon Sep 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Noting that maximu/minimum/maxnum/minnum are missing from this list. I think the code covers them though?

preames
preames approved these changes Sep 9, 2025
View reviewed changes
Copy link
Collaborator

@preames preames left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@topperc topperc merged commit 2422686 into llvm:main Sep 9, 2025
12 checks passed
@topperc topperc deleted the pr/vp-select branch September 9, 2025 22:56
@llvm-ci
Copy link
Collaborator

llvm-ci commented Sep 9, 2025

LLVM Buildbot has detected a new failure on builder sanitizer-x86_64-linux-android running on sanitizer-buildbot-android while building llvm at step 2 "annotate".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/186/builds/12285

Here is the relevant piece of the build log for the reference 
Step 2 (annotate) failure: 'python ../sanitizer_buildbot/sanitizers/zorg/buildbot/builders/sanitizers/buildbot_selector.py' (failure)
...
[       OK ] AddressSanitizer.AtoiAndFriendsOOBTest (2264 ms)
[ RUN      ] AddressSanitizer.HasFeatureAddressSanitizerTest
[       OK ] AddressSanitizer.HasFeatureAddressSanitizerTest (0 ms)
[ RUN      ] AddressSanitizer.CallocReturnsZeroMem
[       OK ] AddressSanitizer.CallocReturnsZeroMem (9 ms)
[ DISABLED ] AddressSanitizer.DISABLED_TSDTest
[ RUN      ] AddressSanitizer.IgnoreTest
[       OK ] AddressSanitizer.IgnoreTest (0 ms)
[ RUN      ] AddressSanitizer.SignalTest
[       OK ] AddressSanitizer.SignalTest (193 ms)
[ RUN      ] AddressSanitizer.ReallocTest
[       OK ] AddressSanitizer.ReallocTest (37 ms)
[ RUN      ] AddressSanitizer.WrongFreeTest
[       OK ] AddressSanitizer.WrongFreeTest (111 ms)
[ RUN      ] AddressSanitizer.LongJmpTest
[       OK ] AddressSanitizer.LongJmpTest (0 ms)
[ RUN      ] AddressSanitizer.ThreadStackReuseTest
[       OK ] AddressSanitizer.ThreadStackReuseTest (1 ms)
[ DISABLED ] AddressSanitizer.DISABLED_MemIntrinsicUnalignedAccessTest
[ DISABLED ] AddressSanitizer.DISABLED_LargeFunctionSymbolizeTest
[ DISABLED ] AddressSanitizer.DISABLED_MallocFreeUnwindAndSymbolizeTest
[ RUN      ] AddressSanitizer.UseThenFreeThenUseTest
[       OK ] AddressSanitizer.UseThenFreeThenUseTest (125 ms)
[ RUN      ] AddressSanitizer.FileNameInGlobalReportTest
[       OK ] AddressSanitizer.FileNameInGlobalReportTest (128 ms)
[ DISABLED ] AddressSanitizer.DISABLED_StressStackReuseAndExceptionsTest
[ RUN      ] AddressSanitizer.MlockTest
[       OK ] AddressSanitizer.MlockTest (0 ms)
[ DISABLED ] AddressSanitizer.DISABLED_DemoThreadedTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoStackTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoThreadStackTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoUAFLowIn
[ DISABLED ] AddressSanitizer.DISABLED_DemoUAFLowLeft
[ DISABLED ] AddressSanitizer.DISABLED_DemoUAFLowRight
[ DISABLED ] AddressSanitizer.DISABLED_DemoUAFHigh
[ DISABLED ] AddressSanitizer.DISABLED_DemoOOM
[ DISABLED ] AddressSanitizer.DISABLED_DemoDoubleFreeTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoNullDerefTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoFunctionStaticTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoTooMuchMemoryTest
[ RUN      ] AddressSanitizer.LongDoubleNegativeTest
[       OK ] AddressSanitizer.LongDoubleNegativeTest (0 ms)
[----------] 19 tests from AddressSanitizer (28083 ms total)

[----------] Global test environment tear-down
[==========] 22 tests from 2 test suites ran. (28096 ms total)
[  PASSED  ] 22 tests.

  YOU HAVE 1 DISABLED TEST

Step 34 (run instrumented asan tests [aarch64/bluejay-userdebug/TQ3A.230805.001]) failure: run instrumented asan tests [aarch64/bluejay-userdebug/TQ3A.230805.001] (failure)
...
[ RUN      ] AddressSanitizer.HasFeatureAddressSanitizerTest
[       OK ] AddressSanitizer.HasFeatureAddressSanitizerTest (0 ms)
[ RUN      ] AddressSanitizer.CallocReturnsZeroMem
[       OK ] AddressSanitizer.CallocReturnsZeroMem (9 ms)
[ DISABLED ] AddressSanitizer.DISABLED_TSDTest
[ RUN      ] AddressSanitizer.IgnoreTest
[       OK ] AddressSanitizer.IgnoreTest (0 ms)
[ RUN      ] AddressSanitizer.SignalTest
[       OK ] AddressSanitizer.SignalTest (193 ms)
[ RUN      ] AddressSanitizer.ReallocTest
[       OK ] AddressSanitizer.ReallocTest (37 ms)
[ RUN      ] AddressSanitizer.WrongFreeTest
[       OK ] AddressSanitizer.WrongFreeTest (111 ms)
[ RUN      ] AddressSanitizer.LongJmpTest
[       OK ] AddressSanitizer.LongJmpTest (0 ms)
[ RUN      ] AddressSanitizer.ThreadStackReuseTest
[       OK ] AddressSanitizer.ThreadStackReuseTest (1 ms)
[ DISABLED ] AddressSanitizer.DISABLED_MemIntrinsicUnalignedAccessTest
[ DISABLED ] AddressSanitizer.DISABLED_LargeFunctionSymbolizeTest
[ DISABLED ] AddressSanitizer.DISABLED_MallocFreeUnwindAndSymbolizeTest
[ RUN      ] AddressSanitizer.UseThenFreeThenUseTest
[       OK ] AddressSanitizer.UseThenFreeThenUseTest (125 ms)
[ RUN      ] AddressSanitizer.FileNameInGlobalReportTest
[       OK ] AddressSanitizer.FileNameInGlobalReportTest (128 ms)
[ DISABLED ] AddressSanitizer.DISABLED_StressStackReuseAndExceptionsTest
[ RUN      ] AddressSanitizer.MlockTest
[       OK ] AddressSanitizer.MlockTest (0 ms)
[ DISABLED ] AddressSanitizer.DISABLED_DemoThreadedTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoStackTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoThreadStackTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoUAFLowIn
[ DISABLED ] AddressSanitizer.DISABLED_DemoUAFLowLeft
[ DISABLED ] AddressSanitizer.DISABLED_DemoUAFLowRight
[ DISABLED ] AddressSanitizer.DISABLED_DemoUAFHigh
[ DISABLED ] AddressSanitizer.DISABLED_DemoOOM
[ DISABLED ] AddressSanitizer.DISABLED_DemoDoubleFreeTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoNullDerefTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoFunctionStaticTest
[ DISABLED ] AddressSanitizer.DISABLED_DemoTooMuchMemoryTest
[ RUN      ] AddressSanitizer.LongDoubleNegativeTest
[       OK ] AddressSanitizer.LongDoubleNegativeTest (0 ms)
[----------] 19 tests from AddressSanitizer (28083 ms total)

[----------] Global test environment tear-down
[==========] 22 tests from 2 test suites ran. (28096 ms total)
[  PASSED  ] 22 tests.

  YOU HAVE 1 DISABLED TEST
program finished with exit code 0
elapsedTime=2298.520977

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@artagnon artagnon artagnon left review comments

@preames preames preames approved these changes

@lukel97 lukel97 Awaiting requested review from lukel97

@wangpc-pp wangpc-pp Awaiting requested review from wangpc-pp

Assignees

No one assigned

Labels

llvm:codegen llvm:transforms

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@topperc @llvmbot @llvm-ci @artagnon @preames