Skip to content

[X86] widenSubVector - widen from smaller build vector if the upper elements are already the same padding elements #122445

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
Jan 10, 2025
Merged

[X86] widenSubVector - widen from smaller build vector if the upper elements are already the same padding elements #122445

merged 1 commit into from
Jan 10, 2025

Conversation

@RKSimon
Copy link
Collaborator

@RKSimon RKSimon commented Jan 10, 2025

No description provided.

@RKSimon
[X86] widenSubVector - widen from smaller build vector if the upper e…
d5282a7 
…lements are already the same padding elements
@llvmbot
Copy link
Member

llvmbot commented Jan 10, 2025

@llvm/pr-subscribers-backend-x86

Author: Simon Pilgrim (RKSimon)

Changes

Full diff: https://github.com/llvm/llvm-project/pull/122445.diff

3 Files Affected:

  • (modified) llvm/lib/Target/X86/X86ISelLowering.cpp (+13-2)
  • (modified) llvm/test/CodeGen/X86/shuffle-vs-trunc-512.ll (+2-2)
  • (modified) llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll (+1-2)
diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
index 6b0eb38e7e0952..fbfcfc700ed62d 100644
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -4144,9 +4144,20 @@ static SDValue insert128BitVector(SDValue Result, SDValue Vec, unsigned IdxVal,
 static SDValue widenSubVector(MVT VT, SDValue Vec, bool ZeroNewElements,
                               const X86Subtarget &Subtarget, SelectionDAG &DAG,
                               const SDLoc &dl) {
-  assert(Vec.getValueSizeInBits().getFixedValue() <= VT.getFixedSizeInBits() &&
-         Vec.getValueType().getScalarType() == VT.getScalarType() &&
+  EVT VecVT = Vec.getValueType();
+  assert(VecVT.getFixedSizeInBits() <= VT.getFixedSizeInBits() &&
+         VecVT.getScalarType() == VT.getScalarType() &&
          "Unsupported vector widening type");
+  // If the upper 128-bits of a build vector are already undef/zero, then try to
+  // widen from the lower 128-bits.
+  if (Vec.getOpcode() == ISD::BUILD_VECTOR && VecVT.is256BitVector()) {
+    unsigned NumSrcElts = VecVT.getVectorNumElements();
+    ArrayRef<SDUse> Hi = Vec->ops().drop_front(NumSrcElts / 2);
+    if (all_of(Hi, [&](SDValue V) {
+          return V.isUndef() || (ZeroNewElements && X86::isZeroNode(V));
+        }))
+      Vec = extract128BitVector(Vec, 0, DAG, dl);
+  }
   SDValue Res = ZeroNewElements ? getZeroVector(VT, Subtarget, DAG, dl)
                                 : DAG.getUNDEF(VT);
   return DAG.getNode(ISD::INSERT_SUBVECTOR, dl, VT, Res, Vec,
diff --git a/llvm/test/CodeGen/X86/shuffle-vs-trunc-512.ll b/llvm/test/CodeGen/X86/shuffle-vs-trunc-512.ll
index 445468d06fb042..e7557134b14864 100644
--- a/llvm/test/CodeGen/X86/shuffle-vs-trunc-512.ll
+++ b/llvm/test/CodeGen/X86/shuffle-vs-trunc-512.ll
@@ -442,7 +442,7 @@ define <4 x double> @PR34175(ptr %p) {
 ;
 ; AVX512BW-LABEL: PR34175:
 ; AVX512BW:       # %bb.0:
-; AVX512BW-NEXT:    vpbroadcastq {{.*#+}} ymm0 = [0,8,32,40,0,8,32,40,0,8,32,40,0,8,32,40]
+; AVX512BW-NEXT:    vmovq {{.*#+}} xmm0 = [0,8,32,40,0,0,0,0]
 ; AVX512BW-NEXT:    vmovdqu (%rdi), %ymm1
 ; AVX512BW-NEXT:    vmovdqu 32(%rdi), %ymm2
 ; AVX512BW-NEXT:    vpermt2w %zmm2, %zmm0, %zmm1
@@ -461,7 +461,7 @@ define <4 x double> @PR34175(ptr %p) {
 ;
 ; AVX512VBMI-LABEL: PR34175:
 ; AVX512VBMI:       # %bb.0:
-; AVX512VBMI-NEXT:    vpbroadcastq {{.*#+}} ymm0 = [0,8,32,40,0,8,32,40,0,8,32,40,0,8,32,40]
+; AVX512VBMI-NEXT:    vmovq {{.*#+}} xmm0 = [0,8,32,40,0,0,0,0]
 ; AVX512VBMI-NEXT:    vmovdqu (%rdi), %ymm1
 ; AVX512VBMI-NEXT:    vmovdqu 32(%rdi), %ymm2
 ; AVX512VBMI-NEXT:    vpermt2w %zmm2, %zmm0, %zmm1
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll b/llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll
index 81ce14132c8799..05071064fc60e1 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll
@@ -598,8 +598,7 @@ define void @PR48908(<4 x double> %v0, <4 x double> %v1, <4 x double> %v2, ptr n
 ; X64-AVX512-NEXT:    vmovapd {{.*#+}} ymm3 = [0,3,10,1]
 ; X64-AVX512-NEXT:    vpermi2pd %zmm0, %zmm4, %zmm3
 ; X64-AVX512-NEXT:    vmovapd %ymm3, (%rsi)
-; X64-AVX512-NEXT:    vbroadcastf128 {{.*#+}} ymm3 = [3,11,3,11]
-; X64-AVX512-NEXT:    # ymm3 = mem[0,1,0,1]
+; X64-AVX512-NEXT:    vmovapd {{.*#+}} xmm3 = [3,11]
 ; X64-AVX512-NEXT:    vpermi2pd %zmm1, %zmm0, %zmm3
 ; X64-AVX512-NEXT:    vmovapd {{.*#+}} ymm0 = [2,8,9,3]
 ; X64-AVX512-NEXT:    vpermi2pd %zmm3, %zmm2, %zmm0

@github-actions
Copy link

github-actions bot commented Jan 10, 2025

⚠️ undef deprecator found issues in your code. ⚠️

You can test this locally with the following command: 
git diff -U0 --pickaxe-regex -S '([^a-zA-Z0-9#_-]undef[^a-zA-Z0-9_-]|UndefValue::get)' 9b49da2b3169544355192dfd8d6909213169d0c1 d5282a773f6bae02a162b710036bd39e47080fe6 llvm/lib/Target/X86/X86ISelLowering.cpp llvm/test/CodeGen/X86/shuffle-vs-trunc-512.ll llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll

The following files introduce new uses of undef:

  • llvm/lib/Target/X86/X86ISelLowering.cpp

Undef is now deprecated and should only be used in the rare cases where no replacement is possible. For example, a load of uninitialized memory yields undef. You should use poison values for placeholders instead.

In tests, avoid using undef and having tests that trigger undefined behavior. If you need an operand with some unimportant value, you can add a new argument to the function and use that instead.

For example, this is considered a bad practice:

define void @fn() {
  ...
  br i1 undef, ...
}

Please use the following instead:

define void @fn(i1 %cond) {
  ...
  br i1 %cond, ...
}

Please refer to the Undefined Behavior Manual for more information.

@RKSimon RKSimon merged commit 35a3925 into llvm:main Jan 10, 2025
9 of 10 checks passed
@RKSimon RKSimon deleted the x86-widen-build-subvector branch January 10, 2025 15:13
@llvm-ci
Copy link
Collaborator

llvm-ci commented Jan 10, 2025

LLVM Buildbot has detected a new failure on builder llvm-clang-x86_64-expensive-checks-debian running on gribozavr4 while building llvm at step 6 "test-build-unified-tree-check-all".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/16/builds/11778

Here is the relevant piece of the build log for the reference 
Step 6 (test-build-unified-tree-check-all) failure: test (failure)
******************** TEST 'LLVM :: tools/llvm-gsymutil/ARM_AArch64/macho-merged-funcs-dwarf.yaml' FAILED ********************
Exit Code: 1

Command Output (stdout):
--
Input file: /b/1/llvm-clang-x86_64-expensive-checks-debian/build/test/tools/llvm-gsymutil/ARM_AArch64/Output/macho-merged-funcs-dwarf.yaml.tmp.dSYM
Output file (aarch64): /b/1/llvm-clang-x86_64-expensive-checks-debian/build/test/tools/llvm-gsymutil/ARM_AArch64/Output/macho-merged-funcs-dwarf.yaml.tmp.default.gSYM
Loaded 3 functions from DWARF.
Loaded 3 functions from symbol table.
warning: same address range contains different debug info. Removing:
[0x0000000000000248 - 0x0000000000000270): Name=0x00000030
addr=0x0000000000000248, file=  2, line=  5
addr=0x0000000000000254, file=  2, line=  7
addr=0x0000000000000258, file=  2, line=  9
addr=0x000000000000025c, file=  2, line=  8
addr=0x0000000000000260, file=  2, line= 11
addr=0x0000000000000264, file=  2, line= 10
addr=0x0000000000000268, file=  2, line=  6


In favor of this one:
[0x0000000000000248 - 0x0000000000000270): Name=0x00000047
addr=0x0000000000000248, file=  3, line=  5
addr=0x0000000000000254, file=  3, line=  7
addr=0x0000000000000258, file=  3, line=  9
addr=0x000000000000025c, file=  3, line=  8
addr=0x0000000000000260, file=  3, line= 11
addr=0x0000000000000264, file=  3, line= 10
addr=0x0000000000000268, file=  3, line=  6


warning: same address range contains different debug info. Removing:
[0x0000000000000248 - 0x0000000000000270): Name=0x00000047
addr=0x0000000000000248, file=  3, line=  5
addr=0x0000000000000254, file=  3, line=  7
addr=0x0000000000000258, file=  3, line=  9
addr=0x000000000000025c, file=  3, line=  8
addr=0x0000000000000260, file=  3, line= 11
addr=0x0000000000000264, file=  3, line= 10
addr=0x0000000000000268, file=  3, line=  6


In favor of this one:
[0x0000000000000248 - 0x0000000000000270): Name=0x00000001
addr=0x0000000000000248, file=  1, line=  5
addr=0x0000000000000254, file=  1, line=  7
addr=0x0000000000000258, file=  1, line=  9
addr=0x000000000000025c, file=  1, line=  8
addr=0x0000000000000260, file=  1, line= 11
addr=0x0000000000000264, file=  1, line= 10
...

BaiXilin pushed a commit to BaiXilin/llvm-project that referenced this pull request Jan 12, 2025
@RKSimon @BaiXilin
[X86] widenSubVector - widen from smaller build vector if the upper e…
a4a25d6 
…lements are already the same padding elements (llvm#122445)

Further simplifies some shuffle masks to help additional combines
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

backend:X86

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@RKSimon @llvmbot @llvm-ci