Skip to content

Commit

Permalink
[AArch64] Don't generate neon integer complex numbers with +sve2. NFC (
Browse files Browse the repository at this point in the history
…llvm#79829)

The condition for allowing integer complex number support could also
allow neon fixed length complex numbers if +sve2 was specified. This
tightens the condition to only allow integer complex number support for
scalable vectors.

We could generalize this in the future to generate SVE intrinsics for
fixed-length vectors, but for the moment this opts for the simpler fix.
  • Loading branch information
davemgreen authored Jan 29, 2024
1 parent 430c1fd commit 9520773
Show file tree
Hide file tree
Showing 2 changed files with 29 additions and 2 deletions.
4 changes: 2 additions & 2 deletions llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -26949,7 +26949,7 @@ bool AArch64TargetLowering::isComplexDeinterleavingOperationSupported(
return false;

// If the vector is scalable, SVE is enabled, implying support for complex
// numbers. Otherwirse, we need to ensure complex number support is avaialble
// numbers. Otherwise, we need to ensure complex number support is available
if (!VTy->isScalableTy() && !Subtarget->hasComplxNum())
return false;

Expand All @@ -26965,7 +26965,7 @@ bool AArch64TargetLowering::isComplexDeinterleavingOperationSupported(
!llvm::isPowerOf2_32(VTyWidth))
return false;

if (ScalarTy->isIntegerTy() && Subtarget->hasSVE2()) {
if (ScalarTy->isIntegerTy() && Subtarget->hasSVE2() && VTy->isScalableTy()) {
unsigned ScalarWidth = ScalarTy->getScalarSizeInBits();
return 8 <= ScalarWidth && ScalarWidth <= 64;
}
Expand Down
27 changes: 27 additions & 0 deletions llvm/test/CodeGen/AArch64/complex-deinterleaving-f16-add.ll
Original file line number Diff line number Diff line change
@@ -1,6 +1,7 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s --mattr=+complxnum,+neon,+fullfp16 -o - | FileCheck %s
; RUN: llc < %s --mattr=+complxnum,+neon,+fullfp16,+sve -o - | FileCheck %s
; RUN: llc < %s --mattr=+complxnum,+neon,+fullfp16,+sve2 -o - | FileCheck %s

target triple = "aarch64"

Expand Down Expand Up @@ -158,6 +159,32 @@ entry:
ret <16 x half> %interleaved.vec
}


; Expected not to transform as it is integer
define <16 x i16> @complex_add_v16i16(<16 x i16> %a, <16 x i16> %b) {
; CHECK-LABEL: complex_add_v16i16:
; CHECK: // %bb.0: // %entry
; CHECK-NEXT: uzp1 v4.8h, v2.8h, v3.8h
; CHECK-NEXT: uzp1 v5.8h, v0.8h, v1.8h
; CHECK-NEXT: uzp2 v0.8h, v0.8h, v1.8h
; CHECK-NEXT: uzp2 v1.8h, v2.8h, v3.8h
; CHECK-NEXT: sub v2.8h, v4.8h, v0.8h
; CHECK-NEXT: add v1.8h, v1.8h, v5.8h
; CHECK-NEXT: zip1 v0.8h, v2.8h, v1.8h
; CHECK-NEXT: zip2 v1.8h, v2.8h, v1.8h
; CHECK-NEXT: ret
entry:
%a.real = shufflevector <16 x i16> %a, <16 x i16> zeroinitializer, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14>
%a.imag = shufflevector <16 x i16> %a, <16 x i16> zeroinitializer, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15>
%b.real = shufflevector <16 x i16> %b, <16 x i16> zeroinitializer, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14>
%b.imag = shufflevector <16 x i16> %b, <16 x i16> zeroinitializer, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15>
%0 = sub <8 x i16> %b.real, %a.imag
%1 = add <8 x i16> %b.imag, %a.real
%interleaved.vec = shufflevector <8 x i16> %0, <8 x i16> %1, <16 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11, i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15>
ret <16 x i16> %interleaved.vec
}


declare { <2 x half>, <2 x half> } @llvm.experimental.vector.deinterleave2.v4f16(<4 x half>)
declare <4 x half> @llvm.experimental.vector.interleave2.v4f16(<2 x half>, <2 x half>)

Expand Down

0 comments on commit 9520773

Please sign in to comment.