[llvm] X86: Add more sincos vector test checks (PR #166041)
via llvm-commits
llvm-commits at lists.llvm.org
Sat Nov 1 23:15:19 PDT 2025
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-x86
Author: Matt Arsenault (arsenm)
<details>
<summary>Changes</summary>
Macos sincos stret was not well covered, and 64-bit wasn't
checked for vector sincos cases.
---
Patch is 29.10 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/166041.diff
2 Files Affected:
- (modified) llvm/test/CodeGen/X86/isel-llvm.sincos.ll (+133)
- (modified) llvm/test/CodeGen/X86/llvm.sincos.vec.ll (+321-83)
``````````diff
diff --git a/llvm/test/CodeGen/X86/isel-llvm.sincos.ll b/llvm/test/CodeGen/X86/isel-llvm.sincos.ll
index 065710f91457b..8576f8f149e9a 100644
--- a/llvm/test/CodeGen/X86/isel-llvm.sincos.ll
+++ b/llvm/test/CodeGen/X86/isel-llvm.sincos.ll
@@ -3,6 +3,9 @@
; RUN: llc < %s -mtriple=x86_64-linux-gnu -fast-isel | FileCheck %s --check-prefixes=X64,FASTISEL-X64
; RUN: llc < %s -mtriple=i686-linux-gnu -global-isel=0 -fast-isel=0 | FileCheck %s --check-prefixes=X86,SDAG-X86
; RUN: llc < %s -mtriple=x86_64-linux-gnu -global-isel=0 -fast-isel=0 | FileCheck %s --check-prefixes=X64,SDAG-X64
+; RUN: llc < %s -mtriple=x86_64-apple-macosx10.9.0 -mcpu=core2 | FileCheck %s --check-prefix=MACOS-SINCOS-STRET
+; RUN: llc < %s -mtriple=x86_64-apple-macosx10.8.0 -mcpu=core2 | FileCheck %s --check-prefix=MACOS-NOSINCOS-STRET
+
; TODO: The below RUN line will fails GISEL selection and will fallback to DAG selection due to lack of support for loads/stores in i686 mode, support is expected soon enough, for this reason the llvm/test/CodeGen/X86/GlobalISel/llvm.sincos.mir test is added for now because of the lack of support for i686 in GlobalISel.
; RUN: llc < %s -mtriple=i686-linux-gnu -global-isel=1 -global-isel-abort=2 | FileCheck %s --check-prefixes=GISEL-X86
; RUN: llc < %s -mtriple=x86_64-linux-gnu -global-isel=1 -global-isel-abort=1 | FileCheck %s --check-prefixes=GISEL-X64
@@ -34,6 +37,29 @@ define { float, float } @test_sincos_f32(float %Val) nounwind {
; X64-NEXT: popq %rax
; X64-NEXT: retq
;
+; MACOS-SINCOS-STRET-LABEL: test_sincos_f32:
+; MACOS-SINCOS-STRET: ## %bb.0:
+; MACOS-SINCOS-STRET-NEXT: pushq %rax
+; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret
+; MACOS-SINCOS-STRET-NEXT: movshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
+; MACOS-SINCOS-STRET-NEXT: popq %rax
+; MACOS-SINCOS-STRET-NEXT: retq
+;
+; MACOS-NOSINCOS-STRET-LABEL: test_sincos_f32:
+; MACOS-NOSINCOS-STRET: ## %bb.0:
+; MACOS-NOSINCOS-STRET-NEXT: pushq %rax
+; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rsp) ## 4-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: callq _sinf
+; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 4-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: movss (%rsp), %xmm0 ## 4-byte Reload
+; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero
+; MACOS-NOSINCOS-STRET-NEXT: callq _cosf
+; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, %xmm1
+; MACOS-NOSINCOS-STRET-NEXT: movss {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 4-byte Reload
+; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero
+; MACOS-NOSINCOS-STRET-NEXT: popq %rax
+; MACOS-NOSINCOS-STRET-NEXT: retq
+;
; GISEL-X86-LABEL: test_sincos_f32:
; GISEL-X86: # %bb.0:
; GISEL-X86-NEXT: subl $28, %esp
@@ -93,6 +119,28 @@ define { double, double } @test_sincos_f64(double %Val) nounwind {
; X64-NEXT: addq $24, %rsp
; X64-NEXT: retq
;
+; MACOS-SINCOS-STRET-LABEL: test_sincos_f64:
+; MACOS-SINCOS-STRET: ## %bb.0:
+; MACOS-SINCOS-STRET-NEXT: pushq %rax
+; MACOS-SINCOS-STRET-NEXT: callq ___sincos_stret
+; MACOS-SINCOS-STRET-NEXT: popq %rax
+; MACOS-SINCOS-STRET-NEXT: retq
+;
+; MACOS-NOSINCOS-STRET-LABEL: test_sincos_f64:
+; MACOS-NOSINCOS-STRET: ## %bb.0:
+; MACOS-NOSINCOS-STRET-NEXT: subq $24, %rsp
+; MACOS-NOSINCOS-STRET-NEXT: movsd %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 8-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: callq _sin
+; MACOS-NOSINCOS-STRET-NEXT: movsd %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 8-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: movsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 8-byte Reload
+; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero
+; MACOS-NOSINCOS-STRET-NEXT: callq _cos
+; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, %xmm1
+; MACOS-NOSINCOS-STRET-NEXT: movsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 8-byte Reload
+; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero
+; MACOS-NOSINCOS-STRET-NEXT: addq $24, %rsp
+; MACOS-NOSINCOS-STRET-NEXT: retq
+;
; GISEL-X86-LABEL: test_sincos_f64:
; GISEL-X86: # %bb.0:
; GISEL-X86-NEXT: subl $44, %esp
@@ -153,6 +201,40 @@ define { x86_fp80, x86_fp80 } @test_sincos_f80(x86_fp80 %Val) nounwind {
; X64-NEXT: addq $56, %rsp
; X64-NEXT: retq
;
+; MACOS-SINCOS-STRET-LABEL: test_sincos_f80:
+; MACOS-SINCOS-STRET: ## %bb.0:
+; MACOS-SINCOS-STRET-NEXT: subq $40, %rsp
+; MACOS-SINCOS-STRET-NEXT: fldt {{[0-9]+}}(%rsp)
+; MACOS-SINCOS-STRET-NEXT: fld %st(0)
+; MACOS-SINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill
+; MACOS-SINCOS-STRET-NEXT: fstpt (%rsp)
+; MACOS-SINCOS-STRET-NEXT: callq _cosl
+; MACOS-SINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill
+; MACOS-SINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload
+; MACOS-SINCOS-STRET-NEXT: fstpt (%rsp)
+; MACOS-SINCOS-STRET-NEXT: callq _sinl
+; MACOS-SINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload
+; MACOS-SINCOS-STRET-NEXT: fxch %st(1)
+; MACOS-SINCOS-STRET-NEXT: addq $40, %rsp
+; MACOS-SINCOS-STRET-NEXT: retq
+;
+; MACOS-NOSINCOS-STRET-LABEL: test_sincos_f80:
+; MACOS-NOSINCOS-STRET: ## %bb.0:
+; MACOS-NOSINCOS-STRET-NEXT: subq $40, %rsp
+; MACOS-NOSINCOS-STRET-NEXT: fldt {{[0-9]+}}(%rsp)
+; MACOS-NOSINCOS-STRET-NEXT: fld %st(0)
+; MACOS-NOSINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill
+; MACOS-NOSINCOS-STRET-NEXT: fstpt (%rsp)
+; MACOS-NOSINCOS-STRET-NEXT: callq _cosl
+; MACOS-NOSINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill
+; MACOS-NOSINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload
+; MACOS-NOSINCOS-STRET-NEXT: fstpt (%rsp)
+; MACOS-NOSINCOS-STRET-NEXT: callq _sinl
+; MACOS-NOSINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload
+; MACOS-NOSINCOS-STRET-NEXT: fxch %st(1)
+; MACOS-NOSINCOS-STRET-NEXT: addq $40, %rsp
+; MACOS-NOSINCOS-STRET-NEXT: retq
+;
; GISEL-X86-LABEL: test_sincos_f80:
; GISEL-X86: # %bb.0:
; GISEL-X86-NEXT: subl $60, %esp
@@ -288,6 +370,57 @@ define void @can_fold_with_call_in_chain(float %x, ptr noalias %a, ptr noalias %
; SDAG-X64-NEXT: popq %r14
; SDAG-X64-NEXT: retq
;
+; MACOS-SINCOS-STRET-LABEL: can_fold_with_call_in_chain:
+; MACOS-SINCOS-STRET: ## %bb.0: ## %entry
+; MACOS-SINCOS-STRET-NEXT: pushq %r14
+; MACOS-SINCOS-STRET-NEXT: pushq %rbx
+; MACOS-SINCOS-STRET-NEXT: subq $40, %rsp
+; MACOS-SINCOS-STRET-NEXT: movq %rsi, %rbx
+; MACOS-SINCOS-STRET-NEXT: movq %rdi, %r14
+; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret
+; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill
+; MACOS-SINCOS-STRET-NEXT: movshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
+; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; MACOS-SINCOS-STRET-NEXT: movq %r14, %rdi
+; MACOS-SINCOS-STRET-NEXT: movq %rbx, %rsi
+; MACOS-SINCOS-STRET-NEXT: callq _foo
+; MACOS-SINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload
+; MACOS-SINCOS-STRET-NEXT: movss %xmm0, (%r14)
+; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
+; MACOS-SINCOS-STRET-NEXT: movss %xmm0, (%rbx)
+; MACOS-SINCOS-STRET-NEXT: addq $40, %rsp
+; MACOS-SINCOS-STRET-NEXT: popq %rbx
+; MACOS-SINCOS-STRET-NEXT: popq %r14
+; MACOS-SINCOS-STRET-NEXT: retq
+;
+; MACOS-NOSINCOS-STRET-LABEL: can_fold_with_call_in_chain:
+; MACOS-NOSINCOS-STRET: ## %bb.0: ## %entry
+; MACOS-NOSINCOS-STRET-NEXT: pushq %r14
+; MACOS-NOSINCOS-STRET-NEXT: pushq %rbx
+; MACOS-NOSINCOS-STRET-NEXT: pushq %rax
+; MACOS-NOSINCOS-STRET-NEXT: movq %rsi, %rbx
+; MACOS-NOSINCOS-STRET-NEXT: movq %rdi, %r14
+; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rsp) ## 4-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: callq _sinf
+; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 4-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: movss (%rsp), %xmm0 ## 4-byte Reload
+; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero
+; MACOS-NOSINCOS-STRET-NEXT: callq _cosf
+; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rsp) ## 4-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: movq %r14, %rdi
+; MACOS-NOSINCOS-STRET-NEXT: movq %rbx, %rsi
+; MACOS-NOSINCOS-STRET-NEXT: callq _foo
+; MACOS-NOSINCOS-STRET-NEXT: movss {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 4-byte Reload
+; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero
+; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%r14)
+; MACOS-NOSINCOS-STRET-NEXT: movss (%rsp), %xmm0 ## 4-byte Reload
+; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero
+; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rbx)
+; MACOS-NOSINCOS-STRET-NEXT: addq $8, %rsp
+; MACOS-NOSINCOS-STRET-NEXT: popq %rbx
+; MACOS-NOSINCOS-STRET-NEXT: popq %r14
+; MACOS-NOSINCOS-STRET-NEXT: retq
+;
; GISEL-X86-LABEL: can_fold_with_call_in_chain:
; GISEL-X86: # %bb.0: # %entry
; GISEL-X86-NEXT: pushl %ebx
diff --git a/llvm/test/CodeGen/X86/llvm.sincos.vec.ll b/llvm/test/CodeGen/X86/llvm.sincos.vec.ll
index 834dd788ff7fb..9b02438952035 100644
--- a/llvm/test/CodeGen/X86/llvm.sincos.vec.ll
+++ b/llvm/test/CodeGen/X86/llvm.sincos.vec.ll
@@ -1,59 +1,213 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --no_x86_scrub_sp --version 5
-; RUN: llc < %s -mtriple=i386-unknown-linux-gnu | FileCheck %s
+; RUN: llc < %s -mtriple=i386-unknown-linux-gnu | FileCheck -check-prefix=X86 %s
+; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu | FileCheck -check-prefix=X64 %s
+; RUN: llc < %s -mtriple=x86_64-apple-macosx10.9.0 | FileCheck --check-prefix=MACOS-SINCOS-STRET %s
+; RUN: llc < %s -mtriple=x86_64-apple-macosx10.8.0 | FileCheck --check-prefix=MACOS-NOSINCOS-STRET %s
define void @test_sincos_v4f32(<4 x float> %x, ptr noalias %out_sin, ptr noalias %out_cos) nounwind {
-; CHECK-LABEL: test_sincos_v4f32:
-; CHECK: # %bb.0:
-; CHECK-NEXT: pushl %edi
-; CHECK-NEXT: pushl %esi
-; CHECK-NEXT: subl $52, %esp
-; CHECK-NEXT: movl 84(%esp), %esi
-; CHECK-NEXT: flds 76(%esp)
-; CHECK-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill
-; CHECK-NEXT: flds 64(%esp)
-; CHECK-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill
-; CHECK-NEXT: flds 72(%esp)
-; CHECK-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill
-; CHECK-NEXT: flds 68(%esp)
-; CHECK-NEXT: movl 80(%esp), %edi
-; CHECK-NEXT: leal 40(%esp), %eax
-; CHECK-NEXT: movl %eax, 8(%esp)
-; CHECK-NEXT: leal 4(%edi), %eax
-; CHECK-NEXT: movl %eax, 4(%esp)
-; CHECK-NEXT: fstps (%esp)
-; CHECK-NEXT: calll sincosf
-; CHECK-NEXT: leal 44(%esp), %eax
-; CHECK-NEXT: movl %eax, 8(%esp)
-; CHECK-NEXT: leal 8(%edi), %eax
-; CHECK-NEXT: movl %eax, 4(%esp)
-; CHECK-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload
-; CHECK-NEXT: fstps (%esp)
-; CHECK-NEXT: calll sincosf
-; CHECK-NEXT: leal 36(%esp), %eax
-; CHECK-NEXT: movl %eax, 8(%esp)
-; CHECK-NEXT: movl %edi, 4(%esp)
-; CHECK-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload
-; CHECK-NEXT: fstps (%esp)
-; CHECK-NEXT: calll sincosf
-; CHECK-NEXT: leal 48(%esp), %eax
-; CHECK-NEXT: movl %eax, 8(%esp)
-; CHECK-NEXT: addl $12, %edi
-; CHECK-NEXT: movl %edi, 4(%esp)
-; CHECK-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload
-; CHECK-NEXT: fstps (%esp)
-; CHECK-NEXT: calll sincosf
-; CHECK-NEXT: flds 36(%esp)
-; CHECK-NEXT: flds 40(%esp)
-; CHECK-NEXT: flds 44(%esp)
-; CHECK-NEXT: flds 48(%esp)
-; CHECK-NEXT: fstps 12(%esi)
-; CHECK-NEXT: fstps 8(%esi)
-; CHECK-NEXT: fstps 4(%esi)
-; CHECK-NEXT: fstps (%esi)
-; CHECK-NEXT: addl $52, %esp
-; CHECK-NEXT: popl %esi
-; CHECK-NEXT: popl %edi
-; CHECK-NEXT: retl
+; X86-LABEL: test_sincos_v4f32:
+; X86: # %bb.0:
+; X86-NEXT: pushl %edi
+; X86-NEXT: pushl %esi
+; X86-NEXT: subl $52, %esp
+; X86-NEXT: movl 84(%esp), %esi
+; X86-NEXT: flds 76(%esp)
+; X86-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill
+; X86-NEXT: flds 64(%esp)
+; X86-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill
+; X86-NEXT: flds 72(%esp)
+; X86-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill
+; X86-NEXT: flds 68(%esp)
+; X86-NEXT: movl 80(%esp), %edi
+; X86-NEXT: leal 40(%esp), %eax
+; X86-NEXT: movl %eax, 8(%esp)
+; X86-NEXT: leal 4(%edi), %eax
+; X86-NEXT: movl %eax, 4(%esp)
+; X86-NEXT: fstps (%esp)
+; X86-NEXT: calll sincosf
+; X86-NEXT: leal 44(%esp), %eax
+; X86-NEXT: movl %eax, 8(%esp)
+; X86-NEXT: leal 8(%edi), %eax
+; X86-NEXT: movl %eax, 4(%esp)
+; X86-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload
+; X86-NEXT: fstps (%esp)
+; X86-NEXT: calll sincosf
+; X86-NEXT: leal 36(%esp), %eax
+; X86-NEXT: movl %eax, 8(%esp)
+; X86-NEXT: movl %edi, 4(%esp)
+; X86-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload
+; X86-NEXT: fstps (%esp)
+; X86-NEXT: calll sincosf
+; X86-NEXT: leal 48(%esp), %eax
+; X86-NEXT: movl %eax, 8(%esp)
+; X86-NEXT: addl $12, %edi
+; X86-NEXT: movl %edi, 4(%esp)
+; X86-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload
+; X86-NEXT: fstps (%esp)
+; X86-NEXT: calll sincosf
+; X86-NEXT: flds 36(%esp)
+; X86-NEXT: flds 40(%esp)
+; X86-NEXT: flds 44(%esp)
+; X86-NEXT: flds 48(%esp)
+; X86-NEXT: fstps 12(%esi)
+; X86-NEXT: fstps 8(%esi)
+; X86-NEXT: fstps 4(%esi)
+; X86-NEXT: fstps (%esi)
+; X86-NEXT: addl $52, %esp
+; X86-NEXT: popl %esi
+; X86-NEXT: popl %edi
+; X86-NEXT: retl
+;
+; X64-LABEL: test_sincos_v4f32:
+; X64: # %bb.0:
+; X64-NEXT: pushq %r14
+; X64-NEXT: pushq %rbx
+; X64-NEXT: subq $56, %rsp
+; X64-NEXT: movq %rsi, %rbx
+; X64-NEXT: movq %rdi, %r14
+; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
+; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
+; X64-NEXT: leaq 4(%rsp), %rdi
+; X64-NEXT: movq %rsp, %rsi
+; X64-NEXT: callq sincosf at PLT
+; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
+; X64-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1]
+; X64-NEXT: leaq 12(%rsp), %rdi
+; X64-NEXT: leaq 8(%rsp), %rsi
+; X64-NEXT: callq sincosf at PLT
+; X64-NEXT: leaq 28(%rsp), %rdi
+; X64-NEXT: leaq 24(%rsp), %rsi
+; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
+; X64-NEXT: callq sincosf at PLT
+; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
+; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
+; X64-NEXT: leaq 20(%rsp), %rdi
+; X64-NEXT: leaq 16(%rsp), %rsi
+; X64-NEXT: callq sincosf at PLT
+; X64-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X64-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
+; X64-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X64-NEXT: movss {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; X64-NEXT: unpcklps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
+; X64-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; X64-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X64-NEXT: movss {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; X64-NEXT: unpcklps {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
+; X64-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X64-NEXT: movss {{.*#+}} xmm3 = mem[0],zero,zero,zero
+; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1]
+; X64-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm2[0]
+; X64-NEXT: movups %xmm1, (%r14)
+; X64-NEXT: movups %xmm0, (%rbx)
+; X64-NEXT: addq $56, %rsp
+; X64-NEXT: popq %rbx
+; X64-NEXT: popq %r14
+; X64-NEXT: retq
+;
+; MACOS-SINCOS-STRET-LABEL: test_sincos_v4f32:
+; MACOS-SINCOS-STRET: ## %bb.0:
+; MACOS-SINCOS-STRET-NEXT: pushq %r14
+; MACOS-SINCOS-STRET-NEXT: pushq %rbx
+; MACOS-SINCOS-STRET-NEXT: subq $104, %rsp
+; MACOS-SINCOS-STRET-NEXT: movq %rsi, %rbx
+; MACOS-SINCOS-STRET-NEXT: movq %rdi, %r14
+; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
+; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret
+; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
+; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill
+; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
+; MACOS-SINCOS-STRET-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1]
+; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret
+; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
+; MACOS-SINCOS-STRET-NEXT: unpcklps (%rsp), %xmm0 ## 16-byte Folded Reload
+; MACOS-SINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
+; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill
+; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
+; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret
+; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, %xmm1
+; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1],xmm0[1,1]
+; MACOS-SINCOS-STRET-NEXT: movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
+; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
+; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret
+; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
+; MACOS-SINCOS-STRET-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
+; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
+; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm2 ## 16-byte Reload
+; MACOS-SINCOS-STRET-NEXT: unpcklps {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1]
+; MACOS-SINCOS-STRET-NEXT: unpcklpd (%rsp), %xmm2 ## 16-byte Folded Reload
+; MACOS-SINCOS-STRET-NEXT: ## xmm2 = xmm2[0],mem[0]
+; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
+; MACOS-SINCOS-STRET-NEXT: unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload
+; MACOS-SINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
+; MACOS-SINCOS-STRET-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
+; MACOS-SINCOS-STRET-NEXT: movups %xmm1, (%r14)
+; MACOS-SINCOS-STRET-NEXT: movups %xmm2, (%rbx)
+; MACOS-SINCOS-STRET-NEXT: addq $104, %rsp
+; MACOS-SINCOS-STRET-NEXT: popq %rbx
+; MACOS-SINCOS-STRET-NEXT: popq %r14
+; MACOS-SINCOS-STRET-NEXT: retq
+;
+; MACOS-NOSINCOS-STRET-LABEL: test_sincos_v4f32:
+; MACOS-NOSINCOS-STRET: ## %bb.0:
+; MACOS-NOSINCOS-STRET-NEXT: pushq %r14
+; MACOS-NOSINCOS-STRET-NEXT: pushq %rbx
+; MACOS-NOSINCOS-STRET-NEXT: subq $104, %rsp
+; MACOS-NOSINCOS-STRET-NEXT: movq %rsi, %rbx
+; MACOS-NOSINCOS-STRET-NEXT: movq %rdi, %r14
+; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3]
+; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: callq _cosf
+; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload
+; MACOS-NOSINCOS-STRET-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1]
+; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: callq _cosf
+; MACOS-NOSINCOS-STRET-NEXT: unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload
+; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
+; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload
+; MACOS-NOSINCOS-STRET-NEXT: callq _cosf
+; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/166041
More information about the llvm-commits
mailing list