mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2025-01-31 20:51:52 +01:00
62c0dbfd59
These typically match the FMA3 equivalents, although the multiply operands sometimes get flipped due to the FMA3 permute variants.
21 lines
903 B
LLVM
21 lines
903 B
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -mcpu=bdver2 -mattr=-fma -mtriple=i686-apple-darwin < %s | FileCheck %s
|
|
; RUN: llc -mcpu=bdver2 -mattr=-fma,-fma4 -mtriple=i686-apple-darwin < %s | FileCheck %s --check-prefix=CHECK-NOFMA
|
|
|
|
define <3 x float> @fmafunc(<3 x float> %a, <3 x float> %b, <3 x float> %c) {
|
|
; CHECK-LABEL: fmafunc:
|
|
; CHECK: ## %bb.0:
|
|
; CHECK-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
|
|
; CHECK-NEXT: retl
|
|
;
|
|
; CHECK-NOFMA-LABEL: fmafunc:
|
|
; CHECK-NOFMA: ## %bb.0:
|
|
; CHECK-NOFMA-NEXT: vmulps %xmm1, %xmm0, %xmm0
|
|
; CHECK-NOFMA-NEXT: vaddps %xmm2, %xmm0, %xmm0
|
|
; CHECK-NOFMA-NEXT: retl
|
|
%ret = tail call <3 x float> @llvm.fmuladd.v3f32(<3 x float> %a, <3 x float> %b, <3 x float> %c)
|
|
ret <3 x float> %ret
|
|
}
|
|
|
|
declare <3 x float> @llvm.fmuladd.v3f32(<3 x float>, <3 x float>, <3 x float>) nounwind readnone
|