71 lines · plain
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py2; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx | FileCheck %s --check-prefix=AVX3; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx2 | FileCheck %s --check-prefix=AVX4 5; Prefer a blend instruction to a vinsert128 instruction because blends6; are simpler (no lane changes) and therefore will have equal or better7; performance.8 9define <8 x float> @castA(<4 x float> %m) nounwind uwtable readnone ssp {10; AVX-LABEL: castA:11; AVX: ## %bb.0:12; AVX-NEXT: vmovaps %xmm0, %xmm013; AVX-NEXT: retq14 %shuffle.i = shufflevector <4 x float> %m, <4 x float> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 4, i32 4, i32 4>15 ret <8 x float> %shuffle.i16}17 18define <4 x double> @castB(<2 x double> %m) nounwind uwtable readnone ssp {19; AVX-LABEL: castB:20; AVX: ## %bb.0:21; AVX-NEXT: vmovaps %xmm0, %xmm022; AVX-NEXT: retq23 %shuffle.i = shufflevector <2 x double> %m, <2 x double> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 2>24 ret <4 x double> %shuffle.i25}26 27; AVX2 is needed for integer types.28 29define <4 x i64> @castC(<2 x i64> %m) nounwind uwtable readnone ssp {30; AVX-LABEL: castC:31; AVX: ## %bb.0:32; AVX-NEXT: vmovaps %xmm0, %xmm033; AVX-NEXT: retq34 %shuffle.i = shufflevector <2 x i64> %m, <2 x i64> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 2>35 ret <4 x i64> %shuffle.i36}37 38; The next three tests don't need any shuffling. There may or may not be a39; vzeroupper before the return, so just check for the absence of shuffles.40 41define <4 x float> @castD(<8 x float> %m) nounwind uwtable readnone ssp {42; AVX-LABEL: castD:43; AVX: ## %bb.0:44; AVX-NEXT: ## kill: def $xmm0 killed $xmm0 killed $ymm045; AVX-NEXT: vzeroupper46; AVX-NEXT: retq47 %shuffle.i = shufflevector <8 x float> %m, <8 x float> %m, <4 x i32> <i32 0, i32 1, i32 2, i32 3>48 ret <4 x float> %shuffle.i49}50 51define <2 x i64> @castE(<4 x i64> %m) nounwind uwtable readnone ssp {52; AVX-LABEL: castE:53; AVX: ## %bb.0:54; AVX-NEXT: ## kill: def $xmm0 killed $xmm0 killed $ymm055; AVX-NEXT: vzeroupper56; AVX-NEXT: retq57 %shuffle.i = shufflevector <4 x i64> %m, <4 x i64> %m, <2 x i32> <i32 0, i32 1>58 ret <2 x i64> %shuffle.i59}60 61define <2 x double> @castF(<4 x double> %m) nounwind uwtable readnone ssp {62; AVX-LABEL: castF:63; AVX: ## %bb.0:64; AVX-NEXT: ## kill: def $xmm0 killed $xmm0 killed $ymm065; AVX-NEXT: vzeroupper66; AVX-NEXT: retq67 %shuffle.i = shufflevector <4 x double> %m, <4 x double> %m, <2 x i32> <i32 0, i32 1>68 ret <2 x double> %shuffle.i69}70 71