From 203783dc96b4c1a326b27d9a1c16150fc7dbf5db Mon Sep 17 00:00:00 2001 From: Simon Pilgrim Date: Sun, 28 Apr 2019 20:04:08 +0000 Subject: [PATCH] [X86][AVX] Add fast-hops target for add/fadd reduction tests llvm-svn: 359408 --- test/CodeGen/X86/vector-reduce-add-widen.ll | 479 +++++++---- test/CodeGen/X86/vector-reduce-add.ll | 507 ++++++++---- test/CodeGen/X86/vector-reduce-fadd-fast.ll | 829 +++++++++++++++----- test/CodeGen/X86/vector-reduce-fadd.ll | 13 +- 4 files changed, 1319 insertions(+), 509 deletions(-) diff --git a/test/CodeGen/X86/vector-reduce-add-widen.ll b/test/CodeGen/X86/vector-reduce-add-widen.ll index a59eabb177e..9a045f7bce5 100644 --- a/test/CodeGen/X86/vector-reduce-add-widen.ll +++ b/test/CodeGen/X86/vector-reduce-add-widen.ll @@ -1,10 +1,11 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py -; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE2 -; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE41 -; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1 -; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2 -; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW -; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512VL +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE,SSE2 +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE,SSE41 +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1,AVX1-SLOW +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx,+fast-hops | FileCheck %s --check-prefixes=AVX,AVX1,AVX1-FAST +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2 +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL ; ; vXi64 @@ -202,12 +203,25 @@ define i32 @test_v2i32(<2 x i32> %a0) { ; SSE-NEXT: movd %xmm1, %eax ; SSE-NEXT: retq ; -; AVX-LABEL: test_v2i32: -; AVX: # %bb.0: -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovd %xmm0, %eax -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v2i32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v2i32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v2i32: +; AVX2: # %bb.0: +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX2-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovd %xmm0, %eax +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v2i32: ; AVX512: # %bb.0: @@ -229,14 +243,31 @@ define i32 @test_v4i32(<4 x i32> %a0) { ; SSE-NEXT: movd %xmm0, %eax ; SSE-NEXT: retq ; -; AVX-LABEL: test_v4i32: -; AVX: # %bb.0: -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovd %xmm0, %eax -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4i32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4i32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4i32: +; AVX2: # %bb.0: +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX2-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX2-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovd %xmm0, %eax +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4i32: ; AVX512: # %bb.0: @@ -261,17 +292,28 @@ define i32 @test_v8i32(<8 x i32> %a0) { ; SSE-NEXT: movd %xmm0, %eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v8i32: -; AVX1: # %bb.0: -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v8i32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8i32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v8i32: ; AVX2: # %bb.0: @@ -313,20 +355,34 @@ define i32 @test_v16i32(<16 x i32> %a0) { ; SSE-NEXT: movd %xmm1, %eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v16i32: -; AVX1: # %bb.0: -; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3 -; AVX1-NEXT: vpaddd %xmm2, %xmm3, %xmm2 -; AVX1-NEXT: vpaddd %xmm2, %xmm1, %xmm1 -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v16i32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-SLOW-NEXT: vpaddd %xmm2, %xmm3, %xmm2 +; AVX1-SLOW-NEXT: vpaddd %xmm2, %xmm1, %xmm1 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v16i32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-FAST-NEXT: vpaddd %xmm2, %xmm3, %xmm2 +; AVX1-FAST-NEXT: vpaddd %xmm2, %xmm1, %xmm1 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v16i32: ; AVX2: # %bb.0: @@ -375,26 +431,46 @@ define i32 @test_v32i32(<32 x i32> %a0) { ; SSE-NEXT: movd %xmm1, %eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v32i32: -; AVX1: # %bb.0: -; AVX1-NEXT: vpaddd %xmm3, %xmm1, %xmm4 -; AVX1-NEXT: vextractf128 $1, %ymm3, %xmm3 -; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1 -; AVX1-NEXT: vpaddd %xmm3, %xmm1, %xmm1 -; AVX1-NEXT: vextractf128 $1, %ymm2, %xmm3 -; AVX1-NEXT: vpaddd %xmm1, %xmm3, %xmm1 -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3 -; AVX1-NEXT: vpaddd %xmm1, %xmm3, %xmm1 -; AVX1-NEXT: vpaddd %xmm4, %xmm2, %xmm2 -; AVX1-NEXT: vpaddd %xmm1, %xmm2, %xmm1 -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v32i32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpaddd %xmm3, %xmm1, %xmm4 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm3, %xmm3 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm1 +; AVX1-SLOW-NEXT: vpaddd %xmm3, %xmm1, %xmm1 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm2, %xmm3 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm3, %xmm1 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm3, %xmm1 +; AVX1-SLOW-NEXT: vpaddd %xmm4, %xmm2, %xmm2 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm2, %xmm1 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v32i32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpaddd %xmm3, %xmm1, %xmm4 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm3, %xmm3 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm1, %xmm1 +; AVX1-FAST-NEXT: vpaddd %xmm3, %xmm1, %xmm1 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm2, %xmm3 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm3, %xmm1 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm3, %xmm1 +; AVX1-FAST-NEXT: vpaddd %xmm4, %xmm2, %xmm2 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm2, %xmm1 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v32i32: ; AVX2: # %bb.0: @@ -443,13 +519,28 @@ define i16 @test_v2i16(<2 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX-LABEL: test_v2i16: -; AVX: # %bb.0: -; AVX-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovd %xmm0, %eax -; AVX-NEXT: # kill: def $ax killed $ax killed $eax -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v2i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v2i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v2i16: +; AVX2: # %bb.0: +; AVX2-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovd %xmm0, %eax +; AVX2-NEXT: # kill: def $ax killed $ax killed $eax +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v2i16: ; AVX512: # %bb.0: @@ -474,15 +565,34 @@ define i16 @test_v4i16(<4 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX-LABEL: test_v4i16: -; AVX: # %bb.0: -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovd %xmm0, %eax -; AVX-NEXT: # kill: def $ax killed $ax killed $eax -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4i16: +; AVX2: # %bb.0: +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovd %xmm0, %eax +; AVX2-NEXT: # kill: def $ax killed $ax killed $eax +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4i16: ; AVX512: # %bb.0: @@ -511,17 +621,40 @@ define i16 @test_v8i16(<8 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX-LABEL: test_v8i16: -; AVX: # %bb.0: -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovd %xmm0, %eax -; AVX-NEXT: # kill: def $ax killed $ax killed $eax -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v8i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v8i16: +; AVX2: # %bb.0: +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovd %xmm0, %eax +; AVX2-NEXT: # kill: def $ax killed $ax killed $eax +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v8i16: ; AVX512: # %bb.0: @@ -553,20 +686,34 @@ define i16 @test_v16i16(<16 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v16i16: -; AVX1: # %bb.0: -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: # kill: def $ax killed $ax killed $eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v16i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v16i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v16i16: ; AVX2: # %bb.0: @@ -618,23 +765,40 @@ define i16 @test_v32i16(<32 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v32i16: -; AVX1: # %bb.0: -; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3 -; AVX1-NEXT: vpaddw %xmm2, %xmm3, %xmm2 -; AVX1-NEXT: vpaddw %xmm2, %xmm1, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: # kill: def $ax killed $ax killed $eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v32i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-SLOW-NEXT: vpaddw %xmm2, %xmm3, %xmm2 +; AVX1-SLOW-NEXT: vpaddw %xmm2, %xmm1, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v32i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-FAST-NEXT: vpaddw %xmm2, %xmm3, %xmm2 +; AVX1-FAST-NEXT: vpaddw %xmm2, %xmm1, %xmm1 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v32i16: ; AVX2: # %bb.0: @@ -693,29 +857,52 @@ define i16 @test_v64i16(<64 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v64i16: -; AVX1: # %bb.0: -; AVX1-NEXT: vpaddw %xmm3, %xmm1, %xmm4 -; AVX1-NEXT: vextractf128 $1, %ymm3, %xmm3 -; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1 -; AVX1-NEXT: vpaddw %xmm3, %xmm1, %xmm1 -; AVX1-NEXT: vextractf128 $1, %ymm2, %xmm3 -; AVX1-NEXT: vpaddw %xmm1, %xmm3, %xmm1 -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3 -; AVX1-NEXT: vpaddw %xmm1, %xmm3, %xmm1 -; AVX1-NEXT: vpaddw %xmm4, %xmm2, %xmm2 -; AVX1-NEXT: vpaddw %xmm1, %xmm2, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: # kill: def $ax killed $ax killed $eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v64i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpaddw %xmm3, %xmm1, %xmm4 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm3, %xmm3 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm3, %xmm1, %xmm1 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm2, %xmm3 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm3, %xmm1 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm3, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm4, %xmm2, %xmm2 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm2, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v64i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpaddw %xmm3, %xmm1, %xmm4 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm3, %xmm3 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm1, %xmm1 +; AVX1-FAST-NEXT: vpaddw %xmm3, %xmm1, %xmm1 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm2, %xmm3 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm3, %xmm1 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm3, %xmm1 +; AVX1-FAST-NEXT: vpaddw %xmm4, %xmm2, %xmm2 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm2, %xmm1 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v64i16: ; AVX2: # %bb.0: diff --git a/test/CodeGen/X86/vector-reduce-add.ll b/test/CodeGen/X86/vector-reduce-add.ll index 4c98680dc4f..7abdf1cb037 100644 --- a/test/CodeGen/X86/vector-reduce-add.ll +++ b/test/CodeGen/X86/vector-reduce-add.ll @@ -1,10 +1,11 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE2 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE41 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512VL +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE,SSE2 +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE,SSE41 +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1,AVX1-SLOW +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+fast-hops | FileCheck %s --check-prefixes=AVX,AVX1,AVX1-FAST +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2 +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL ; ; vXi64 @@ -229,14 +230,31 @@ define i32 @test_v4i32(<4 x i32> %a0) { ; SSE-NEXT: movd %xmm0, %eax ; SSE-NEXT: retq ; -; AVX-LABEL: test_v4i32: -; AVX: # %bb.0: -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovd %xmm0, %eax -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4i32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4i32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4i32: +; AVX2: # %bb.0: +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX2-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX2-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovd %xmm0, %eax +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4i32: ; AVX512: # %bb.0: @@ -261,17 +279,28 @@ define i32 @test_v8i32(<8 x i32> %a0) { ; SSE-NEXT: movd %xmm0, %eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v8i32: -; AVX1: # %bb.0: -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v8i32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8i32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v8i32: ; AVX2: # %bb.0: @@ -313,20 +342,34 @@ define i32 @test_v16i32(<16 x i32> %a0) { ; SSE-NEXT: movd %xmm1, %eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v16i32: -; AVX1: # %bb.0: -; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3 -; AVX1-NEXT: vpaddd %xmm2, %xmm3, %xmm2 -; AVX1-NEXT: vpaddd %xmm2, %xmm1, %xmm1 -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v16i32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-SLOW-NEXT: vpaddd %xmm2, %xmm3, %xmm2 +; AVX1-SLOW-NEXT: vpaddd %xmm2, %xmm1, %xmm1 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v16i32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-FAST-NEXT: vpaddd %xmm2, %xmm3, %xmm2 +; AVX1-FAST-NEXT: vpaddd %xmm2, %xmm1, %xmm1 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v16i32: ; AVX2: # %bb.0: @@ -375,26 +418,46 @@ define i32 @test_v32i32(<32 x i32> %a0) { ; SSE-NEXT: movd %xmm1, %eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v32i32: -; AVX1: # %bb.0: -; AVX1-NEXT: vpaddd %xmm3, %xmm1, %xmm4 -; AVX1-NEXT: vextractf128 $1, %ymm3, %xmm3 -; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1 -; AVX1-NEXT: vpaddd %xmm3, %xmm1, %xmm1 -; AVX1-NEXT: vextractf128 $1, %ymm2, %xmm3 -; AVX1-NEXT: vpaddd %xmm1, %xmm3, %xmm1 -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3 -; AVX1-NEXT: vpaddd %xmm1, %xmm3, %xmm1 -; AVX1-NEXT: vpaddd %xmm4, %xmm2, %xmm2 -; AVX1-NEXT: vpaddd %xmm1, %xmm2, %xmm1 -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v32i32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpaddd %xmm3, %xmm1, %xmm4 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm3, %xmm3 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm1 +; AVX1-SLOW-NEXT: vpaddd %xmm3, %xmm1, %xmm1 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm2, %xmm3 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm3, %xmm1 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm3, %xmm1 +; AVX1-SLOW-NEXT: vpaddd %xmm4, %xmm2, %xmm2 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm2, %xmm1 +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v32i32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpaddd %xmm3, %xmm1, %xmm4 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm3, %xmm3 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm1, %xmm1 +; AVX1-FAST-NEXT: vpaddd %xmm3, %xmm1, %xmm1 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm2, %xmm3 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm3, %xmm1 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm3, %xmm1 +; AVX1-FAST-NEXT: vpaddd %xmm4, %xmm2, %xmm2 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm2, %xmm1 +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v32i32: ; AVX2: # %bb.0: @@ -472,15 +535,34 @@ define i16 @test_v4i16(<4 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX-LABEL: test_v4i16: -; AVX: # %bb.0: -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovd %xmm0, %eax -; AVX-NEXT: # kill: def $ax killed $ax killed $eax -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4i16: +; AVX2: # %bb.0: +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX2-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX2-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovd %xmm0, %eax +; AVX2-NEXT: # kill: def $ax killed $ax killed $eax +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4i16: ; AVX512: # %bb.0: @@ -509,17 +591,40 @@ define i16 @test_v8i16(<8 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX-LABEL: test_v8i16: -; AVX: # %bb.0: -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovd %xmm0, %eax -; AVX-NEXT: # kill: def $ax killed $ax killed $eax -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v8i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v8i16: +; AVX2: # %bb.0: +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovd %xmm0, %eax +; AVX2-NEXT: # kill: def $ax killed $ax killed $eax +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v8i16: ; AVX512: # %bb.0: @@ -551,20 +656,34 @@ define i16 @test_v16i16(<16 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v16i16: -; AVX1: # %bb.0: -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: # kill: def $ax killed $ax killed $eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v16i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v16i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v16i16: ; AVX2: # %bb.0: @@ -616,23 +735,40 @@ define i16 @test_v32i16(<32 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v32i16: -; AVX1: # %bb.0: -; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3 -; AVX1-NEXT: vpaddw %xmm2, %xmm3, %xmm2 -; AVX1-NEXT: vpaddw %xmm2, %xmm1, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: # kill: def $ax killed $ax killed $eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v32i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-SLOW-NEXT: vpaddw %xmm2, %xmm3, %xmm2 +; AVX1-SLOW-NEXT: vpaddw %xmm2, %xmm1, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v32i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-FAST-NEXT: vpaddw %xmm2, %xmm3, %xmm2 +; AVX1-FAST-NEXT: vpaddw %xmm2, %xmm1, %xmm1 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v32i16: ; AVX2: # %bb.0: @@ -691,29 +827,52 @@ define i16 @test_v64i16(<64 x i16> %a0) { ; SSE-NEXT: # kill: def $ax killed $ax killed $eax ; SSE-NEXT: retq ; -; AVX1-LABEL: test_v64i16: -; AVX1: # %bb.0: -; AVX1-NEXT: vpaddw %xmm3, %xmm1, %xmm4 -; AVX1-NEXT: vextractf128 $1, %ymm3, %xmm3 -; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1 -; AVX1-NEXT: vpaddw %xmm3, %xmm1, %xmm1 -; AVX1-NEXT: vextractf128 $1, %ymm2, %xmm3 -; AVX1-NEXT: vpaddw %xmm1, %xmm3, %xmm1 -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3 -; AVX1-NEXT: vpaddw %xmm1, %xmm3, %xmm1 -; AVX1-NEXT: vpaddw %xmm4, %xmm2, %xmm2 -; AVX1-NEXT: vpaddw %xmm1, %xmm2, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vmovd %xmm0, %eax -; AVX1-NEXT: # kill: def $ax killed $ax killed $eax -; AVX1-NEXT: vzeroupper -; AVX1-NEXT: retq +; AVX1-SLOW-LABEL: test_v64i16: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpaddw %xmm3, %xmm1, %xmm4 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm3, %xmm3 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm3, %xmm1, %xmm1 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm2, %xmm3 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm3, %xmm1 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm3, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm4, %xmm2, %xmm2 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm2, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovd %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v64i16: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpaddw %xmm3, %xmm1, %xmm4 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm3, %xmm3 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm1, %xmm1 +; AVX1-FAST-NEXT: vpaddw %xmm3, %xmm1, %xmm1 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm2, %xmm3 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm3, %xmm1 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm3 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm3, %xmm1 +; AVX1-FAST-NEXT: vpaddw %xmm4, %xmm2, %xmm2 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm2, %xmm1 +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vmovd %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $ax killed $ax killed $eax +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq ; ; AVX2-LABEL: test_v64i16: ; AVX2: # %bb.0: @@ -815,15 +974,34 @@ define i8 @test_v4i8(<4 x i8> %a0) { ; SSE41-NEXT: # kill: def $al killed $al killed $eax ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v4i8: -; AVX: # %bb.0: -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpextrb $0, %xmm0, %eax -; AVX-NEXT: # kill: def $al killed $al killed $eax -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4i8: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpextrb $0, %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $al killed $al killed $eax +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4i8: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpextrb $0, %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $al killed $al killed $eax +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4i8: +; AVX2: # %bb.0: +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX2-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX2-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpextrb $0, %xmm0, %eax +; AVX2-NEXT: # kill: def $al killed $al killed $eax +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4i8: ; AVX512: # %bb.0: @@ -865,17 +1043,40 @@ define i8 @test_v8i8(<8 x i8> %a0) { ; SSE41-NEXT: # kill: def $al killed $al killed $eax ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v8i8: -; AVX: # %bb.0: -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpsrld $16, %xmm0, %xmm1 -; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpextrb $0, %xmm0, %eax -; AVX-NEXT: # kill: def $al killed $al killed $eax -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v8i8: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX1-SLOW-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpextrb $0, %xmm0, %eax +; AVX1-SLOW-NEXT: # kill: def $al killed $al killed $eax +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8i8: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX1-FAST-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vphaddw %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpextrb $0, %xmm0, %eax +; AVX1-FAST-NEXT: # kill: def $al killed $al killed $eax +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v8i8: +; AVX2: # %bb.0: +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1] +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3] +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpsrld $16, %xmm0, %xmm1 +; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpextrb $0, %xmm0, %eax +; AVX2-NEXT: # kill: def $al killed $al killed $eax +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v8i8: ; AVX512: # %bb.0: diff --git a/test/CodeGen/X86/vector-reduce-fadd-fast.ll b/test/CodeGen/X86/vector-reduce-fadd-fast.ll index 7344a34ba8a..54513af0804 100644 --- a/test/CodeGen/X86/vector-reduce-fadd-fast.ll +++ b/test/CodeGen/X86/vector-reduce-fadd-fast.ll @@ -1,10 +1,11 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE2 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE41 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512VL +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE,SSE2 +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE,SSE41 +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1,AVX1-SLOW +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+fast-hops | FileCheck %s --check-prefixes=AVX,AVX1,AVX1-FAST +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2 +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL ; ; vXf32 (accum) @@ -24,11 +25,22 @@ define float @test_v2f32(float %a0, <2 x float> %a1) { ; SSE41-NEXT: addss %xmm1, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v2f32: -; AVX: # %bb.0: -; AVX-NEXT: vmovshdup {{.*#+}} xmm0 = xmm1[1,1,3,3] -; AVX-NEXT: vaddss %xmm0, %xmm1, %xmm0 -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v2f32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm0 = xmm1[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm0, %xmm1, %xmm0 +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v2f32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vhaddps %xmm1, %xmm1, %xmm0 +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v2f32: +; AVX2: # %bb.0: +; AVX2-NEXT: vmovshdup {{.*#+}} xmm0 = xmm1[1,1,3,3] +; AVX2-NEXT: vaddss %xmm0, %xmm1, %xmm0 +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v2f32: ; AVX512: # %bb.0: @@ -59,13 +71,28 @@ define float @test_v4f32(float %a0, <4 x float> %a1) { ; SSE41-NEXT: addss %xmm2, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v4f32: -; AVX: # %bb.0: -; AVX-NEXT: vpermilpd {{.*#+}} xmm0 = xmm1[1,0] -; AVX-NEXT: vaddps %xmm0, %xmm1, %xmm0 -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4f32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm0 = xmm1[1,0] +; AVX1-SLOW-NEXT: vaddps %xmm0, %xmm1, %xmm0 +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4f32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpermilpd {{.*#+}} xmm0 = xmm1[1,0] +; AVX1-FAST-NEXT: vaddps %xmm0, %xmm1, %xmm0 +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4f32: +; AVX2: # %bb.0: +; AVX2-NEXT: vpermilpd {{.*#+}} xmm0 = xmm1[1,0] +; AVX2-NEXT: vaddps %xmm0, %xmm1, %xmm0 +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4f32: ; AVX512: # %bb.0: @@ -100,16 +127,37 @@ define float @test_v8f32(float %a0, <8 x float> %a1) { ; SSE41-NEXT: addss %xmm2, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v8f32: -; AVX: # %bb.0: -; AVX-NEXT: vextractf128 $1, %ymm1, %xmm0 -; AVX-NEXT: vaddps %xmm0, %xmm1, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v8f32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm0 +; AVX1-SLOW-NEXT: vaddps %xmm0, %xmm1, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8f32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm1, %xmm0 +; AVX1-FAST-NEXT: vaddps %xmm0, %xmm1, %xmm0 +; AVX1-FAST-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v8f32: +; AVX2: # %bb.0: +; AVX2-NEXT: vextractf128 $1, %ymm1, %xmm0 +; AVX2-NEXT: vaddps %xmm0, %xmm1, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v8f32: ; AVX512: # %bb.0: @@ -151,17 +199,40 @@ define float @test_v16f32(float %a0, <16 x float> %a1) { ; SSE41-NEXT: addss %xmm2, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v16f32: -; AVX: # %bb.0: -; AVX-NEXT: vaddps %ymm2, %ymm1, %ymm0 -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v16f32: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vaddps %ymm2, %ymm1, %ymm0 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v16f32: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vaddps %ymm2, %ymm1, %ymm0 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v16f32: +; AVX2: # %bb.0: +; AVX2-NEXT: vaddps %ymm2, %ymm1, %ymm0 +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v16f32: ; AVX512: # %bb.0: @@ -198,11 +269,22 @@ define float @test_v2f32_zero(<2 x float> %a0) { ; SSE41-NEXT: addss %xmm1, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v2f32_zero: -; AVX: # %bb.0: -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v2f32_zero: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v2f32_zero: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v2f32_zero: +; AVX2: # %bb.0: +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v2f32_zero: ; AVX512: # %bb.0: @@ -234,13 +316,28 @@ define float @test_v4f32_zero(<4 x float> %a0) { ; SSE41-NEXT: movaps %xmm1, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v4f32_zero: -; AVX: # %bb.0: -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4f32_zero: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4f32_zero: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4f32_zero: +; AVX2: # %bb.0: +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4f32_zero: ; AVX512: # %bb.0: @@ -276,16 +373,37 @@ define float @test_v8f32_zero(<8 x float> %a0) { ; SSE41-NEXT: movaps %xmm1, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v8f32_zero: -; AVX: # %bb.0: -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v8f32_zero: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8f32_zero: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v8f32_zero: +; AVX2: # %bb.0: +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v8f32_zero: ; AVX512: # %bb.0: @@ -328,17 +446,40 @@ define float @test_v16f32_zero(<16 x float> %a0) { ; SSE41-NEXT: movaps %xmm1, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v16f32_zero: -; AVX: # %bb.0: -; AVX-NEXT: vaddps %ymm1, %ymm0, %ymm0 -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v16f32_zero: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vaddps %ymm1, %ymm0, %ymm0 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v16f32_zero: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vaddps %ymm1, %ymm0, %ymm0 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v16f32_zero: +; AVX2: # %bb.0: +; AVX2-NEXT: vaddps %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v16f32_zero: ; AVX512: # %bb.0: @@ -375,11 +516,22 @@ define float @test_v2f32_undef(<2 x float> %a0) { ; SSE41-NEXT: addss %xmm1, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v2f32_undef: -; AVX: # %bb.0: -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v2f32_undef: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v2f32_undef: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v2f32_undef: +; AVX2: # %bb.0: +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v2f32_undef: ; AVX512: # %bb.0: @@ -411,13 +563,28 @@ define float @test_v4f32_undef(<4 x float> %a0) { ; SSE41-NEXT: movaps %xmm1, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v4f32_undef: -; AVX: # %bb.0: -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4f32_undef: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4f32_undef: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4f32_undef: +; AVX2: # %bb.0: +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4f32_undef: ; AVX512: # %bb.0: @@ -453,16 +620,37 @@ define float @test_v8f32_undef(<8 x float> %a0) { ; SSE41-NEXT: movaps %xmm1, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v8f32_undef: -; AVX: # %bb.0: -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v8f32_undef: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8f32_undef: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v8f32_undef: +; AVX2: # %bb.0: +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v8f32_undef: ; AVX512: # %bb.0: @@ -505,17 +693,40 @@ define float @test_v16f32_undef(<16 x float> %a0) { ; SSE41-NEXT: movaps %xmm1, %xmm0 ; SSE41-NEXT: retq ; -; AVX-LABEL: test_v16f32_undef: -; AVX: # %bb.0: -; AVX-NEXT: vaddps %ymm1, %ymm0, %ymm0 -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] -; AVX-NEXT: vaddss %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v16f32_undef: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vaddps %ymm1, %ymm0, %ymm0 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX1-SLOW-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v16f32_undef: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vaddps %ymm1, %ymm0, %ymm0 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-FAST-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddps %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v16f32_undef: +; AVX2: # %bb.0: +; AVX2-NEXT: vaddps %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; AVX2-NEXT: vaddss %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v16f32_undef: ; AVX512: # %bb.0: @@ -545,11 +756,22 @@ define double @test_v2f64(double %a0, <2 x double> %a1) { ; SSE-NEXT: addsd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v2f64: -; AVX: # %bb.0: -; AVX-NEXT: vpermilpd {{.*#+}} xmm0 = xmm1[1,0] -; AVX-NEXT: vaddsd %xmm0, %xmm1, %xmm0 -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v2f64: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm0 = xmm1[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm0, %xmm1, %xmm0 +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v2f64: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vhaddpd %xmm1, %xmm1, %xmm0 +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v2f64: +; AVX2: # %bb.0: +; AVX2-NEXT: vpermilpd {{.*#+}} xmm0 = xmm1[1,0] +; AVX2-NEXT: vaddsd %xmm0, %xmm1, %xmm0 +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v2f64: ; AVX512: # %bb.0: @@ -569,14 +791,31 @@ define double @test_v4f64(double %a0, <4 x double> %a1) { ; SSE-NEXT: addsd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v4f64: -; AVX: # %bb.0: -; AVX-NEXT: vextractf128 $1, %ymm1, %xmm0 -; AVX-NEXT: vaddpd %xmm0, %xmm1, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4f64: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm0 +; AVX1-SLOW-NEXT: vaddpd %xmm0, %xmm1, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4f64: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm1, %xmm0 +; AVX1-FAST-NEXT: vaddpd %xmm0, %xmm1, %xmm0 +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4f64: +; AVX2: # %bb.0: +; AVX2-NEXT: vextractf128 $1, %ymm1, %xmm0 +; AVX2-NEXT: vaddpd %xmm0, %xmm1, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4f64: ; AVX512: # %bb.0: @@ -601,15 +840,34 @@ define double @test_v8f64(double %a0, <8 x double> %a1) { ; SSE-NEXT: addsd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v8f64: -; AVX: # %bb.0: -; AVX-NEXT: vaddpd %ymm2, %ymm1, %ymm0 -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddpd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v8f64: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vaddpd %ymm2, %ymm1, %ymm0 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8f64: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vaddpd %ymm2, %ymm1, %ymm0 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v8f64: +; AVX2: # %bb.0: +; AVX2-NEXT: vaddpd %ymm2, %ymm1, %ymm0 +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v8f64: ; AVX512: # %bb.0: @@ -640,17 +898,40 @@ define double @test_v16f64(double %a0, <16 x double> %a1) { ; SSE-NEXT: addsd %xmm4, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v16f64: -; AVX: # %bb.0: -; AVX-NEXT: vaddpd %ymm4, %ymm2, %ymm0 -; AVX-NEXT: vaddpd %ymm3, %ymm1, %ymm1 -; AVX-NEXT: vaddpd %ymm0, %ymm1, %ymm0 -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddpd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v16f64: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vaddpd %ymm4, %ymm2, %ymm0 +; AVX1-SLOW-NEXT: vaddpd %ymm3, %ymm1, %ymm1 +; AVX1-SLOW-NEXT: vaddpd %ymm0, %ymm1, %ymm0 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v16f64: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vaddpd %ymm4, %ymm2, %ymm0 +; AVX1-FAST-NEXT: vaddpd %ymm3, %ymm1, %ymm1 +; AVX1-FAST-NEXT: vaddpd %ymm0, %ymm1, %ymm0 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v16f64: +; AVX2: # %bb.0: +; AVX2-NEXT: vaddpd %ymm4, %ymm2, %ymm0 +; AVX2-NEXT: vaddpd %ymm3, %ymm1, %ymm1 +; AVX2-NEXT: vaddpd %ymm0, %ymm1, %ymm0 +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v16f64: ; AVX512: # %bb.0: @@ -680,11 +961,22 @@ define double @test_v2f64_zero(<2 x double> %a0) { ; SSE-NEXT: movapd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v2f64_zero: -; AVX: # %bb.0: -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v2f64_zero: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v2f64_zero: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v2f64_zero: +; AVX2: # %bb.0: +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v2f64_zero: ; AVX512: # %bb.0: @@ -705,14 +997,31 @@ define double @test_v4f64_zero(<4 x double> %a0) { ; SSE-NEXT: movapd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v4f64_zero: -; AVX: # %bb.0: -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddpd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4f64_zero: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4f64_zero: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4f64_zero: +; AVX2: # %bb.0: +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4f64_zero: ; AVX512: # %bb.0: @@ -738,15 +1047,34 @@ define double @test_v8f64_zero(<8 x double> %a0) { ; SSE-NEXT: movapd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v8f64_zero: -; AVX: # %bb.0: -; AVX-NEXT: vaddpd %ymm1, %ymm0, %ymm0 -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddpd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v8f64_zero: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8f64_zero: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v8f64_zero: +; AVX2: # %bb.0: +; AVX2-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v8f64_zero: ; AVX512: # %bb.0: @@ -777,17 +1105,40 @@ define double @test_v16f64_zero(<16 x double> %a0) { ; SSE-NEXT: addsd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v16f64_zero: -; AVX: # %bb.0: -; AVX-NEXT: vaddpd %ymm3, %ymm1, %ymm1 -; AVX-NEXT: vaddpd %ymm2, %ymm0, %ymm0 -; AVX-NEXT: vaddpd %ymm1, %ymm0, %ymm0 -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddpd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v16f64_zero: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vaddpd %ymm3, %ymm1, %ymm1 +; AVX1-SLOW-NEXT: vaddpd %ymm2, %ymm0, %ymm0 +; AVX1-SLOW-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v16f64_zero: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vaddpd %ymm3, %ymm1, %ymm1 +; AVX1-FAST-NEXT: vaddpd %ymm2, %ymm0, %ymm0 +; AVX1-FAST-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v16f64_zero: +; AVX2: # %bb.0: +; AVX2-NEXT: vaddpd %ymm3, %ymm1, %ymm1 +; AVX2-NEXT: vaddpd %ymm2, %ymm0, %ymm0 +; AVX2-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v16f64_zero: ; AVX512: # %bb.0: @@ -817,11 +1168,22 @@ define double @test_v2f64_undef(<2 x double> %a0) { ; SSE-NEXT: movapd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v2f64_undef: -; AVX: # %bb.0: -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v2f64_undef: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v2f64_undef: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v2f64_undef: +; AVX2: # %bb.0: +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v2f64_undef: ; AVX512: # %bb.0: @@ -842,14 +1204,31 @@ define double @test_v4f64_undef(<4 x double> %a0) { ; SSE-NEXT: movapd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v4f64_undef: -; AVX: # %bb.0: -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddpd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v4f64_undef: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v4f64_undef: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v4f64_undef: +; AVX2: # %bb.0: +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v4f64_undef: ; AVX512: # %bb.0: @@ -875,15 +1254,34 @@ define double @test_v8f64_undef(<8 x double> %a0) { ; SSE-NEXT: movapd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v8f64_undef: -; AVX: # %bb.0: -; AVX-NEXT: vaddpd %ymm1, %ymm0, %ymm0 -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddpd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v8f64_undef: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v8f64_undef: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v8f64_undef: +; AVX2: # %bb.0: +; AVX2-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v8f64_undef: ; AVX512: # %bb.0: @@ -914,17 +1312,40 @@ define double @test_v16f64_undef(<16 x double> %a0) { ; SSE-NEXT: addsd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: test_v16f64_undef: -; AVX: # %bb.0: -; AVX-NEXT: vaddpd %ymm3, %ymm1, %ymm1 -; AVX-NEXT: vaddpd %ymm2, %ymm0, %ymm0 -; AVX-NEXT: vaddpd %ymm1, %ymm0, %ymm0 -; AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 -; AVX-NEXT: vaddpd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] -; AVX-NEXT: vaddsd %xmm1, %xmm0, %xmm0 -; AVX-NEXT: vzeroupper -; AVX-NEXT: retq +; AVX1-SLOW-LABEL: test_v16f64_undef: +; AVX1-SLOW: # %bb.0: +; AVX1-SLOW-NEXT: vaddpd %ymm3, %ymm1, %ymm1 +; AVX1-SLOW-NEXT: vaddpd %ymm2, %ymm0, %ymm0 +; AVX1-SLOW-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX1-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-SLOW-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX1-SLOW-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX1-SLOW-NEXT: vzeroupper +; AVX1-SLOW-NEXT: retq +; +; AVX1-FAST-LABEL: test_v16f64_undef: +; AVX1-FAST: # %bb.0: +; AVX1-FAST-NEXT: vaddpd %ymm3, %ymm1, %ymm1 +; AVX1-FAST-NEXT: vaddpd %ymm2, %ymm0, %ymm0 +; AVX1-FAST-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX1-FAST-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-FAST-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vhaddpd %xmm0, %xmm0, %xmm0 +; AVX1-FAST-NEXT: vzeroupper +; AVX1-FAST-NEXT: retq +; +; AVX2-LABEL: test_v16f64_undef: +; AVX2: # %bb.0: +; AVX2-NEXT: vaddpd %ymm3, %ymm1, %ymm1 +; AVX2-NEXT: vaddpd %ymm2, %ymm0, %ymm0 +; AVX2-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vaddpd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0] +; AVX2-NEXT: vaddsd %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq ; ; AVX512-LABEL: test_v16f64_undef: ; AVX512: # %bb.0: diff --git a/test/CodeGen/X86/vector-reduce-fadd.ll b/test/CodeGen/X86/vector-reduce-fadd.ll index f025c57d6b1..cae5f4c1252 100644 --- a/test/CodeGen/X86/vector-reduce-fadd.ll +++ b/test/CodeGen/X86/vector-reduce-fadd.ll @@ -1,10 +1,11 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE2 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE41 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512VL +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE,SSE2 +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE,SSE41 +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1,AVX1-SLOW +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+fast-hops | FileCheck %s --check-prefixes=AVX,AVX1,AVX1-FAST +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2 +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL ; ; vXf32 (accum)