From: Simon Pilgrim Date: Tue, 17 Nov 2015 21:19:45 +0000 (+0000) Subject: [X86][SSE] Share AVX1/AVX2 shuffle tests with AVX512 where possible X-Git-Url: http://plrg.eecs.uci.edu/git/?a=commitdiff_plain;h=d82a1c70eb2d088855d36c829dbaa81aff8c80ac;p=oota-llvm.git [X86][SSE] Share AVX1/AVX2 shuffle tests with AVX512 where possible git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@253379 91177308-0d34-0410-b5e6-96231b3b80d8 --- diff --git a/test/CodeGen/X86/vector-shuffle-128-v2.ll b/test/CodeGen/X86/vector-shuffle-128-v2.ll index 4d6acbef989..ad33ef7645b 100644 --- a/test/CodeGen/X86/vector-shuffle-128-v2.ll +++ b/test/CodeGen/X86/vector-shuffle-128-v2.ll @@ -4,7 +4,7 @@ ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+sse4.1 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE41 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512vl | FileCheck %s --check-prefix=AVX512VL +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX512VL target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" target triple = "x86_64-unknown-unknown" @@ -42,11 +42,6 @@ define <2 x i64> @shuffle_v2i64_10(<2 x i64> %a, <2 x i64> %b) { ; AVX: # BB#0: ; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2i64_10: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> ret <2 x i64> %shuffle } @@ -60,11 +55,6 @@ define <2 x i64> @shuffle_v2i64_11(<2 x i64> %a, <2 x i64> %b) { ; AVX: # BB#0: ; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,2,3] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2i64_11: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,2,3] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> ret <2 x i64> %shuffle } @@ -101,11 +91,6 @@ define <2 x i64> @shuffle_v2i64_32(<2 x i64> %a, <2 x i64> %b) { ; AVX: # BB#0: ; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm1[2,3,0,1] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2i64_32: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vpshufd {{.*#+}} xmm0 = xmm1[2,3,0,1] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> ret <2 x i64> %shuffle } @@ -119,11 +104,6 @@ define <2 x i64> @shuffle_v2i64_33(<2 x i64> %a, <2 x i64> %b) { ; AVX: # BB#0: ; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm1[2,3,2,3] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2i64_33: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vpshufd {{.*#+}} xmm0 = xmm1[2,3,2,3] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> ret <2 x i64> %shuffle } @@ -153,11 +133,6 @@ define <2 x double> @shuffle_v2f64_00(<2 x double> %a, <2 x double> %b) { ; AVX: # BB#0: ; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2f64_00: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> ret <2 x double> %shuffle } @@ -167,10 +142,15 @@ define <2 x double> @shuffle_v2f64_10(<2 x double> %a, <2 x double> %b) { ; SSE-NEXT: shufpd {{.*#+}} xmm0 = xmm0[1,0] ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2f64_10: -; AVX: # BB#0: -; AVX-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2f64_10: +; AVX1: # BB#0: +; AVX1-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2f64_10: +; AVX2: # BB#0: +; AVX2-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2f64_10: ; AVX512VL: # BB#0: @@ -186,10 +166,15 @@ define <2 x double> @shuffle_v2f64_11(<2 x double> %a, <2 x double> %b) { ; SSE-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2f64_11: -; AVX: # BB#0: -; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2f64_11: +; AVX1: # BB#0: +; AVX1-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2f64_11: +; AVX2: # BB#0: +; AVX2-NEXT: vmovhlps {{.*#+}} xmm0 = xmm0[1,1] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2f64_11: ; AVX512VL: # BB#0: @@ -224,11 +209,6 @@ define <2 x double> @shuffle_v2f64_22(<2 x double> %a, <2 x double> %b) { ; AVX: # BB#0: ; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm1[0,0] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2f64_22: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vmovddup {{.*#+}} xmm0 = xmm1[0,0] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> ret <2 x double> %shuffle } @@ -239,10 +219,15 @@ define <2 x double> @shuffle_v2f64_32(<2 x double> %a, <2 x double> %b) { ; SSE-NEXT: movapd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2f64_32: -; AVX: # BB#0: -; AVX-NEXT: vpermilpd {{.*#+}} xmm0 = xmm1[1,0] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2f64_32: +; AVX1: # BB#0: +; AVX1-NEXT: vpermilpd {{.*#+}} xmm0 = xmm1[1,0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2f64_32: +; AVX2: # BB#0: +; AVX2-NEXT: vpermilpd {{.*#+}} xmm0 = xmm1[1,0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2f64_32: ; AVX512VL: # BB#0: @@ -259,10 +244,15 @@ define <2 x double> @shuffle_v2f64_33(<2 x double> %a, <2 x double> %b) { ; SSE-NEXT: movaps %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2f64_33: -; AVX: # BB#0: -; AVX-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1,1] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2f64_33: +; AVX1: # BB#0: +; AVX1-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1,1] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2f64_33: +; AVX2: # BB#0: +; AVX2-NEXT: vmovhlps {{.*#+}} xmm0 = xmm1[1,1] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2f64_33: ; AVX512VL: # BB#0: @@ -299,11 +289,6 @@ define <2 x double> @shuffle_v2f64_03(<2 x double> %a, <2 x double> %b) { ; AVX: # BB#0: ; AVX-NEXT: vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2f64_03: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> ret <2 x double> %shuffle } @@ -332,11 +317,6 @@ define <2 x double> @shuffle_v2f64_21(<2 x double> %a, <2 x double> %b) { ; AVX: # BB#0: ; AVX-NEXT: vblendpd {{.*#+}} xmm0 = xmm1[0],xmm0[1] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2f64_21: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vblendpd {{.*#+}} xmm0 = xmm1[0],xmm0[1] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> ret <2 x double> %shuffle } @@ -348,10 +328,15 @@ define <2 x i64> @shuffle_v2i64_02(<2 x i64> %a, <2 x i64> %b) { ; SSE-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2i64_02: -; AVX: # BB#0: -; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2i64_02: +; AVX1: # BB#0: +; AVX1-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2i64_02: +; AVX2: # BB#0: +; AVX2-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2i64_02: ; AVX512VL: # BB#0: @@ -367,10 +352,15 @@ define <2 x i64> @shuffle_v2i64_02_copy(<2 x i64> %nonce, <2 x i64> %a, <2 x i64 ; SSE-NEXT: movdqa %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2i64_02_copy: -; AVX: # BB#0: -; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm2[0] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2i64_02_copy: +; AVX1: # BB#0: +; AVX1-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm2[0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2i64_02_copy: +; AVX2: # BB#0: +; AVX2-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm2[0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2i64_02_copy: ; AVX512VL: # BB#0: @@ -489,11 +479,6 @@ define <2 x i64> @shuffle_v2i64_12(<2 x i64> %a, <2 x i64> %b) { ; AVX: # BB#0: ; AVX-NEXT: vpalignr {{.*#+}} xmm0 = xmm0[8,9,10,11,12,13,14,15],xmm1[0,1,2,3,4,5,6,7] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2i64_12: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vpalignr {{.*#+}} xmm0 = xmm0[8,9,10,11,12,13,14,15],xmm1[0,1,2,3,4,5,6,7] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> ret <2 x i64> %shuffle } @@ -526,11 +511,6 @@ define <2 x i64> @shuffle_v2i64_12_copy(<2 x i64> %nonce, <2 x i64> %a, <2 x i64 ; AVX: # BB#0: ; AVX-NEXT: vpalignr {{.*#+}} xmm0 = xmm1[8,9,10,11,12,13,14,15],xmm2[0,1,2,3,4,5,6,7] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2i64_12_copy: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vpalignr {{.*#+}} xmm0 = xmm1[8,9,10,11,12,13,14,15],xmm2[0,1,2,3,4,5,6,7] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> ret <2 x i64> %shuffle } @@ -540,10 +520,15 @@ define <2 x i64> @shuffle_v2i64_13(<2 x i64> %a, <2 x i64> %b) { ; SSE-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1] ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2i64_13: -; AVX: # BB#0: -; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2i64_13: +; AVX1: # BB#0: +; AVX1-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2i64_13: +; AVX2: # BB#0: +; AVX2-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2i64_13: ; AVX512VL: # BB#0: @@ -559,10 +544,15 @@ define <2 x i64> @shuffle_v2i64_13_copy(<2 x i64> %nonce, <2 x i64> %a, <2 x i64 ; SSE-NEXT: movdqa %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2i64_13_copy: -; AVX: # BB#0: -; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm2[1] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2i64_13_copy: +; AVX1: # BB#0: +; AVX1-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm2[1] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2i64_13_copy: +; AVX2: # BB#0: +; AVX2-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm2[1] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2i64_13_copy: ; AVX512VL: # BB#0: @@ -578,10 +568,15 @@ define <2 x i64> @shuffle_v2i64_20(<2 x i64> %a, <2 x i64> %b) { ; SSE-NEXT: movdqa %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2i64_20: -; AVX: # BB#0: -; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2i64_20: +; AVX1: # BB#0: +; AVX1-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2i64_20: +; AVX2: # BB#0: +; AVX2-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2i64_20: ; AVX512VL: # BB#0: @@ -597,10 +592,15 @@ define <2 x i64> @shuffle_v2i64_20_copy(<2 x i64> %nonce, <2 x i64> %a, <2 x i64 ; SSE-NEXT: movdqa %xmm2, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2i64_20_copy: -; AVX: # BB#0: -; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm2[0],xmm1[0] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2i64_20_copy: +; AVX1: # BB#0: +; AVX1-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm2[0],xmm1[0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2i64_20_copy: +; AVX2: # BB#0: +; AVX2-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm2[0],xmm1[0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2i64_20_copy: ; AVX512VL: # BB#0: @@ -716,11 +716,6 @@ define <2 x i64> @shuffle_v2i64_30(<2 x i64> %a, <2 x i64> %b) { ; AVX: # BB#0: ; AVX-NEXT: vpalignr {{.*#+}} xmm0 = xmm1[8,9,10,11,12,13,14,15],xmm0[0,1,2,3,4,5,6,7] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2i64_30: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vpalignr {{.*#+}} xmm0 = xmm1[8,9,10,11,12,13,14,15],xmm0[0,1,2,3,4,5,6,7] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> ret <2 x i64> %shuffle } @@ -753,11 +748,6 @@ define <2 x i64> @shuffle_v2i64_30_copy(<2 x i64> %nonce, <2 x i64> %a, <2 x i64 ; AVX: # BB#0: ; AVX-NEXT: vpalignr {{.*#+}} xmm0 = xmm2[8,9,10,11,12,13,14,15],xmm1[0,1,2,3,4,5,6,7] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2i64_30_copy: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vpalignr {{.*#+}} xmm0 = xmm2[8,9,10,11,12,13,14,15],xmm1[0,1,2,3,4,5,6,7] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> ret <2 x i64> %shuffle } @@ -768,10 +758,15 @@ define <2 x i64> @shuffle_v2i64_31(<2 x i64> %a, <2 x i64> %b) { ; SSE-NEXT: movdqa %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2i64_31: -; AVX: # BB#0: -; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2i64_31: +; AVX1: # BB#0: +; AVX1-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2i64_31: +; AVX2: # BB#0: +; AVX2-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2i64_31: ; AVX512VL: # BB#0: @@ -787,10 +782,15 @@ define <2 x i64> @shuffle_v2i64_31_copy(<2 x i64> %nonce, <2 x i64> %a, <2 x i64 ; SSE-NEXT: movdqa %xmm2, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2i64_31_copy: -; AVX: # BB#0: -; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm2[1],xmm1[1] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2i64_31_copy: +; AVX1: # BB#0: +; AVX1-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm2[1],xmm1[1] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2i64_31_copy: +; AVX2: # BB#0: +; AVX2-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm2[1],xmm1[1] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2i64_31_copy: ; AVX512VL: # BB#0: @@ -806,10 +806,15 @@ define <2 x i64> @shuffle_v2i64_0z(<2 x i64> %a) { ; SSE-NEXT: movq {{.*#+}} xmm0 = xmm0[0],zero ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2i64_0z: -; AVX: # BB#0: -; AVX-NEXT: vmovq {{.*#+}} xmm0 = xmm0[0],zero -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2i64_0z: +; AVX1: # BB#0: +; AVX1-NEXT: vmovq {{.*#+}} xmm0 = xmm0[0],zero +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2i64_0z: +; AVX2: # BB#0: +; AVX2-NEXT: vmovq {{.*#+}} xmm0 = xmm0[0],zero +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2i64_0z: ; AVX512VL: # BB#0: @@ -829,11 +834,6 @@ define <2 x i64> @shuffle_v2i64_1z(<2 x i64> %a) { ; AVX: # BB#0: ; AVX-NEXT: vpsrldq {{.*#+}} xmm0 = xmm0[8,9,10,11,12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2i64_1z: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vpsrldq {{.*#+}} xmm0 = xmm0[8,9,10,11,12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x i64> %a, <2 x i64> zeroinitializer, <2 x i32> ret <2 x i64> %shuffle } @@ -848,11 +848,6 @@ define <2 x i64> @shuffle_v2i64_z0(<2 x i64> %a) { ; AVX: # BB#0: ; AVX-NEXT: vpslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6,7] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2i64_z0: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vpslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6,7] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x i64> %a, <2 x i64> zeroinitializer, <2 x i32> ret <2 x i64> %shuffle } @@ -909,10 +904,15 @@ define <2 x double> @shuffle_v2f64_0z(<2 x double> %a) { ; SSE-NEXT: movq {{.*#+}} xmm0 = xmm0[0],zero ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2f64_0z: -; AVX: # BB#0: -; AVX-NEXT: vmovq {{.*#+}} xmm0 = xmm0[0],zero -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2f64_0z: +; AVX1: # BB#0: +; AVX1-NEXT: vmovq {{.*#+}} xmm0 = xmm0[0],zero +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2f64_0z: +; AVX2: # BB#0: +; AVX2-NEXT: vmovq {{.*#+}} xmm0 = xmm0[0],zero +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2f64_0z: ; AVX512VL: # BB#0: @@ -929,11 +929,17 @@ define <2 x double> @shuffle_v2f64_1z(<2 x double> %a) { ; SSE-NEXT: unpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1] ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2f64_1z: -; AVX: # BB#0: -; AVX-NEXT: vxorpd %xmm1, %xmm1, %xmm1 -; AVX-NEXT: vunpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2f64_1z: +; AVX1: # BB#0: +; AVX1-NEXT: vxorpd %xmm1, %xmm1, %xmm1 +; AVX1-NEXT: vunpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2f64_1z: +; AVX2: # BB#0: +; AVX2-NEXT: vxorpd %xmm1, %xmm1, %xmm1 +; AVX2-NEXT: vunpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2f64_1z: ; AVX512VL: # BB#0: @@ -952,11 +958,17 @@ define <2 x double> @shuffle_v2f64_z0(<2 x double> %a) { ; SSE-NEXT: movapd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2f64_z0: -; AVX: # BB#0: -; AVX-NEXT: vxorpd %xmm1, %xmm1, %xmm1 -; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2f64_z0: +; AVX1: # BB#0: +; AVX1-NEXT: vxorpd %xmm1, %xmm1, %xmm1 +; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2f64_z0: +; AVX2: # BB#0: +; AVX2-NEXT: vxorpd %xmm1, %xmm1, %xmm1 +; AVX2-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2f64_z0: ; AVX512VL: # BB#0: @@ -997,12 +1009,6 @@ define <2 x double> @shuffle_v2f64_z1(<2 x double> %a) { ; AVX-NEXT: vxorpd %xmm1, %xmm1, %xmm1 ; AVX-NEXT: vblendpd {{.*#+}} xmm0 = xmm1[0],xmm0[1] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: shuffle_v2f64_z1: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vxorpd %xmm1, %xmm1, %xmm1 -; AVX512VL-NEXT: vblendpd {{.*#+}} xmm0 = xmm1[0],xmm0[1] -; AVX512VL-NEXT: retq %shuffle = shufflevector <2 x double> %a, <2 x double> zeroinitializer, <2 x i32> ret <2 x double> %shuffle } @@ -1014,11 +1020,17 @@ define <2 x double> @shuffle_v2f64_bitcast_1z(<2 x double> %a) { ; SSE-NEXT: shufpd {{.*#+}} xmm0 = xmm0[1],xmm1[0] ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_v2f64_bitcast_1z: -; AVX: # BB#0: -; AVX-NEXT: vxorpd %xmm1, %xmm1, %xmm1 -; AVX-NEXT: vshufpd {{.*#+}} xmm0 = xmm0[1],xmm1[0] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v2f64_bitcast_1z: +; AVX1: # BB#0: +; AVX1-NEXT: vxorpd %xmm1, %xmm1, %xmm1 +; AVX1-NEXT: vshufpd {{.*#+}} xmm0 = xmm0[1],xmm1[0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v2f64_bitcast_1z: +; AVX2: # BB#0: +; AVX2-NEXT: vxorpd %xmm1, %xmm1, %xmm1 +; AVX2-NEXT: vshufpd {{.*#+}} xmm0 = xmm0[1],xmm1[0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v2f64_bitcast_1z: ; AVX512VL: # BB#0: @@ -1102,11 +1114,6 @@ define <2 x i64> @insert_reg_and_zero_v2i64(i64 %a) { ; AVX: # BB#0: ; AVX-NEXT: vmovq %rdi, %xmm0 ; AVX-NEXT: retq -; -; AVX512VL-LABEL: insert_reg_and_zero_v2i64: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vmovq %rdi, %xmm0 -; AVX512VL-NEXT: retq %v = insertelement <2 x i64> undef, i64 %a, i32 0 %shuffle = shufflevector <2 x i64> %v, <2 x i64> zeroinitializer, <2 x i32> ret <2 x i64> %shuffle @@ -1118,10 +1125,15 @@ define <2 x i64> @insert_mem_and_zero_v2i64(i64* %ptr) { ; SSE-NEXT: movq {{.*#+}} xmm0 = mem[0],zero ; SSE-NEXT: retq ; -; AVX-LABEL: insert_mem_and_zero_v2i64: -; AVX: # BB#0: -; AVX-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero -; AVX-NEXT: retq +; AVX1-LABEL: insert_mem_and_zero_v2i64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_mem_and_zero_v2i64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_mem_and_zero_v2i64: ; AVX512VL: # BB#0: @@ -1139,10 +1151,15 @@ define <2 x double> @insert_reg_and_zero_v2f64(double %a) { ; SSE-NEXT: movq {{.*#+}} xmm0 = xmm0[0],zero ; SSE-NEXT: retq ; -; AVX-LABEL: insert_reg_and_zero_v2f64: -; AVX: # BB#0: -; AVX-NEXT: vmovq {{.*#+}} xmm0 = xmm0[0],zero -; AVX-NEXT: retq +; AVX1-LABEL: insert_reg_and_zero_v2f64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovq {{.*#+}} xmm0 = xmm0[0],zero +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_reg_and_zero_v2f64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovq {{.*#+}} xmm0 = xmm0[0],zero +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_reg_and_zero_v2f64: ; AVX512VL: # BB#0: @@ -1159,10 +1176,15 @@ define <2 x double> @insert_mem_and_zero_v2f64(double* %ptr) { ; SSE-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero ; SSE-NEXT: retq ; -; AVX-LABEL: insert_mem_and_zero_v2f64: -; AVX: # BB#0: -; AVX-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero -; AVX-NEXT: retq +; AVX1-LABEL: insert_mem_and_zero_v2f64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_mem_and_zero_v2f64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_mem_and_zero_v2f64: ; AVX512VL: # BB#0: @@ -1273,11 +1295,17 @@ define <2 x i64> @insert_reg_hi_v2i64(i64 %a, <2 x i64> %b) { ; SSE-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] ; SSE-NEXT: retq ; -; AVX-LABEL: insert_reg_hi_v2i64: -; AVX: # BB#0: -; AVX-NEXT: vmovq %rdi, %xmm1 -; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] -; AVX-NEXT: retq +; AVX1-LABEL: insert_reg_hi_v2i64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovq %rdi, %xmm1 +; AVX1-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_reg_hi_v2i64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovq %rdi, %xmm1 +; AVX2-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_reg_hi_v2i64: ; AVX512VL: # BB#0: @@ -1296,11 +1324,17 @@ define <2 x i64> @insert_mem_hi_v2i64(i64* %ptr, <2 x i64> %b) { ; SSE-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] ; SSE-NEXT: retq ; -; AVX-LABEL: insert_mem_hi_v2i64: -; AVX: # BB#0: -; AVX-NEXT: vmovq {{.*#+}} xmm1 = mem[0],zero -; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] -; AVX-NEXT: retq +; AVX1-LABEL: insert_mem_hi_v2i64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovq {{.*#+}} xmm1 = mem[0],zero +; AVX1-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_mem_hi_v2i64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovq {{.*#+}} xmm1 = mem[0],zero +; AVX2-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_mem_hi_v2i64: ; AVX512VL: # BB#0: @@ -1320,10 +1354,15 @@ define <2 x double> @insert_reg_lo_v2f64(double %a, <2 x double> %b) { ; SSE-NEXT: movapd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: insert_reg_lo_v2f64: -; AVX: # BB#0: -; AVX-NEXT: vmovsd {{.*#+}} xmm0 = xmm0[0],xmm1[1] -; AVX-NEXT: retq +; AVX1-LABEL: insert_reg_lo_v2f64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovsd {{.*#+}} xmm0 = xmm0[0],xmm1[1] +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_reg_lo_v2f64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovsd {{.*#+}} xmm0 = xmm0[0],xmm1[1] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_reg_lo_v2f64: ; AVX512VL: # BB#0: @@ -1344,11 +1383,6 @@ define <2 x double> @insert_mem_lo_v2f64(double* %ptr, <2 x double> %b) { ; AVX: # BB#0: ; AVX-NEXT: vmovlpd (%rdi), %xmm0, %xmm0 ; AVX-NEXT: retq -; -; AVX512VL-LABEL: insert_mem_lo_v2f64: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vmovlpd (%rdi), %xmm0, %xmm0 -; AVX512VL-NEXT: retq %a = load double, double* %ptr %v = insertelement <2 x double> undef, double %a, i32 0 %shuffle = shufflevector <2 x double> %v, <2 x double> %b, <2 x i32> @@ -1362,10 +1396,15 @@ define <2 x double> @insert_reg_hi_v2f64(double %a, <2 x double> %b) { ; SSE-NEXT: movapd %xmm1, %xmm0 ; SSE-NEXT: retq ; -; AVX-LABEL: insert_reg_hi_v2f64: -; AVX: # BB#0: -; AVX-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0] -; AVX-NEXT: retq +; AVX1-LABEL: insert_reg_hi_v2f64: +; AVX1: # BB#0: +; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_reg_hi_v2f64: +; AVX2: # BB#0: +; AVX2-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_reg_hi_v2f64: ; AVX512VL: # BB#0: @@ -1386,11 +1425,6 @@ define <2 x double> @insert_mem_hi_v2f64(double* %ptr, <2 x double> %b) { ; AVX: # BB#0: ; AVX-NEXT: vmovhpd (%rdi), %xmm0, %xmm0 ; AVX-NEXT: retq -; -; AVX512VL-LABEL: insert_mem_hi_v2f64: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vmovhpd (%rdi), %xmm0, %xmm0 -; AVX512VL-NEXT: retq %a = load double, double* %ptr %v = insertelement <2 x double> undef, double %a, i32 0 %shuffle = shufflevector <2 x double> %v, <2 x double> %b, <2 x i32> @@ -1422,11 +1456,6 @@ define <2 x double> @insert_dup_reg_v2f64(double %a) { ; AVX: # BB#0: ; AVX-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0] ; AVX-NEXT: retq -; -; AVX512VL-LABEL: insert_dup_reg_v2f64: -; AVX512VL: # BB#0: -; AVX512VL-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0] -; AVX512VL-NEXT: retq %v = insertelement <2 x double> undef, double %a, i32 0 %shuffle = shufflevector <2 x double> %v, <2 x double> undef, <2 x i32> ret <2 x double> %shuffle @@ -1454,10 +1483,15 @@ define <2 x double> @insert_dup_mem_v2f64(double* %ptr) { ; SSE41-NEXT: movddup {{.*#+}} xmm0 = mem[0,0] ; SSE41-NEXT: retq ; -; AVX-LABEL: insert_dup_mem_v2f64: -; AVX: # BB#0: -; AVX-NEXT: vmovddup {{.*#+}} xmm0 = mem[0,0] -; AVX-NEXT: retq +; AVX1-LABEL: insert_dup_mem_v2f64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = mem[0,0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_dup_mem_v2f64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovddup {{.*#+}} xmm0 = mem[0,0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_dup_mem_v2f64: ; AVX512VL: # BB#0: @@ -1505,10 +1539,15 @@ define <2 x double> @shuffle_mem_v2f64_10(<2 x double>* %ptr) { ; SSE-NEXT: shufpd {{.*#+}} xmm0 = xmm0[1,0] ; SSE-NEXT: retq ; -; AVX-LABEL: shuffle_mem_v2f64_10: -; AVX: # BB#0: -; AVX-NEXT: vpermilpd {{.*#+}} xmm0 = mem[1,0] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_mem_v2f64_10: +; AVX1: # BB#0: +; AVX1-NEXT: vpermilpd {{.*#+}} xmm0 = mem[1,0] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_mem_v2f64_10: +; AVX2: # BB#0: +; AVX2-NEXT: vpermilpd {{.*#+}} xmm0 = mem[1,0] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_mem_v2f64_10: ; AVX512VL: # BB#0: diff --git a/test/CodeGen/X86/vector-shuffle-256-v4.ll b/test/CodeGen/X86/vector-shuffle-256-v4.ll index f6dd6c773da..3d9661084e9 100644 --- a/test/CodeGen/X86/vector-shuffle-256-v4.ll +++ b/test/CodeGen/X86/vector-shuffle-256-v4.ll @@ -1,5 +1,5 @@ -; RUN: llc < %s -mcpu=x86-64 -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1 -; RUN: llc < %s -mcpu=x86-64 -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2 +; RUN: llc < %s -mcpu=x86-64 -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX1 +; RUN: llc < %s -mcpu=x86-64 -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX2 ; RUN: llc < %s -mcpu=knl -mattr=+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL target triple = "x86_64-unknown-unknown" @@ -170,10 +170,15 @@ define <4 x double> @shuffle_v4f64_3210(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_0023(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_0023: -; AVX: # BB#0: -; AVX-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,3] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_0023: +; AVX1: # BB#0: +; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,3] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_0023: +; AVX2: # BB#0: +; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,3] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_0023: ; AVX512VL: # BB#0: @@ -194,10 +199,15 @@ define <4 x double> @shuffle_v4f64_0022(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_1032(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_1032: -; AVX: # BB#0: -; AVX-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_1032: +; AVX1: # BB#0: +; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_1032: +; AVX2: # BB#0: +; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_1032: ; AVX512VL: # BB#0: @@ -208,10 +218,15 @@ define <4 x double> @shuffle_v4f64_1032(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_1133(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_1133: -; AVX: # BB#0: -; AVX-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,3] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_1133: +; AVX1: # BB#0: +; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,3] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_1133: +; AVX2: # BB#0: +; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,3] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_1133: ; AVX512VL: # BB#0: @@ -222,10 +237,15 @@ define <4 x double> @shuffle_v4f64_1133(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_1023(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_1023: -; AVX: # BB#0: -; AVX-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,3] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_1023: +; AVX1: # BB#0: +; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,3] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_1023: +; AVX2: # BB#0: +; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,3] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_1023: ; AVX512VL: # BB#0: @@ -236,10 +256,15 @@ define <4 x double> @shuffle_v4f64_1023(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_1022(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_1022: -; AVX: # BB#0: -; AVX-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,2] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_1022: +; AVX1: # BB#0: +; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,2] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_1022: +; AVX2: # BB#0: +; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,2] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_1022: ; AVX512VL: # BB#0: @@ -283,10 +308,15 @@ define <4 x double> @shuffle_v4f64_0462(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_0426(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_0426: -; AVX: # BB#0: -; AVX-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_0426: +; AVX1: # BB#0: +; AVX1-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_0426: +; AVX2: # BB#0: +; AVX2-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_0426: ; AVX512VL: # BB#0: @@ -297,10 +327,15 @@ define <4 x double> @shuffle_v4f64_0426(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_1537(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_1537: -; AVX: # BB#0: -; AVX-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_1537: +; AVX1: # BB#0: +; AVX1-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_1537: +; AVX2: # BB#0: +; AVX2-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_1537: ; AVX512VL: # BB#0: @@ -311,10 +346,15 @@ define <4 x double> @shuffle_v4f64_1537(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_4062(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_4062: -; AVX: # BB#0: -; AVX-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_4062: +; AVX1: # BB#0: +; AVX1-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_4062: +; AVX2: # BB#0: +; AVX2-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_4062: ; AVX512VL: # BB#0: @@ -325,10 +365,15 @@ define <4 x double> @shuffle_v4f64_4062(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_5173(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_5173: -; AVX: # BB#0: -; AVX-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[3],ymm0[3] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_5173: +; AVX1: # BB#0: +; AVX1-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[3],ymm0[3] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_5173: +; AVX2: # BB#0: +; AVX2-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[3],ymm0[3] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_5173: ; AVX512VL: # BB#0: @@ -339,10 +384,15 @@ define <4 x double> @shuffle_v4f64_5173(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_5163(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_5163: -; AVX: # BB#0: -; AVX-NEXT: vshufpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[2],ymm0[3] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_5163: +; AVX1: # BB#0: +; AVX1-NEXT: vshufpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[2],ymm0[3] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_5163: +; AVX2: # BB#0: +; AVX2-NEXT: vshufpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[2],ymm0[3] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_5163: ; AVX512VL: # BB#0: @@ -371,10 +421,15 @@ define <4 x double> @shuffle_v4f64_4163(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_0145(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_0145: -; AVX: # BB#0: -; AVX-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_0145: +; AVX1: # BB#0: +; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_0145: +; AVX2: # BB#0: +; AVX2-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_0145: ; AVX512VL: # BB#0: @@ -385,10 +440,15 @@ define <4 x double> @shuffle_v4f64_0145(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_4501(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_4501: -; AVX: # BB#0: -; AVX-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0 -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_4501: +; AVX1: # BB#0: +; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_4501: +; AVX2: # BB#0: +; AVX2-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0 +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_4501: ; AVX512VL: # BB#0: @@ -408,11 +468,17 @@ define <4 x double> @shuffle_v4f64_0167(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_1054(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_1054: -; AVX: # BB#0: -; AVX-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 -; AVX-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_1054: +; AVX1: # BB#0: +; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_1054: +; AVX2: # BB#0: +; AVX2-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_1054: ; AVX512VL: # BB#0: @@ -424,11 +490,17 @@ define <4 x double> @shuffle_v4f64_1054(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_3254(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_3254: -; AVX: # BB#0: -; AVX-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1] -; AVX-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_3254: +; AVX1: # BB#0: +; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1] +; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_3254: +; AVX2: # BB#0: +; AVX2-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1] +; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_3254: ; AVX512VL: # BB#0: @@ -440,11 +512,17 @@ define <4 x double> @shuffle_v4f64_3254(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_3276(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_3276: -; AVX: # BB#0: -; AVX-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3] -; AVX-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_3276: +; AVX1: # BB#0: +; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3] +; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_3276: +; AVX2: # BB#0: +; AVX2-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3] +; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_3276: ; AVX512VL: # BB#0: @@ -456,11 +534,17 @@ define <4 x double> @shuffle_v4f64_3276(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_1076(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_1076: -; AVX: # BB#0: -; AVX-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3] -; AVX-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_1076: +; AVX1: # BB#0: +; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3] +; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_1076: +; AVX2: # BB#0: +; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3] +; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_1076: ; AVX512VL: # BB#0: @@ -497,10 +581,15 @@ define <4 x double> @shuffle_v4f64_0415(<4 x double> %a, <4 x double> %b) { } define <4 x double> @shuffle_v4f64_u062(<4 x double> %a, <4 x double> %b) { -; AVX-LABEL: shuffle_v4f64_u062: -; AVX: # BB#0: -; AVX-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2] -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4f64_u062: +; AVX1: # BB#0: +; AVX1-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2] +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4f64_u062: +; AVX2: # BB#0: +; AVX2-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4f64_u062: ; AVX512VL: # BB#0: @@ -797,10 +886,15 @@ define <4 x i64> @shuffle_v4i64_4012(<4 x i64> %a, <4 x i64> %b) { } define <4 x i64> @shuffle_v4i64_0145(<4 x i64> %a, <4 x i64> %b) { -; AVX-LABEL: shuffle_v4i64_0145: -; AVX: # BB#0: -; AVX-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4i64_0145: +; AVX1: # BB#0: +; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4i64_0145: +; AVX2: # BB#0: +; AVX2-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4i64_0145: ; AVX512VL: # BB#0: @@ -836,10 +930,15 @@ define <4 x i64> @shuffle_v4i64_0451(<4 x i64> %a, <4 x i64> %b) { } define <4 x i64> @shuffle_v4i64_4501(<4 x i64> %a, <4 x i64> %b) { -; AVX-LABEL: shuffle_v4i64_4501: -; AVX: # BB#0: -; AVX-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0 -; AVX-NEXT: retq +; AVX1-LABEL: shuffle_v4i64_4501: +; AVX1: # BB#0: +; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: shuffle_v4i64_4501: +; AVX2: # BB#0: +; AVX2-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0 +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: shuffle_v4i64_4501: ; AVX512VL: # BB#0: @@ -1118,10 +1217,15 @@ define <4 x i64> @insert_reg_and_zero_v4i64(i64 %a) { } define <4 x i64> @insert_mem_and_zero_v4i64(i64* %ptr) { -; AVX-LABEL: insert_mem_and_zero_v4i64: -; AVX: # BB#0: -; AVX-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero -; AVX-NEXT: retq +; AVX1-LABEL: insert_mem_and_zero_v4i64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_mem_and_zero_v4i64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_mem_and_zero_v4i64: ; AVX512VL: # BB#0: @@ -1134,11 +1238,17 @@ define <4 x i64> @insert_mem_and_zero_v4i64(i64* %ptr) { } define <4 x double> @insert_reg_and_zero_v4f64(double %a) { -; AVX-LABEL: insert_reg_and_zero_v4f64: -; AVX: # BB#0: -; AVX-NEXT: vxorpd %ymm1, %ymm1, %ymm1 -; AVX-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3] -; AVX-NEXT: retq +; AVX1-LABEL: insert_reg_and_zero_v4f64: +; AVX1: # BB#0: +; AVX1-NEXT: vxorpd %ymm1, %ymm1, %ymm1 +; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3] +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_reg_and_zero_v4f64: +; AVX2: # BB#0: +; AVX2-NEXT: vxorpd %ymm1, %ymm1, %ymm1 +; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3] +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_reg_and_zero_v4f64: ; AVX512VL: # BB#0: @@ -1151,10 +1261,15 @@ define <4 x double> @insert_reg_and_zero_v4f64(double %a) { } define <4 x double> @insert_mem_and_zero_v4f64(double* %ptr) { -; AVX-LABEL: insert_mem_and_zero_v4f64: -; AVX: # BB#0: -; AVX-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero -; AVX-NEXT: retq +; AVX1-LABEL: insert_mem_and_zero_v4f64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_mem_and_zero_v4f64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_mem_and_zero_v4f64: ; AVX512VL: # BB#0: @@ -1178,10 +1293,15 @@ define <4 x double> @splat_mem_v4f64(double* %ptr) { } define <4 x i64> @splat_mem_v4i64(i64* %ptr) { -; AVX-LABEL: splat_mem_v4i64: -; AVX: # BB#0: -; AVX-NEXT: vbroadcastsd (%rdi), %ymm0 -; AVX-NEXT: retq +; AVX1-LABEL: splat_mem_v4i64: +; AVX1: # BB#0: +; AVX1-NEXT: vbroadcastsd (%rdi), %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: splat_mem_v4i64: +; AVX2: # BB#0: +; AVX2-NEXT: vbroadcastsd (%rdi), %ymm0 +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: splat_mem_v4i64: ; AVX512VL: # BB#0: @@ -1268,11 +1388,17 @@ define <4 x double> @splat_mem_v4f64_from_v2f64(<2 x double>* %ptr) { } define <4 x i64> @splat128_mem_v4i64_from_v2i64(<2 x i64>* %ptr) { -; AVX-LABEL: splat128_mem_v4i64_from_v2i64: -; AVX: # BB#0: -; AVX-NEXT: vmovaps (%rdi), %xmm0 -; AVX-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 -; AVX-NEXT: retq +; AVX1-LABEL: splat128_mem_v4i64_from_v2i64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovaps (%rdi), %xmm0 +; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: splat128_mem_v4i64_from_v2i64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovaps (%rdi), %xmm0 +; AVX2-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: splat128_mem_v4i64_from_v2i64: ; AVX512VL: # BB#0: @@ -1285,11 +1411,17 @@ define <4 x i64> @splat128_mem_v4i64_from_v2i64(<2 x i64>* %ptr) { } define <4 x double> @splat128_mem_v4f64_from_v2f64(<2 x double>* %ptr) { -; AVX-LABEL: splat128_mem_v4f64_from_v2f64: -; AVX: # BB#0: -; AVX-NEXT: vmovaps (%rdi), %xmm0 -; AVX-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 -; AVX-NEXT: retq +; AVX1-LABEL: splat128_mem_v4f64_from_v2f64: +; AVX1: # BB#0: +; AVX1-NEXT: vmovaps (%rdi), %xmm0 +; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: splat128_mem_v4f64_from_v2f64: +; AVX2: # BB#0: +; AVX2-NEXT: vmovaps (%rdi), %xmm0 +; AVX2-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: splat128_mem_v4f64_from_v2f64: ; AVX512VL: # BB#0: @@ -1347,10 +1479,15 @@ define <4 x i64> @concat_v4i64_0167(<4 x i64> %a0, <4 x i64> %a1) { } define <4 x i64> @concat_v4i64_0145_bc(<4 x i64> %a0, <4 x i64> %a1) { -; AVX-LABEL: concat_v4i64_0145_bc: -; AVX: # BB#0: -; AVX-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 -; AVX-NEXT: retq +; AVX1-LABEL: concat_v4i64_0145_bc: +; AVX1: # BB#0: +; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: concat_v4i64_0145_bc: +; AVX2: # BB#0: +; AVX2-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: concat_v4i64_0145_bc: ; AVX512VL: # BB#0: @@ -1366,10 +1503,15 @@ define <4 x i64> @concat_v4i64_0145_bc(<4 x i64> %a0, <4 x i64> %a1) { } define <4 x i64> @insert_dup_mem_v4i64(i64* %ptr) { -; AVX-LABEL: insert_dup_mem_v4i64: -; AVX: # BB#0: -; AVX-NEXT: vbroadcastsd (%rdi), %ymm0 -; AVX-NEXT: retq +; AVX1-LABEL: insert_dup_mem_v4i64: +; AVX1: # BB#0: +; AVX1-NEXT: vbroadcastsd (%rdi), %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: insert_dup_mem_v4i64: +; AVX2: # BB#0: +; AVX2-NEXT: vbroadcastsd (%rdi), %ymm0 +; AVX2-NEXT: retq ; ; AVX512VL-LABEL: insert_dup_mem_v4i64: ; AVX512VL: # BB#0: