[llvm] r365265 - [X86] Copy some test cases from vector-shuffle-sse1.ll to vector-shuffle-128-v4.ll and v8 where sse1 did better load folding. NFC

[Craig Topper via llvm-commits]

Author: ctopper
Date: Sat Jul  6 10:59:41 2019
New Revision: 365265

URL: http://llvm.org/viewvc/llvm-project?rev=365265&view=rev
Log:
[X86] Copy some test cases from vector-shuffle-sse1.ll to vector-shuffle-128-v4.ll and v8 where sse1 did better load folding. NFC

Modified:
    llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v2.ll
    llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v4.ll

Modified: llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v2.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v2.ll?rev=365265&r1=365264&r2=365265&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v2.ll (original)
+++ llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v2.ll Sat Jul  6 10:59:41 2019
@@ -1305,3 +1305,53 @@ define <2 x double> @shuffle_mem_v2f64_3
   %f = shufflevector <2 x double> %a, <2 x double> %c, <2 x i32> <i32 3, i32 1>
   ret <2 x double> %f
 }
+
+define <2 x double> @shuffle_mem_v2f64_02(<2 x double> %a, <2 x double>* %pb) {
+; SSE-LABEL: shuffle_mem_v2f64_02:
+; SSE:       # %bb.0:
+; SSE-NEXT:    movups (%rdi), %xmm1
+; SSE-NEXT:    movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT:    retq
+;
+; AVX-LABEL: shuffle_mem_v2f64_02:
+; AVX:       # %bb.0:
+; AVX-NEXT:    vunpcklpd {{.*#+}} xmm0 = xmm0[0],mem[0]
+; AVX-NEXT:    retq
+  %b = load <2 x double>, <2 x double>* %pb, align 1
+  %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 0, i32 2>
+  ret <2 x double> %shuffle
+}
+
+define <2 x double> @shuffle_mem_v2f64_21(<2 x double> %a, <2 x double>* %pb) {
+; SSE2-LABEL: shuffle_mem_v2f64_21:
+; SSE2:       # %bb.0:
+; SSE2-NEXT:    movupd (%rdi), %xmm1
+; SSE2-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSE2-NEXT:    retq
+;
+; SSE3-LABEL: shuffle_mem_v2f64_21:
+; SSE3:       # %bb.0:
+; SSE3-NEXT:    movupd (%rdi), %xmm1
+; SSE3-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSE3-NEXT:    retq
+;
+; SSSE3-LABEL: shuffle_mem_v2f64_21:
+; SSSE3:       # %bb.0:
+; SSSE3-NEXT:    movupd (%rdi), %xmm1
+; SSSE3-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSSE3-NEXT:    retq
+;
+; SSE41-LABEL: shuffle_mem_v2f64_21:
+; SSE41:       # %bb.0:
+; SSE41-NEXT:    movups (%rdi), %xmm1
+; SSE41-NEXT:    blendps {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3]
+; SSE41-NEXT:    retq
+;
+; AVX-LABEL: shuffle_mem_v2f64_21:
+; AVX:       # %bb.0:
+; AVX-NEXT:    vblendps {{.*#+}} xmm0 = mem[0,1],xmm0[2,3]
+; AVX-NEXT:    retq
+  %b = load <2 x double>, <2 x double>* %pb, align 1
+  %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 2, i32 1>
+  ret <2 x double> %shuffle
+}

Modified: llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v4.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v4.ll?rev=365265&r1=365264&r2=365265&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v4.ll (original)
+++ llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v4.ll Sat Jul  6 10:59:41 2019
@@ -2437,3 +2437,53 @@ define <4 x i32> @shuffle_v4i32_1z3z(<4
   %shuffle = shufflevector <4 x i32> %a, <4 x i32> zeroinitializer, <4 x i32> <i32 1, i32 4, i32 3, i32 4>
   ret <4 x i32> %shuffle
 }
+
+define <4 x float> @shuffle_mem_v4f32_0145(<4 x float> %a, <4 x float>* %pb) {
+; SSE-LABEL: shuffle_mem_v4f32_0145:
+; SSE:       # %bb.0:
+; SSE-NEXT:    movups (%rdi), %xmm1
+; SSE-NEXT:    movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; SSE-NEXT:    retq
+;
+; AVX-LABEL: shuffle_mem_v4f32_0145:
+; AVX:       # %bb.0:
+; AVX-NEXT:    vunpcklpd {{.*#+}} xmm0 = xmm0[0],mem[0]
+; AVX-NEXT:    retq
+  %b = load <4 x float>, <4 x float>* %pb, align 1
+  %shuffle = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
+  ret <4 x float> %shuffle
+}
+
+define <4 x float> @shuffle_mem_v4f32_4523(<4 x float> %a, <4 x float>* %pb) {
+; SSE2-LABEL: shuffle_mem_v4f32_4523:
+; SSE2:       # %bb.0:
+; SSE2-NEXT:    movupd (%rdi), %xmm1
+; SSE2-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSE2-NEXT:    retq
+;
+; SSE3-LABEL: shuffle_mem_v4f32_4523:
+; SSE3:       # %bb.0:
+; SSE3-NEXT:    movupd (%rdi), %xmm1
+; SSE3-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSE3-NEXT:    retq
+;
+; SSSE3-LABEL: shuffle_mem_v4f32_4523:
+; SSSE3:       # %bb.0:
+; SSSE3-NEXT:    movupd (%rdi), %xmm1
+; SSSE3-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSSE3-NEXT:    retq
+;
+; SSE41-LABEL: shuffle_mem_v4f32_4523:
+; SSE41:       # %bb.0:
+; SSE41-NEXT:    movups (%rdi), %xmm1
+; SSE41-NEXT:    blendps {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3]
+; SSE41-NEXT:    retq
+;
+; AVX-LABEL: shuffle_mem_v4f32_4523:
+; AVX:       # %bb.0:
+; AVX-NEXT:    vblendps {{.*#+}} xmm0 = mem[0,1],xmm0[2,3]
+; AVX-NEXT:    retq
+  %b = load <4 x float>, <4 x float>* %pb, align 1
+  %shuffle = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+  ret <4 x float> %shuffle
+}