[X86][SSE42] Added fast-isel tests to sync with clang/test/CodeGen/sse42-builtins.c

llvm-svn: 269929

2 files changed, 411 insertions, 0 deletions

diff --git a/llvm/test/CodeGen/X86/sse42-intrinsics-fast-isel-x86_64.ll b/llvm/test/CodeGen/X86/sse42-intrinsics-fast-isel-x86_64.ll
new file mode 100644
index 00000000000..0a69d263212
--- /dev/null
+++ b/llvm/test/CodeGen/X86/sse42-intrinsics-fast-isel-x86_64.ll
@@ -0,0 +1,26 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -fast-isel -mtriple=x86_64-unknown-unknown -mattr=+sse4.2 | FileCheck %s --check-prefix=X64
+
+; NOTE: This should use IR equivalent to what is generated by clang/test/CodeGen/sse42-builtins.c
+
+define i64 @test_mm_crc64_u8(i64 %a0, i8 %a1) nounwind{
+; X64-LABEL: test_mm_crc64_u8:
+; X64:       # BB#0:
+; X64-NEXT:    crc32b %sil, %edi
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    retq
+  %res = call i64 @llvm.x86.sse42.crc32.64.8(i64 %a0, i8 %a1)
+  ret i64 %res
+}
+declare i64 @llvm.x86.sse42.crc32.64.8(i64, i8) nounwind readnone
+
+define i64 @test_mm_crc64_u64(i64 %a0, i64 %a1) nounwind{
+; X64-LABEL: test_mm_crc64_u64:
+; X64:       # BB#0:
+; X64-NEXT:    crc32q %rsi, %rdi
+; X64-NEXT:    movq %rdi, %rax
+; X64-NEXT:    retq
+  %res = call i64 @llvm.x86.sse42.crc32.64.64(i64 %a0, i64 %a1)
+  ret i64 %res
+}
+declare i64 @llvm.x86.sse42.crc32.64.64(i64, i64) nounwind readnone
diff --git a/llvm/test/CodeGen/X86/sse42-intrinsics-fast-isel.ll b/llvm/test/CodeGen/X86/sse42-intrinsics-fast-isel.ll
new file mode 100644
index 00000000000..926ace80917
--- /dev/null
+++ b/llvm/test/CodeGen/X86/sse42-intrinsics-fast-isel.ll
@@ -0,0 +1,385 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -fast-isel -mtriple=i386-unknown-unknown -mattr=+sse4.2 | FileCheck %s --check-prefix=ALL --check-prefix=X32
+; RUN: llc < %s -fast-isel -mtriple=x86_64-unknown-unknown -mattr=+sse4.2 | FileCheck %s --check-prefix=ALL --check-prefix=X64
+
+; NOTE: This should use IR equivalent to what is generated by clang/test/CodeGen/sse42-builtins.c
+
+define i32 @test_mm_cmpestra(<2 x i64> %a0, i32 %a1, <2 x i64> %a2, i32 %a3) {
+; X32-LABEL: test_mm_cmpestra:
+; X32:       # BB#0:
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %edx
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X32-NEXT:    seta %al
+; X32-NEXT:    movzbl %al, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpestra:
+; X64:       # BB#0:
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %esi, %edx
+; X64-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X64-NEXT:    seta %al
+; X64-NEXT:    movzbl %al, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg2 = bitcast <2 x i64> %a2 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpestria128(<16 x i8> %arg0, i32 %a1, <16 x i8> %arg2, i32 %a3, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpestria128(<16 x i8>, i32, <16 x i8>, i32, i8) nounwind readnone
+
+define i32 @test_mm_cmpestrc(<2 x i64> %a0, i32 %a1, <2 x i64> %a2, i32 %a3) {
+; X32-LABEL: test_mm_cmpestrc:
+; X32:       # BB#0:
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %edx
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X32-NEXT:    sbbl %eax, %eax
+; X32-NEXT:    andl $1, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpestrc:
+; X64:       # BB#0:
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %esi, %edx
+; X64-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X64-NEXT:    sbbl %eax, %eax
+; X64-NEXT:    andl $1, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg2 = bitcast <2 x i64> %a2 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpestric128(<16 x i8> %arg0, i32 %a1, <16 x i8> %arg2, i32 %a3, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpestric128(<16 x i8>, i32, <16 x i8>, i32, i8) nounwind readnone
+
+define i32 @test_mm_cmpestri(<2 x i64> %a0, i32 %a1, <2 x i64> %a2, i32 %a3) {
+; X32-LABEL: test_mm_cmpestri:
+; X32:       # BB#0:
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %edx
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X32-NEXT:    movl %ecx, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpestri:
+; X64:       # BB#0:
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %esi, %edx
+; X64-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg2 = bitcast <2 x i64> %a2 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpestri128(<16 x i8> %arg0, i32 %a1, <16 x i8> %arg2, i32 %a3, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpestri128(<16 x i8>, i32, <16 x i8>, i32, i8) nounwind readnone
+
+define <2 x i64> @test_mm_cmpestrm(<2 x i64> %a0, i32 %a1, <2 x i64> %a2, i32 %a3) {
+; X32-LABEL: test_mm_cmpestrm:
+; X32:       # BB#0:
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %edx
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    pcmpestrm $7, %xmm1, %xmm0
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpestrm:
+; X64:       # BB#0:
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %esi, %edx
+; X64-NEXT:    pcmpestrm $7, %xmm1, %xmm0
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg2 = bitcast <2 x i64> %a2 to <16 x i8>
+  %res = call <16 x i8> @llvm.x86.sse42.pcmpestrm128(<16 x i8> %arg0, i32 %a1, <16 x i8> %arg2, i32 %a3, i8 7)
+  %bc = bitcast <16 x i8> %res to <2 x i64>
+  ret <2 x i64> %bc
+}
+declare <16 x i8> @llvm.x86.sse42.pcmpestrm128(<16 x i8>, i32, <16 x i8>, i32, i8) nounwind readnone
+
+define i32 @test_mm_cmpestro(<2 x i64> %a0, i32 %a1, <2 x i64> %a2, i32 %a3) {
+; X32-LABEL: test_mm_cmpestro:
+; X32:       # BB#0:
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %edx
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X32-NEXT:    seto %al
+; X32-NEXT:    movzbl %al, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpestro:
+; X64:       # BB#0:
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %esi, %edx
+; X64-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X64-NEXT:    seto %al
+; X64-NEXT:    movzbl %al, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg2 = bitcast <2 x i64> %a2 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpestrio128(<16 x i8> %arg0, i32 %a1, <16 x i8> %arg2, i32 %a3, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpestrio128(<16 x i8>, i32, <16 x i8>, i32, i8) nounwind readnone
+
+define i32 @test_mm_cmpestrs(<2 x i64> %a0, i32 %a1, <2 x i64> %a2, i32 %a3) {
+; X32-LABEL: test_mm_cmpestrs:
+; X32:       # BB#0:
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %edx
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X32-NEXT:    sets %al
+; X32-NEXT:    movzbl %al, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpestrs:
+; X64:       # BB#0:
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %esi, %edx
+; X64-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X64-NEXT:    sets %al
+; X64-NEXT:    movzbl %al, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg2 = bitcast <2 x i64> %a2 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpestris128(<16 x i8> %arg0, i32 %a1, <16 x i8> %arg2, i32 %a3, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpestris128(<16 x i8>, i32, <16 x i8>, i32, i8) nounwind readnone
+
+define i32 @test_mm_cmpestrz(<2 x i64> %a0, i32 %a1, <2 x i64> %a2, i32 %a3) {
+; X32-LABEL: test_mm_cmpestrz:
+; X32:       # BB#0:
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %edx
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X32-NEXT:    sete %al
+; X32-NEXT:    movzbl %al, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpestrz:
+; X64:       # BB#0:
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    movl %esi, %edx
+; X64-NEXT:    pcmpestri $7, %xmm1, %xmm0
+; X64-NEXT:    sete %al
+; X64-NEXT:    movzbl %al, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg2 = bitcast <2 x i64> %a2 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpestriz128(<16 x i8> %arg0, i32 %a1, <16 x i8> %arg2, i32 %a3, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpestriz128(<16 x i8>, i32, <16 x i8>, i32, i8) nounwind readnone
+
+define <2 x i64> @test_mm_cmpgt_epi64(<2 x i64> %a0, <2 x i64> %a1) {
+; X32-LABEL: test_mm_cmpgt_epi64:
+; X32:       # BB#0:
+; X32-NEXT:    pcmpgtq %xmm1, %xmm0
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpgt_epi64:
+; X64:       # BB#0:
+; X64-NEXT:    pcmpgtq %xmm1, %xmm0
+; X64-NEXT:    retq
+  %cmp = icmp sgt <2 x i64> %a0, %a1
+  %res = sext <2 x i1> %cmp to <2 x i64>
+  ret <2 x i64> %res
+}
+
+define i32 @test_mm_cmpistra(<2 x i64> %a0, <2 x i64> %a1) {
+; X32-LABEL: test_mm_cmpistra:
+; X32:       # BB#0:
+; X32-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X32-NEXT:    seta %al
+; X32-NEXT:    movzbl %al, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpistra:
+; X64:       # BB#0:
+; X64-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X64-NEXT:    seta %al
+; X64-NEXT:    movzbl %al, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg1 = bitcast <2 x i64> %a1 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpistria128(<16 x i8> %arg0, <16 x i8> %arg1, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpistria128(<16 x i8>, <16 x i8>, i8) nounwind readnone
+
+define i32 @test_mm_cmpistrc(<2 x i64> %a0, <2 x i64> %a1) {
+; X32-LABEL: test_mm_cmpistrc:
+; X32:       # BB#0:
+; X32-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X32-NEXT:    sbbl %eax, %eax
+; X32-NEXT:    andl $1, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpistrc:
+; X64:       # BB#0:
+; X64-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X64-NEXT:    sbbl %eax, %eax
+; X64-NEXT:    andl $1, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg1 = bitcast <2 x i64> %a1 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpistric128(<16 x i8> %arg0, <16 x i8> %arg1, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpistric128(<16 x i8>, <16 x i8>, i8) nounwind readnone
+
+define i32 @test_mm_cmpistri(<2 x i64> %a0, <2 x i64> %a1) {
+; X32-LABEL: test_mm_cmpistri:
+; X32:       # BB#0:
+; X32-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X32-NEXT:    movl %ecx, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpistri:
+; X64:       # BB#0:
+; X64-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg1 = bitcast <2 x i64> %a1 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpistri128(<16 x i8> %arg0, <16 x i8> %arg1, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpistri128(<16 x i8>, <16 x i8>, i8) nounwind readnone
+
+define <2 x i64> @test_mm_cmpistrm(<2 x i64> %a0, <2 x i64> %a1) {
+; X32-LABEL: test_mm_cmpistrm:
+; X32:       # BB#0:
+; X32-NEXT:    pcmpistrm $7, %xmm1, %xmm0
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpistrm:
+; X64:       # BB#0:
+; X64-NEXT:    pcmpistrm $7, %xmm1, %xmm0
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg1 = bitcast <2 x i64> %a1 to <16 x i8>
+  %res = call <16 x i8> @llvm.x86.sse42.pcmpistrm128(<16 x i8> %arg0, <16 x i8> %arg1, i8 7)
+  %bc = bitcast <16 x i8> %res to <2 x i64>
+  ret <2 x i64> %bc
+}
+declare <16 x i8> @llvm.x86.sse42.pcmpistrm128(<16 x i8>, <16 x i8>, i8) nounwind readnone
+
+define i32 @test_mm_cmpistro(<2 x i64> %a0, <2 x i64> %a1) {
+; X32-LABEL: test_mm_cmpistro:
+; X32:       # BB#0:
+; X32-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X32-NEXT:    seto %al
+; X32-NEXT:    movzbl %al, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpistro:
+; X64:       # BB#0:
+; X64-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X64-NEXT:    seto %al
+; X64-NEXT:    movzbl %al, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg1 = bitcast <2 x i64> %a1 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpistrio128(<16 x i8> %arg0, <16 x i8> %arg1, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpistrio128(<16 x i8>, <16 x i8>, i8) nounwind readnone
+
+define i32 @test_mm_cmpistrs(<2 x i64> %a0, <2 x i64> %a1) {
+; X32-LABEL: test_mm_cmpistrs:
+; X32:       # BB#0:
+; X32-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X32-NEXT:    sets %al
+; X32-NEXT:    movzbl %al, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpistrs:
+; X64:       # BB#0:
+; X64-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X64-NEXT:    sets %al
+; X64-NEXT:    movzbl %al, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg1 = bitcast <2 x i64> %a1 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpistris128(<16 x i8> %arg0, <16 x i8> %arg1, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpistris128(<16 x i8>, <16 x i8>, i8) nounwind readnone
+
+define i32 @test_mm_cmpistrz(<2 x i64> %a0, <2 x i64> %a1) {
+; X32-LABEL: test_mm_cmpistrz:
+; X32:       # BB#0:
+; X32-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X32-NEXT:    sete %al
+; X32-NEXT:    movzbl %al, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_cmpistrz:
+; X64:       # BB#0:
+; X64-NEXT:    pcmpistri $7, %xmm1, %xmm0
+; X64-NEXT:    sete %al
+; X64-NEXT:    movzbl %al, %eax
+; X64-NEXT:    retq
+  %arg0 = bitcast <2 x i64> %a0 to <16 x i8>
+  %arg1 = bitcast <2 x i64> %a1 to <16 x i8>
+  %res = call i32 @llvm.x86.sse42.pcmpistriz128(<16 x i8> %arg0, <16 x i8> %arg1, i8 7)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.pcmpistriz128(<16 x i8>, <16 x i8>, i8) nounwind readnone
+
+define i32 @test_mm_crc32_u8(i32 %a0, i8 %a1) {
+; X32-LABEL: test_mm_crc32_u8:
+; X32:       # BB#0:
+; X32-NEXT:    movb {{[0-9]+}}(%esp), %cl
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    crc32b %cl, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_crc32_u8:
+; X64:       # BB#0:
+; X64-NEXT:    crc32b %sil, %edi
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    retq
+  %res = call i32 @llvm.x86.sse42.crc32.32.8(i32 %a0, i8 %a1)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.crc32.32.8(i32, i8) nounwind readnone
+
+define i32 @test_mm_crc32_u16(i32 %a0, i16 %a1) {
+; X32-LABEL: test_mm_crc32_u16:
+; X32:       # BB#0:
+; X32-NEXT:    movzwl {{[0-9]+}}(%esp), %ecx
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    crc32w %cx, %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_crc32_u16:
+; X64:       # BB#0:
+; X64-NEXT:    crc32w %si, %edi
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    retq
+  %res = call i32 @llvm.x86.sse42.crc32.32.16(i32 %a0, i16 %a1)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.crc32.32.16(i32, i16) nounwind readnone
+
+define i32 @test_mm_crc32_u32(i32 %a0, i32 %a1) {
+; X32-LABEL: test_mm_crc32_u32:
+; X32:       # BB#0:
+; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    crc32l {{[0-9]+}}(%esp), %eax
+; X32-NEXT:    retl
+;
+; X64-LABEL: test_mm_crc32_u32:
+; X64:       # BB#0:
+; X64-NEXT:    crc32l %esi, %edi
+; X64-NEXT:    movl %edi, %eax
+; X64-NEXT:    retq
+  %res = call i32 @llvm.x86.sse42.crc32.32.32(i32 %a0, i32 %a1)
+  ret i32 %res
+}
+declare i32 @llvm.x86.sse42.crc32.32.32(i32, i32) nounwind readnone