diff options
Diffstat (limited to 'clang/test/CodeGen/X86/sse2-builtins.c')
| -rw-r--r-- | clang/test/CodeGen/X86/sse2-builtins.c | 57 |
1 files changed, 57 insertions, 0 deletions
diff --git a/clang/test/CodeGen/X86/sse2-builtins.c b/clang/test/CodeGen/X86/sse2-builtins.c index 38d5e877a503..0ba32bb230cd 100644 --- a/clang/test/CodeGen/X86/sse2-builtins.c +++ b/clang/test/CodeGen/X86/sse2-builtins.c @@ -9,6 +9,17 @@ // RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse2 -fno-signed-char -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,X86 // RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -fms-extensions -fms-compatibility -ffreestanding %s -triple=x86_64-windows-msvc -target-feature +sse2 -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,X64 +// RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse2 -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s --check-prefixes=CHECK,X64 +// RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse2 -fno-signed-char -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s --check-prefixes=CHECK,X64 +// RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse2 -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s --check-prefixes=CHECK,X86 +// RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse2 -fno-signed-char -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s --check-prefixes=CHECK,X86 +// RUN: %clang_cc1 -x c -flax-vector-conversions=none -fms-extensions -fms-compatibility -ffreestanding %s -triple=x86_64-windows-msvc -target-feature +sse2 -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s --check-prefixes=CHECK,X64 +// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse2 -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s --check-prefixes=CHECK,X64 +// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse2 -fno-signed-char -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s --check-prefixes=CHECK,X64 +// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse2 -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s --check-prefixes=CHECK,X86 +// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse2 -fno-signed-char -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s --check-prefixes=CHECK,X86 +// RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -fms-extensions -fms-compatibility -ffreestanding %s -triple=x86_64-windows-msvc -target-feature +sse2 -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s --check-prefixes=CHECK,X64 + #include <immintrin.h> #include "builtin_test_helpers.h" @@ -123,12 +134,14 @@ __m128i test_mm_avg_epu8(__m128i A, __m128i B) { // CHECK: call <16 x i8> @llvm.x86.sse2.pavg.b(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}) return _mm_avg_epu8(A, B); } +TEST_CONSTEXPR(match_v16qu(_mm_avg_epu8((__m128i)(__v16qu){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16}, (__m128i)(__v16qu){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16}), 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)); __m128i test_mm_avg_epu16(__m128i A, __m128i B) { // CHECK-LABEL: test_mm_avg_epu16 // CHECK: call <8 x i16> @llvm.x86.sse2.pavg.w(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}) return _mm_avg_epu16(A, B); } +TEST_CONSTEXPR(match_v8hu(_mm_avg_epu16((__m128i)(__v8hu){1, 2, 3, 4, 5, 6, 7, 8}, (__m128i)(__v8hu){1, 2, 3, 4, 5, 6, 7, 8}), 1, 2, 3, 4, 5, 6, 7, 8)); __m128i test_mm_bslli_si128(__m128i A) { // CHECK-LABEL: test_mm_bslli_si128 @@ -243,18 +256,21 @@ __m128i test_mm_cmpeq_epi8(__m128i A, __m128i B) { // CHECK: icmp eq <16 x i8> return _mm_cmpeq_epi8(A, B); } +TEST_CONSTEXPR(match_v16qi(_mm_cmpeq_epi8((__m128i)(__v16qs){1,-2,3,-4,-5,6,-7,8,-9,10,-11,12,-13,14,-15,16}, (__m128i)(__v16qs){10,-2,6,-4,-5,12,-14,8,-9,20,-22,12,-26,14,-30,16}), 0,-1,0,-1,-1,0,0,-1,-1,0,0,-1,0,-1,0,-1)); __m128i test_mm_cmpeq_epi16(__m128i A, __m128i B) { // CHECK-LABEL: test_mm_cmpeq_epi16 // CHECK: icmp eq <8 x i16> return _mm_cmpeq_epi16(A, B); } +TEST_CONSTEXPR(match_v8hi(_mm_cmpeq_epi16((__m128i)(__v8hi){+1, -2, +3, -4, +5, -6, +7, -8}, (__m128i)(__v8hi){-10, -2, +6, -4, +5, -12, +14, -8}), 0, -1, 0, -1, -1, 0, 0, -1)); __m128i test_mm_cmpeq_epi32(__m128i A, __m128i B) { // CHECK-LABEL: test_mm_cmpeq_epi32 // CHECK: icmp eq <4 x i32> return _mm_cmpeq_epi32(A, B); } +TEST_CONSTEXPR(match_v4si(_mm_cmpeq_epi32((__m128i)(__v4si){+1, -2, +3, -4}, (__m128i)(__v4si){-10, -2, +6, -4}), 0, -1, 0, -1)); __m128d test_mm_cmpeq_pd(__m128d A, __m128d B) { // CHECK-LABEL: test_mm_cmpeq_pd @@ -293,18 +309,24 @@ __m128i test_mm_cmpgt_epi8(__m128i A, __m128i B) { // CHECK: icmp sgt <16 x i8> return _mm_cmpgt_epi8(A, B); } +TEST_CONSTEXPR(match_v16qi(_mm_cmpgt_epi8( + (__m128i)(__v16qs){15,-2,8,-4,12,6,-20,8,25,-10,30,12,-35,14,40,-16}, + (__m128i)(__v16qs){10,-2,6,-4,5,12,-14,8,9,-20,22,12,-26,14,30,-16}), + -1, 0, -1, 0, -1, 0, 0, 0,-1, -1, -1, 0, 0, 0, -1, 0)); __m128i test_mm_cmpgt_epi16(__m128i A, __m128i B) { // CHECK-LABEL: test_mm_cmpgt_epi16 // CHECK: icmp sgt <8 x i16> return _mm_cmpgt_epi16(A, B); } +TEST_CONSTEXPR(match_v8hi(_mm_cmpgt_epi16((__m128i)(__v8hi){15,2,8,4,12,6,20,8}, (__m128i)(__v8hi){10,2,6,4,5,12,14,8}), -1,0,-1,0,-1,0,-1,0)); __m128i test_mm_cmpgt_epi32(__m128i A, __m128i B) { // CHECK-LABEL: test_mm_cmpgt_epi32 // CHECK: icmp sgt <4 x i32> return _mm_cmpgt_epi32(A, B); } +TEST_CONSTEXPR(match_v4si(_mm_cmpgt_epi32((__m128i)(__v4si){15,2,8,4}, (__m128i)(__v4si){10,2,6,4}), -1,0,-1,0)); __m128d test_mm_cmpgt_pd(__m128d A, __m128d B) { // CHECK-LABEL: test_mm_cmpgt_pd @@ -343,18 +365,24 @@ __m128i test_mm_cmplt_epi8(__m128i A, __m128i B) { // CHECK: icmp sgt <16 x i8> return _mm_cmplt_epi8(A, B); } +TEST_CONSTEXPR(match_v16qi(_mm_cmplt_epi8( + (__m128i)(__v16qs){15,-2,8,-4,12,6,-20,8,25,-10,30,12,-35,14,40,-16}, + (__m128i)(__v16qs){10,-2,6,-4,5,12,-14,8,9,-20,22,12,-26,14,30,-16}), + 0, 0, 0, 0, 0, -1, -1, 0,0, 0, 0, 0, -1, 0, 0, 0)); __m128i test_mm_cmplt_epi16(__m128i A, __m128i B) { // CHECK-LABEL: test_mm_cmplt_epi16 // CHECK: icmp sgt <8 x i16> return _mm_cmplt_epi16(A, B); } +TEST_CONSTEXPR(match_v8hi(_mm_cmplt_epi16((__m128i)(__v8hi){5,2,3,4,1,6,7,8}, (__m128i)(__v8hi){10,2,6,4,5,12,14,8}), -1, 0, -1, 0, -1, -1, -1, 0)); __m128i test_mm_cmplt_epi32(__m128i A, __m128i B) { // CHECK-LABEL: test_mm_cmplt_epi32 // CHECK: icmp sgt <4 x i32> return _mm_cmplt_epi32(A, B); } +TEST_CONSTEXPR(match_v4si(_mm_cmplt_epi32((__m128i)(__v4si){5,2,3,4}, (__m128i)(__v4si){10,2,6,4}), -1,0,-1,0)); __m128d test_mm_cmplt_pd(__m128d A, __m128d B) { // CHECK-LABEL: test_mm_cmplt_pd @@ -829,12 +857,16 @@ __m128i test_mm_max_epi16(__m128i A, __m128i B) { return _mm_max_epi16(A, B); } +TEST_CONSTEXPR(match_v8hi(_mm_max_epi16((__m128i)(__v8hi){1, 2, 3, 4, 5, 6, 7, 8}, (__m128i)(__v8hi){1, 2, 3, 5, 8, 12, 20, 32}), 1, 2, 3, 5, 8, 12, 20, 32)); + __m128i test_mm_max_epu8(__m128i A, __m128i B) { // CHECK-LABEL: test_mm_max_epu8 // CHECK: call <16 x i8> @llvm.umax.v16i8(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}) return _mm_max_epu8(A, B); } +TEST_CONSTEXPR(match_v16qu(_mm_max_epu8((__m128i)(__v16qu){9, 10, 11, 12, 13, 14, 15, 16, 1, 2, 3, 4, 5, 6, 7, 8}, (__m128i)(__v16qu){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16}), 9, 10, 11, 12, 13, 14, 15, 16, 9, 10, 11, 12, 13, 14, 15, 16)); + __m128d test_mm_max_pd(__m128d A, __m128d B) { // CHECK-LABEL: test_mm_max_pd // CHECK: call {{.*}}<2 x double> @llvm.x86.sse2.max.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}}) @@ -859,12 +891,16 @@ __m128i test_mm_min_epi16(__m128i A, __m128i B) { return _mm_min_epi16(A, B); } +TEST_CONSTEXPR(match_v8hi(_mm_min_epi16((__m128i)(__v8hi){1, 2, 3, 4, 5, 6, 7, 8}, (__m128i)(__v8hi){1, 2, 3, 5, 8, 12, 20, 32}), 1, 2, 3, 4, 5, 6, 7, 8)); + __m128i test_mm_min_epu8(__m128i A, __m128i B) { // CHECK-LABEL: test_mm_min_epu8 // CHECK: call <16 x i8> @llvm.umin.v16i8(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}) return _mm_min_epu8(A, B); } +TEST_CONSTEXPR(match_v16qu(_mm_min_epu8((__m128i)(__v16qu){9, 10, 11, 12, 13, 14, 15, 16, 1, 2, 3, 4, 5, 6, 7, 8}, (__m128i)(__v16qu){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16}), 1, 2, 3, 4, 5, 6, 7, 8, 1, 2, 3, 4, 5, 6, 7, 8)); + __m128d test_mm_min_pd(__m128d A, __m128d B) { // CHECK-LABEL: test_mm_min_pd // CHECK: call {{.*}}<2 x double> @llvm.x86.sse2.min.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}}) @@ -1296,6 +1332,10 @@ __m128i test_mm_slli_epi16(__m128i A) { // CHECK: call <8 x i16> @llvm.x86.sse2.pslli.w(<8 x i16> %{{.*}}, i32 %{{.*}}) return _mm_slli_epi16(A, 1); } +TEST_CONSTEXPR(match_v8hi(_mm_slli_epi16((__m128i)(__v8hi){0, 1, 2, 3, 4, 5, 6, 7}, 0), 0, 1, 2, 3, 4, 5, 6, 7)); +TEST_CONSTEXPR(match_v8hi(_mm_slli_epi16((__m128i)(__v8hi){0, 1, 2, 3, 4, 5, 6, 7}, 1), 0, 0x2, 0x4, 0x6, 0x8, 0xa, 0xc, 0xe)); +TEST_CONSTEXPR(match_v8hi(_mm_slli_epi16((__m128i)(__v8hi){0, 8, 2, 3, 4, 5, 6, 7}, 8), 0, 0x800, 0x200, 0x300, 0x400, 0x500, 0x600, 0x700)); +TEST_CONSTEXPR(match_v8hi(_mm_slli_epi16((__m128i)(__v8hi){0, 8, 2, 3, 4, 5, 6, 7}, 16), 0, 0, 0, 0, 0, 0, 0, 0)); __m128i test_mm_slli_epi16_1(__m128i A) { // CHECK-LABEL: test_mm_slli_epi16_1 @@ -1314,6 +1354,11 @@ __m128i test_mm_slli_epi32(__m128i A) { // CHECK: call <4 x i32> @llvm.x86.sse2.pslli.d(<4 x i32> %{{.*}}, i32 %{{.*}}) return _mm_slli_epi32(A, 1); } +TEST_CONSTEXPR(match_v4si(_mm_slli_epi32((__m128i)(__v4si){0, 1, 2, 3}, 0), 0, 1, 2, 3)); +TEST_CONSTEXPR(match_v4si(_mm_slli_epi32((__m128i)(__v4si){0, 1, 2, 3}, 1), 0, 0x2, 0x4, 0x6)); +TEST_CONSTEXPR(match_v4su(_mm_slli_epi32((__m128i)(__v4su){0, 1, 2, 3}, 31), 0, 0x80000000, 0x0, 0x80000000)); +TEST_CONSTEXPR(match_v4si(_mm_slli_epi32((__m128i)(__v4si){0, 1, 2, 3}, 32), 0, 0, 0, 0)); +TEST_CONSTEXPR(match_v4si(_mm_slli_epi32((__m128i)(__v4si){0, 1, 2, 3}, 33), 0, 0, 0, 0)); __m128i test_mm_slli_epi32_1(__m128i A) { // CHECK-LABEL: test_mm_slli_epi32_1 @@ -1332,6 +1377,12 @@ __m128i test_mm_slli_epi64(__m128i A) { // CHECK: call {{.*}}<2 x i64> @llvm.x86.sse2.pslli.q(<2 x i64> %{{.*}}, i32 %{{.*}}) return _mm_slli_epi64(A, 1); } +TEST_CONSTEXPR(match_v2di(_mm_slli_epi64((__m128i)(__v2di){0, 1}, 0), 0, 1)); +TEST_CONSTEXPR(match_v2di(_mm_slli_epi64((__m128i)(__v2di){0, 1}, 1), 0, 0x2)); +TEST_CONSTEXPR(match_v2di(_mm_slli_epi64((__m128i)(__v2di){5, 8}, 6), 0x140, 0x200)); +TEST_CONSTEXPR(match_v2du(_mm_slli_epi64((__m128i)(__v2du){0, 1}, 63), 0, 0x8000000000000000ULL)); +TEST_CONSTEXPR(match_v2di(_mm_slli_epi64((__m128i)(__v2di){0, 1}, 64), 0, 0)); +TEST_CONSTEXPR(match_v2di(_mm_slli_epi64((__m128i)(__v2di){0, 1}, 65), 0, 0)); __m128i test_mm_slli_epi64_1(__m128i A) { // CHECK-LABEL: test_mm_slli_epi64_1 @@ -1388,6 +1439,8 @@ __m128i test_mm_srai_epi16(__m128i A) { // CHECK: call <8 x i16> @llvm.x86.sse2.psrai.w(<8 x i16> %{{.*}}, i32 %{{.*}}) return _mm_srai_epi16(A, 1); } +TEST_CONSTEXPR(match_v8hi(_mm_srai_epi16((__m128i)(__v8hi){-32768, 32767, -3, -2, -1, 0, 1, 2}, 1), -16384, 16383, -2, -1, -1, 0, 0, 1)); +TEST_CONSTEXPR(match_v8hi(_mm_srai_epi16((__m128i)(__v8hi){-32768, 32767, -3, -2, -1, 0, 1, 2}, 1), -16384, 16383, -2, -1, -1, 0, 0, 1)); __m128i test_mm_srai_epi16_1(__m128i A) { // CHECK-LABEL: test_mm_srai_epi16_1 @@ -1406,6 +1459,7 @@ __m128i test_mm_srai_epi32(__m128i A) { // CHECK: call <4 x i32> @llvm.x86.sse2.psrai.d(<4 x i32> %{{.*}}, i32 %{{.*}}) return _mm_srai_epi32(A, 1); } +TEST_CONSTEXPR(match_v4si(_mm_srai_epi32((__m128i)(__v4si){-32768, 32767, -3, 2}, 1), -16384, 16383, -2, 1)); __m128i test_mm_srai_epi32_1(__m128i A) { // CHECK-LABEL: test_mm_srai_epi32_1 @@ -1442,6 +1496,7 @@ __m128i test_mm_srli_epi16(__m128i A) { // CHECK: call <8 x i16> @llvm.x86.sse2.psrli.w(<8 x i16> %{{.*}}, i32 %{{.*}}) return _mm_srli_epi16(A, 1); } +TEST_CONSTEXPR(match_v8hi(_mm_srli_epi16((__m128i)(__v8hi){0, 1, 2, 3, 4, 5, 6, 7}, 1), 0, 0x0, 0x1, 0x1, 0x2, 0x2, 0x3, 0x3)); __m128i test_mm_srli_epi16_1(__m128i A) { // CHECK-LABEL: test_mm_srli_epi16_1 @@ -1460,6 +1515,7 @@ __m128i test_mm_srli_epi32(__m128i A) { // CHECK: call <4 x i32> @llvm.x86.sse2.psrli.d(<4 x i32> %{{.*}}, i32 %{{.*}}) return _mm_srli_epi32(A, 1); } +TEST_CONSTEXPR(match_v4si(_mm_srli_epi32((__m128i)(__v4si){0, 1, 2, 3}, 8), 0, 0x0, 0x0, 0x0)); __m128i test_mm_srli_epi32_1(__m128i A) { // CHECK-LABEL: test_mm_srli_epi32_1 @@ -1478,6 +1534,7 @@ __m128i test_mm_srli_epi64(__m128i A) { // CHECK: call {{.*}}<2 x i64> @llvm.x86.sse2.psrli.q(<2 x i64> %{{.*}}, i32 %{{.*}}) return _mm_srli_epi64(A, 1); } +TEST_CONSTEXPR(match_v2di(_mm_srli_epi64((__m128i)(__v2di){100005, 100008}, 6), 0x61a, 0x61a)); __m128i test_mm_srli_epi64_1(__m128i A) { // CHECK-LABEL: test_mm_srli_epi64_1 |