Skip to content

Commit 539b86e

Browse files
ahmednoursphinxahmednoursphinx
committed
chore: Move tests from avx512vl-builtins
1 parent b4101ff commit 539b86e

File tree

2 files changed

+32
-44
lines changed

2 files changed

+32
-44
lines changed

‎clang/test/CodeGen/X86/avx512vl-builtins.c‎

Lines changed: 16 additions & 44 deletions
Original file line numberDiff line numberDiff line change
@@ -7,50 +7,6 @@
77
#include <immintrin.h>
88
#include "builtin_test_helpers.h"
99

10-
// Constexpr coverage for VLDQ broadcast features (i32x2/f32x2/f64x2/i64x2) that don't have test functions in this file.
11-
// The corresponding test_mm*_*_broadcast_* functions are in avx512vldq-builtins.c.
12-
TEST_CONSTEXPR(match_v4si(_mm_mask_broadcast_i32x2(_mm_setzero_si128(), 0xF, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1));
13-
14-
TEST_CONSTEXPR(match_m256(_mm256_mask_broadcast_f32x2(_mm256_setzero_ps(), 0xFF, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 1.f,2.f,1.f,2.f,1.f,2.f,1.f,2.f));
15-
16-
TEST_CONSTEXPR(match_m256(_mm256_maskz_broadcast_f32x2(0xFF, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 1.f,2.f,1.f,2.f,1.f,2.f,1.f,2.f));
17-
18-
TEST_CONSTEXPR(match_m256d(_mm256_mask_broadcast_f64x2(_mm256_setzero_pd(), 0xFF, (__m128d)(__v2df){1.0,2.0}), 1.0,2.0,1.0,2.0));
19-
20-
TEST_CONSTEXPR(match_m256d(_mm256_maskz_broadcast_f64x2(0xFF, (__m128d)(__v2df){1.0,2.0}), 1.0,2.0,1.0,2.0));
21-
22-
TEST_CONSTEXPR(match_v4di(_mm256_maskz_broadcast_i64x2(0xF, (__m128i)(__v2di){1,2}), 1,2,1,2));
23-
24-
25-
// i32x2 maskz (128/256)
26-
27-
TEST_CONSTEXPR(match_v4si(_mm_maskz_broadcast_i32x2(0xF, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1));
28-
29-
30-
31-
TEST_CONSTEXPR(match_v8si(_mm256_maskz_broadcast_i32x2(0xFF, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1,0,1,0,1));
32-
33-
34-
// unpackhi/lo with full mask behaves like the underlying unpack
35-
36-
TEST_CONSTEXPR(match_m128d(_mm_mask_unpackhi_pd(_mm_setzero_pd(), 0x3, (__m128d)(__v2df){1.0,2.0}, (__m128d)(__v2df){3.0,4.0}), 2.0,4.0));
37-
TEST_CONSTEXPR(match_m128d(_mm_mask_unpacklo_pd(_mm_setzero_pd(), 0x3, (__m128d)(__v2df){1.0,2.0}, (__m128d)(__v2df){3.0,4.0}), 1.0,3.0));
38-
39-
40-
TEST_CONSTEXPR(match_m256d(_mm256_mask_unpackhi_pd(_mm256_setzero_pd(), 0xFF, (__m256d)(__v4df){1.0,2.0,3.0,4.0}, (__m256d)(__v4df){5.0,6.0,7.0,8.0}), 2.0,6.0,4.0,8.0));
41-
TEST_CONSTEXPR(match_m256d(_mm256_mask_unpacklo_pd(_mm256_setzero_pd(), 0xFF, (__m256d)(__v4df){1.0,2.0,3.0,4.0}, (__m256d)(__v4df){5.0,6.0,7.0,8.0}), 1.0,5.0,3.0,7.0));
42-
43-
44-
// movehdup/moveldup with full mask equals the underlying *_move*dup
45-
46-
TEST_CONSTEXPR(match_m128(_mm_mask_movehdup_ps(_mm_setzero_ps(), 0xF, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 2.f,2.f,4.f,4.f));
47-
TEST_CONSTEXPR(match_m128(_mm_mask_moveldup_ps(_mm_setzero_ps(), 0xF, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 1.f,1.f,3.f,3.f));
48-
49-
50-
TEST_CONSTEXPR(match_m256(_mm256_mask_movehdup_ps(_mm256_setzero_ps(), 0xFF, (__m256)(__v8sf){1,2,3,4,5,6,7,8}), 2,2,4,4,6,6,8,8));
51-
TEST_CONSTEXPR(match_m256(_mm256_mask_moveldup_ps(_mm256_setzero_ps(), 0xFF, (__m256)(__v8sf){1,2,3,4,5,6,7,8}), 1,1,3,3,5,5,7,7));
52-
53-
5410
__mmask8 test_mm_cmpeq_epu32_mask(__m128i __a, __m128i __b) {
5511
// CHECK-LABEL: test_mm_cmpeq_epu32_mask
5612
// CHECK: icmp eq <4 x i32> %{{.*}}, %{{.*}}
@@ -7675,6 +7631,8 @@ __m128d test_mm_mask_unpackhi_pd(__m128d __W, __mmask8 __U, __m128d __A, __m128d
76757631
return _mm_mask_unpackhi_pd(__W, __U, __A, __B);
76767632
}
76777633

7634+
TEST_CONSTEXPR(match_m128d(_mm_mask_unpackhi_pd(_mm_setzero_pd(), 0x3, (__m128d)(__v2df){1.0,2.0}, (__m128d)(__v2df){3.0,4.0}), 2.0,4.0));
7635+
76787636
__m128d test_mm_maskz_unpackhi_pd(__mmask8 __U, __m128d __A, __m128d __B) {
76797637
// CHECK-LABEL: test_mm_maskz_unpackhi_pd
76807638
// CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x i32> <i32 1, i32 3>
@@ -7689,6 +7647,8 @@ __m256d test_mm256_mask_unpackhi_pd(__m256d __W, __mmask8 __U, __m256d __A, __m2
76897647
return _mm256_mask_unpackhi_pd(__W, __U, __A, __B);
76907648
}
76917649

7650+
TEST_CONSTEXPR(match_m256d(_mm256_mask_unpackhi_pd(_mm256_setzero_pd(), 0xFF, (__m256d)(__v4df){1.0,2.0,3.0,4.0}, (__m256d)(__v4df){5.0,6.0,7.0,8.0}), 2.0,6.0,4.0,8.0));
7651+
76927652
__m256d test_mm256_maskz_unpackhi_pd(__mmask8 __U, __m256d __A, __m256d __B) {
76937653
// CHECK-LABEL: test_mm256_maskz_unpackhi_pd
76947654
// CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
@@ -7731,6 +7691,8 @@ __m128d test_mm_mask_unpacklo_pd(__m128d __W, __mmask8 __U, __m128d __A, __m128d
77317691
return _mm_mask_unpacklo_pd(__W, __U, __A, __B);
77327692
}
77337693

7694+
TEST_CONSTEXPR(match_m128d(_mm_mask_unpacklo_pd(_mm_setzero_pd(), 0x3, (__m128d)(__v2df){1.0,2.0}, (__m128d)(__v2df){3.0,4.0}), 1.0,3.0));
7695+
77347696
__m128d test_mm_maskz_unpacklo_pd(__mmask8 __U, __m128d __A, __m128d __B) {
77357697
// CHECK-LABEL: test_mm_maskz_unpacklo_pd
77367698
// CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x i32> <i32 0, i32 2>
@@ -7745,6 +7707,8 @@ __m256d test_mm256_mask_unpacklo_pd(__m256d __W, __mmask8 __U, __m256d __A, __m2
77457707
return _mm256_mask_unpacklo_pd(__W, __U, __A, __B);
77467708
}
77477709

7710+
TEST_CONSTEXPR(match_m256d(_mm256_mask_unpacklo_pd(_mm256_setzero_pd(), 0xFF, (__m256d)(__v4df){1.0,2.0,3.0,4.0}, (__m256d)(__v4df){5.0,6.0,7.0,8.0}), 1.0,5.0,3.0,7.0));
7711+
77487712
__m256d test_mm256_maskz_unpacklo_pd(__mmask8 __U, __m256d __A, __m256d __B) {
77497713
// CHECK-LABEL: test_mm256_maskz_unpacklo_pd
77507714
// CHECK: shufflevector <4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x i32> <i32 0, i32 4, i32 2, i32 6>
@@ -10380,6 +10344,8 @@ __m128 test_mm_mask_movehdup_ps(__m128 __W, __mmask8 __U, __m128 __A) {
1038010344
return _mm_mask_movehdup_ps(__W, __U, __A);
1038110345
}
1038210346

10347+
TEST_CONSTEXPR(match_m128(_mm_mask_movehdup_ps(_mm_setzero_ps(), 0xF, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 2.f,2.f,4.f,4.f));
10348+
1038310349
__m128 test_mm_maskz_movehdup_ps(__mmask8 __U, __m128 __A) {
1038410350
// CHECK-LABEL: test_mm_maskz_movehdup_ps
1038510351
// CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
@@ -10394,6 +10360,8 @@ __m256 test_mm256_mask_movehdup_ps(__m256 __W, __mmask8 __U, __m256 __A) {
1039410360
return _mm256_mask_movehdup_ps(__W, __U, __A);
1039510361
}
1039610362

10363+
TEST_CONSTEXPR(match_m256(_mm256_mask_movehdup_ps(_mm256_setzero_ps(), 0xFF, (__m256)(__v8sf){1,2,3,4,5,6,7,8}), 2,2,4,4,6,6,8,8));
10364+
1039710365
__m256 test_mm256_maskz_movehdup_ps(__mmask8 __U, __m256 __A) {
1039810366
// CHECK-LABEL: test_mm256_maskz_movehdup_ps
1039910367
// CHECK: shufflevector <8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7>
@@ -10408,6 +10376,8 @@ __m128 test_mm_mask_moveldup_ps(__m128 __W, __mmask8 __U, __m128 __A) {
1040810376
return _mm_mask_moveldup_ps(__W, __U, __A);
1040910377
}
1041010378

10379+
TEST_CONSTEXPR(match_m128(_mm_mask_moveldup_ps(_mm_setzero_ps(), 0xF, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 1.f,1.f,3.f,3.f));
10380+
1041110381
__m128 test_mm_maskz_moveldup_ps(__mmask8 __U, __m128 __A) {
1041210382
// CHECK-LABEL: test_mm_maskz_moveldup_ps
1041310383
// CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
@@ -10422,6 +10392,8 @@ __m256 test_mm256_mask_moveldup_ps(__m256 __W, __mmask8 __U, __m256 __A) {
1042210392
return _mm256_mask_moveldup_ps(__W, __U, __A);
1042310393
}
1042410394

10395+
TEST_CONSTEXPR(match_m256(_mm256_mask_moveldup_ps(_mm256_setzero_ps(), 0xFF, (__m256)(__v8sf){1,2,3,4,5,6,7,8}), 1,1,3,3,5,5,7,7));
10396+
1042510397
__m256 test_mm256_maskz_moveldup_ps(__mmask8 __U, __m256 __A) {
1042610398
// CHECK-LABEL: test_mm256_maskz_moveldup_ps
1042710399
// CHECK: shufflevector <8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>

‎clang/test/CodeGen/X86/avx512vldq-builtins.c‎

Lines changed: 16 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -987,13 +987,17 @@ __m256 test_mm256_mask_broadcast_f32x2(__m256 __O, __mmask8 __M, __m128 __A) {
987987
return _mm256_mask_broadcast_f32x2(__O, __M, __A);
988988
}
989989

990+
TEST_CONSTEXPR(match_m256(_mm256_mask_broadcast_f32x2(_mm256_setzero_ps(), 0xFF, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 1.f,2.f,1.f,2.f,1.f,2.f,1.f,2.f));
991+
990992
__m256 test_mm256_maskz_broadcast_f32x2(__mmask8 __M, __m128 __A) {
991993
// CHECK-LABEL: test_mm256_maskz_broadcast_f32x2
992994
// CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
993995
// CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
994996
return _mm256_maskz_broadcast_f32x2(__M, __A);
995997
}
996998

999+
TEST_CONSTEXPR(match_m256(_mm256_maskz_broadcast_f32x2(0xFF, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 1.f,2.f,1.f,2.f,1.f,2.f,1.f,2.f));
1000+
9971001
__m256d test_mm256_broadcast_f64x2(double const* __A) {
9981002
// CHECK-LABEL: test_mm256_broadcast_f64x2
9991003
// CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
@@ -1008,13 +1012,17 @@ __m256d test_mm256_mask_broadcast_f64x2(__m256d __O, __mmask8 __M, double const*
10081012
return _mm256_mask_broadcast_f64x2(__O, __M, _mm_loadu_pd(__A));
10091013
}
10101014

1015+
TEST_CONSTEXPR(match_m256d(_mm256_mask_broadcast_f64x2(_mm256_setzero_pd(), 0xFF, (__m128d)(__v2df){1.0,2.0}), 1.0,2.0,1.0,2.0));
1016+
10111017
__m256d test_mm256_maskz_broadcast_f64x2(__mmask8 __M, double const* __A) {
10121018
// CHECK-LABEL: test_mm256_maskz_broadcast_f64x2
10131019
// CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
10141020
// CHECK: select <4 x i1> %{{.*}}, <4 x double> %{{.*}}, <4 x double> %{{.*}}
10151021
return _mm256_maskz_broadcast_f64x2(__M, _mm_loadu_pd(__A));
10161022
}
10171023

1024+
TEST_CONSTEXPR(match_m256d(_mm256_maskz_broadcast_f64x2(0xFF, (__m128d)(__v2df){1.0,2.0}), 1.0,2.0,1.0,2.0));
1025+
10181026
__m128i test_mm_broadcast_i32x2(__m128i __A) {
10191027
// CHECK-LABEL: test_mm_broadcast_i32x2
10201028
// CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
@@ -1029,13 +1037,17 @@ __m128i test_mm_mask_broadcast_i32x2(__m128i __O, __mmask8 __M, __m128i __A) {
10291037
return _mm_mask_broadcast_i32x2(__O, __M, __A);
10301038
}
10311039

1040+
TEST_CONSTEXPR(match_v4si(_mm_mask_broadcast_i32x2(_mm_setzero_si128(), 0xF, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1));
1041+
10321042
__m128i test_mm_maskz_broadcast_i32x2(__mmask8 __M, __m128i __A) {
10331043
// CHECK-LABEL: test_mm_maskz_broadcast_i32x2
10341044
// CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
10351045
// CHECK: select <4 x i1> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}
10361046
return _mm_maskz_broadcast_i32x2(__M, __A);
10371047
}
10381048

1049+
TEST_CONSTEXPR(match_v4si(_mm_maskz_broadcast_i32x2(0xF, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1));
1050+
10391051
__m256i test_mm256_broadcast_i32x2(__m128i __A) {
10401052
// CHECK-LABEL: test_mm256_broadcast_i32x2
10411053
// CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
@@ -1057,6 +1069,8 @@ __m256i test_mm256_maskz_broadcast_i32x2(__mmask8 __M, __m128i __A) {
10571069
return _mm256_maskz_broadcast_i32x2(__M, __A);
10581070
}
10591071

1072+
TEST_CONSTEXPR(match_v8si(_mm256_maskz_broadcast_i32x2(0xFF, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1,0,1,0,1));
1073+
10601074
__m256i test_mm256_broadcast_i64x2(__m128i const* __A) {
10611075
// CHECK-LABEL: test_mm256_broadcast_i64x2
10621076
// CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
@@ -1078,6 +1092,8 @@ __m256i test_mm256_maskz_broadcast_i64x2(__mmask8 __M, __m128i const* __A) {
10781092
return _mm256_maskz_broadcast_i64x2(__M, _mm_loadu_si128(__A));
10791093
}
10801094

1095+
TEST_CONSTEXPR(match_v4di(_mm256_maskz_broadcast_i64x2(0xF, (__m128i)(__v2di){1,2}), 1,2,1,2));
1096+
10811097
__m128d test_mm256_extractf64x2_pd(__m256d __A) {
10821098
// CHECK-LABEL: test_mm256_extractf64x2_pd
10831099
// CHECK: shufflevector <4 x double> %{{.*}}, <4 x double> poison, <2 x i32> <i32 2, i32 3>

0 commit comments

Comments
 (0)