diff --git a/gcc/config/i386/avx512fintrin.h b/gcc/config/i386/avx512fintrin.h index 016b01b..c0012a9 100644 --- a/gcc/config/i386/avx512fintrin.h +++ b/gcc/config/i386/avx512fintrin.h @@ -10774,6 +10774,29 @@ _mm_max_round_sd (__m128d __A, __m128d __B, const int __R) __R); } +extern __inline __m128d +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm_mask_max_round_sd (__m128d __W, __mmask8 __U, __m128d __A, + __m128d __B, const int __R) +{ + return (__m128d) __builtin_ia32_maxsd_mask_round ((__v2df) __A, + (__v2df) __B, + (__v2df) __W, + (__mmask8) __U, __R); +} + +extern __inline __m128d +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm_maskz_max_round_sd (__mmask8 __U, __m128d __A, __m128d __B, + const int __R) +{ + return (__m128d) __builtin_ia32_maxsd_mask_round ((__v2df) __A, + (__v2df) __B, + (__v2df) + _mm_setzero_pd (), + (__mmask8) __U, __R); +} + extern __inline __m128 __attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) _mm_max_round_ss (__m128 __A, __m128 __B, const int __R) @@ -10783,6 +10806,29 @@ _mm_max_round_ss (__m128 __A, __m128 __B, const int __R) __R); } +extern __inline __m128 +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm_mask_max_round_ss (__m128 __W, __mmask8 __U, __m128 __A, + __m128 __B, const int __R) +{ + return (__m128) __builtin_ia32_maxss_mask_round ((__v4sf) __A, + (__v4sf) __B, + (__v4sf) __W, + (__mmask8) __U, __R); +} + +extern __inline __m128 +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm_maskz_max_round_ss (__mmask8 __U, __m128 __A, __m128 __B, + const int __R) +{ + return (__m128) __builtin_ia32_maxss_mask_round ((__v4sf) __A, + (__v4sf) __B, + (__v4sf) + _mm_setzero_ps (), + (__mmask8) __U, __R); +} + extern __inline __m128d __attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) _mm_min_round_sd (__m128d __A, __m128d __B, const int __R) @@ -10792,6 +10838,29 @@ _mm_min_round_sd (__m128d __A, __m128d __B, const int __R) __R); } +extern __inline __m128d +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm_mask_min_round_sd (__m128d __W, __mmask8 __U, __m128d __A, + __m128d __B, const int __R) +{ + return (__m128d) __builtin_ia32_minsd_mask_round ((__v2df) __A, + (__v2df) __B, + (__v2df) __W, + (__mmask8) __U, __R); +} + +extern __inline __m128d +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm_maskz_min_round_sd (__mmask8 __U, __m128d __A, __m128d __B, + const int __R) +{ + return (__m128d) __builtin_ia32_minsd_mask_round ((__v2df) __A, + (__v2df) __B, + (__v2df) + _mm_setzero_pd (), + (__mmask8) __U, __R); +} + extern __inline __m128 __attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) _mm_min_round_ss (__m128 __A, __m128 __B, const int __R) @@ -10801,18 +10870,66 @@ _mm_min_round_ss (__m128 __A, __m128 __B, const int __R) __R); } +extern __inline __m128 +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm_mask_min_round_ss (__m128 __W, __mmask8 __U, __m128 __A, + __m128 __B, const int __R) +{ + return (__m128) __builtin_ia32_minss_mask_round ((__v4sf) __A, + (__v4sf) __B, + (__v4sf) __W, + (__mmask8) __U, __R); +} + +extern __inline __m128 +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) +_mm_maskz_min_round_ss (__mmask8 __U, __m128 __A, __m128 __B, + const int __R) +{ + return (__m128) __builtin_ia32_minss_mask_round ((__v4sf) __A, + (__v4sf) __B, + (__v4sf) + _mm_setzero_ps (), + (__mmask8) __U, __R); +} + #else #define _mm_max_round_sd(A, B, C) \ - (__m128d)__builtin_ia32_addsd_round(A, B, C) + (__m128d)__builtin_ia32_maxsd_round(A, B, C) + +#define _mm_mask_max_round_sd(W, U, A, B, C) \ + (__m128d)__builtin_ia32_maxsd_mask_round(A, B, W, U, C) + +#define _mm_maskz_max_round_sd(U, A, B, C) \ + (__m128d)__builtin_ia32_maxsd_mask_round(A, B, (__v2df)_mm_setzero_pd(), U, C) #define _mm_max_round_ss(A, B, C) \ - (__m128)__builtin_ia32_addss_round(A, B, C) + (__m128)__builtin_ia32_maxss_round(A, B, C) + +#define _mm_mask_max_round_ss(W, U, A, B, C) \ + (__m128)__builtin_ia32_maxss_mask_round(A, B, W, U, C) + +#define _mm_maskz_max_round_ss(U, A, B, C) \ + (__m128)__builtin_ia32_maxss_mask_round(A, B, (__v4sf)_mm_setzero_ps(), U, C) #define _mm_min_round_sd(A, B, C) \ - (__m128d)__builtin_ia32_subsd_round(A, B, C) + (__m128d)__builtin_ia32_minsd_round(A, B, C) + +#define _mm_mask_min_round_sd(W, U, A, B, C) \ + (__m128d)__builtin_ia32_minsd_mask_round(A, B, W, U, C) + +#define _mm_maskz_min_round_sd(U, A, B, C) \ + (__m128d)__builtin_ia32_minsd_mask_round(A, B, (__v2df)_mm_setzero_pd(), U, C) #define _mm_min_round_ss(A, B, C) \ - (__m128)__builtin_ia32_subss_round(A, B, C) + (__m128)__builtin_ia32_minss_round(A, B, C) + +#define _mm_mask_min_round_ss(W, U, A, B, C) \ + (__m128)__builtin_ia32_minss_mask_round(A, B, W, U, C) + +#define _mm_maskz_min_round_ss(U, A, B, C) \ + (__m128)__builtin_ia32_minss_mask_round(A, B, (__v4sf)_mm_setzero_ps(), U, C) + #endif extern __inline __m512d diff --git a/gcc/config/i386/i386-builtin.def b/gcc/config/i386/i386-builtin.def index fce32ef..c94abd3 100644 --- a/gcc/config/i386/i386-builtin.def +++ b/gcc/config/i386/i386-builtin.def @@ -2442,11 +2442,15 @@ BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_vgetmantv4sf_round, "__builtin_ BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_smaxv8df3_mask_round, "__builtin_ia32_maxpd512_mask", IX86_BUILTIN_MAXPD512, UNKNOWN, (int) V8DF_FTYPE_V8DF_V8DF_V8DF_UQI_INT) BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_smaxv16sf3_mask_round, "__builtin_ia32_maxps512_mask", IX86_BUILTIN_MAXPS512, UNKNOWN, (int) V16SF_FTYPE_V16SF_V16SF_V16SF_HI_INT) BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sse2_vmsmaxv2df3_round, "__builtin_ia32_maxsd_round", IX86_BUILTIN_MAXSD_ROUND, UNKNOWN, (int) V2DF_FTYPE_V2DF_V2DF_INT) +BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sse2_vmsmaxv2df3_mask_round, "__builtin_ia32_maxsd_mask_round", IX86_BUILTIN_MAXSD_MASK_ROUND, UNKNOWN, (int) V2DF_FTYPE_V2DF_V2DF_V2DF_UQI_INT) BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sse_vmsmaxv4sf3_round, "__builtin_ia32_maxss_round", IX86_BUILTIN_MAXSS_ROUND, UNKNOWN, (int) V4SF_FTYPE_V4SF_V4SF_INT) +BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sse_vmsmaxv4sf3_mask_round, "__builtin_ia32_maxss_mask_round", IX86_BUILTIN_MAXSS_MASK_ROUND, UNKNOWN, (int) V4SF_FTYPE_V4SF_V4SF_V4SF_UQI_INT) BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sminv8df3_mask_round, "__builtin_ia32_minpd512_mask", IX86_BUILTIN_MINPD512, UNKNOWN, (int) V8DF_FTYPE_V8DF_V8DF_V8DF_UQI_INT) BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sminv16sf3_mask_round, "__builtin_ia32_minps512_mask", IX86_BUILTIN_MINPS512, UNKNOWN, (int) V16SF_FTYPE_V16SF_V16SF_V16SF_HI_INT) BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sse2_vmsminv2df3_round, "__builtin_ia32_minsd_round", IX86_BUILTIN_MINSD_ROUND, UNKNOWN, (int) V2DF_FTYPE_V2DF_V2DF_INT) +BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sse2_vmsminv2df3_mask_round, "__builtin_ia32_minsd_mask_round", IX86_BUILTIN_MINSD_MASK_ROUND, UNKNOWN, (int) V2DF_FTYPE_V2DF_V2DF_V2DF_UQI_INT) BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sse_vmsminv4sf3_round, "__builtin_ia32_minss_round", IX86_BUILTIN_MINSS_ROUND, UNKNOWN, (int) V4SF_FTYPE_V4SF_V4SF_INT) +BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sse_vmsminv4sf3_mask_round, "__builtin_ia32_minss_mask_round", IX86_BUILTIN_MINSS_MASK_ROUND, UNKNOWN, (int) V4SF_FTYPE_V4SF_V4SF_V4SF_UQI_INT) BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_mulv8df3_mask_round, "__builtin_ia32_mulpd512_mask", IX86_BUILTIN_MULPD512, UNKNOWN, (int) V8DF_FTYPE_V8DF_V8DF_V8DF_UQI_INT) BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_mulv16sf3_mask_round, "__builtin_ia32_mulps512_mask", IX86_BUILTIN_MULPS512, UNKNOWN, (int) V16SF_FTYPE_V16SF_V16SF_V16SF_HI_INT) BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sse2_vmmulv2df3_round, "__builtin_ia32_mulsd_round", IX86_BUILTIN_MULSD_ROUND, UNKNOWN, (int) V2DF_FTYPE_V2DF_V2DF_INT) diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md index 850970f..cf9f102 100644 --- a/gcc/config/i386/sse.md +++ b/gcc/config/i386/sse.md @@ -1910,7 +1910,7 @@ (set_attr "prefix" "") (set_attr "mode" "")]) -(define_insn "_vm3" +(define_insn "_vm3" [(set (match_operand:VF_128 0 "register_operand" "=x,v") (vec_merge:VF_128 (smaxmin:VF_128 @@ -1921,7 +1921,7 @@ "TARGET_SSE" "@ \t{%2, %0|%0, %2} - v\t{%2, %1, %0|%0, %1, %2}" + v\t{%2, %1, %0|%0, %1, %2}" [(set_attr "isa" "noavx,avx") (set_attr "type" "sse") (set_attr "btver2_sse_attr" "maxmin") diff --git a/gcc/testsuite/gcc.target/i386/avx-1.c b/gcc/testsuite/gcc.target/i386/avx-1.c index 6672039..8377555 100644 --- a/gcc/testsuite/gcc.target/i386/avx-1.c +++ b/gcc/testsuite/gcc.target/i386/avx-1.c @@ -245,11 +245,15 @@ #define __builtin_ia32_maxpd512_mask(A, B, C, D, E) __builtin_ia32_maxpd512_mask(A, B, C, D, 8) #define __builtin_ia32_maxps512_mask(A, B, C, D, E) __builtin_ia32_maxps512_mask(A, B, C, D, 8) #define __builtin_ia32_maxsd_round(A, B, C) __builtin_ia32_maxsd_round(A, B, 4) +#define __builtin_ia32_maxsd_mask_round(A, B, C, D, E) __builtin_ia32_maxsd_mask_round(A, B, C, D, 8) #define __builtin_ia32_maxss_round(A, B, C) __builtin_ia32_maxss_round(A, B, 4) +#define __builtin_ia32_maxss_mask_round(A, B, C, D, E) __builtin_ia32_maxss_mask_round(A, B, C, D, 8) #define __builtin_ia32_minpd512_mask(A, B, C, D, E) __builtin_ia32_minpd512_mask(A, B, C, D, 8) #define __builtin_ia32_minps512_mask(A, B, C, D, E) __builtin_ia32_minps512_mask(A, B, C, D, 8) #define __builtin_ia32_minsd_round(A, B, C) __builtin_ia32_minsd_round(A, B, 4) +#define __builtin_ia32_minsd_mask_round(A, B, C, D, E) __builtin_ia32_minsd_mask_round(A, B, C, D, 4) #define __builtin_ia32_minss_round(A, B, C) __builtin_ia32_minss_round(A, B, 4) +#define __builtin_ia32_minss_mask_round(A, B, C, D, E) __builtin_ia32_minss_mask_round(A, B, C, D, 4) #define __builtin_ia32_mulpd512_mask(A, B, C, D, E) __builtin_ia32_mulpd512_mask(A, B, C, D, 8) #define __builtin_ia32_mulps512_mask(A, B, C, D, E) __builtin_ia32_mulps512_mask(A, B, C, D, 8) #define __builtin_ia32_mulsd_round(A, B, C) __builtin_ia32_mulsd_round(A, B, 8) diff --git a/gcc/testsuite/gcc.target/i386/avx512f-vmaxsd-1.c b/gcc/testsuite/gcc.target/i386/avx512f-vmaxsd-1.c index 20a0a3d..c1c8f8d 100644 --- a/gcc/testsuite/gcc.target/i386/avx512f-vmaxsd-1.c +++ b/gcc/testsuite/gcc.target/i386/avx512f-vmaxsd-1.c @@ -1,13 +1,18 @@ /* { dg-do compile } */ /* { dg-options "-mavx512f -O2" } */ /* { dg-final { scan-assembler-times "vmaxsd\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vmaxsd\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vmaxsd\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */ #include -volatile __m128d x1, x2; +volatile __m128d x1, x2, x3; +volatile __mmask8 m; void extern avx512f_test (void) { x1 = _mm_max_round_sd (x1, x2, _MM_FROUND_NO_EXC); + x1 = _mm_mask_max_round_sd (x1, m, x2, x3, _MM_FROUND_NO_EXC); + x1 = _mm_maskz_max_round_sd (m, x1, x2, _MM_FROUND_NO_EXC); } diff --git a/gcc/testsuite/gcc.target/i386/avx512f-vmaxsd-2.c b/gcc/testsuite/gcc.target/i386/avx512f-vmaxsd-2.c new file mode 100644 index 0000000..29db777 --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/avx512f-vmaxsd-2.c @@ -0,0 +1,54 @@ +/* { dg-do run } */ +/* { dg-options "-O2 -mavx512f" } */ +/* { dg-require-effective-target avx512f } */ + +#include "avx512f-check.h" + +#include "avx512f-helper.h" + +#define SIZE (128 / 64) +#include "avx512f-mask-type.h" + +static void +calc_max (double *r, double *s1, double *s2) +{ + r[0] = s1[0] > s2[0] ? s1[0] : s2[0]; + r[1] = s1[1]; +} + +void +avx512f_test (void) +{ + int i, sign; + union128d res1, res2, res3, src1, src2; + MASK_TYPE mask = MASK_VALUE; + double res_ref[SIZE]; + + sign = -1; + for (i = 0; i < SIZE; i++) + { + src1.a[i] = 1.5 + 34.67 * i * sign; + src2.a[i] = -22.17 * i * sign + 1.0; + sign = sign * -1; + } + for (i = 0; i < SIZE; i++) + res2.a[i] = DEFAULT_VALUE; + + res1.x = _mm_max_round_sd (src1.x, src2.x, _MM_FROUND_NO_EXC); + res2.x = _mm_mask_max_round_sd (res2.x, mask, src1.x, src2.x, _MM_FROUND_NO_EXC); + res3.x = _mm_maskz_max_round_sd (mask, src1.x, src2.x, _MM_FROUND_NO_EXC); + + calc_max (res_ref, src1.a, src2.a); + + if (check_union128d (res1, res_ref)) + abort(); + + MASK_MERGE (d) (res_ref, mask, 1); + if (check_union128d (res2, res_ref)) + abort (); + + MASK_ZERO (d) (res_ref, mask, 1); + if (check_union128d (res3, res_ref)) + abort (); +} + diff --git a/gcc/testsuite/gcc.target/i386/avx512f-vmaxss-1.c b/gcc/testsuite/gcc.target/i386/avx512f-vmaxss-1.c index 0c692aa..ecd8757 100644 --- a/gcc/testsuite/gcc.target/i386/avx512f-vmaxss-1.c +++ b/gcc/testsuite/gcc.target/i386/avx512f-vmaxss-1.c @@ -1,13 +1,18 @@ /* { dg-do compile } */ /* { dg-options "-mavx512f -O2" } */ /* { dg-final { scan-assembler-times "vmaxss\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vmaxss\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vmaxss\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */ #include -volatile __m128 x1, x2; +volatile __m128 x1, x2, x3; +volatile __mmask8 m; void extern avx512f_test (void) { x1 = _mm_max_round_ss (x1, x2, _MM_FROUND_NO_EXC); + x1 = _mm_mask_max_round_ss (x1, m, x2, x3, _MM_FROUND_NO_EXC); + x1 = _mm_maskz_max_round_ss (m, x1, x2, _MM_FROUND_NO_EXC); } diff --git a/gcc/testsuite/gcc.target/i386/avx512f-vmaxss-2.c b/gcc/testsuite/gcc.target/i386/avx512f-vmaxss-2.c new file mode 100644 index 0000000..7676ae5 --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/avx512f-vmaxss-2.c @@ -0,0 +1,58 @@ +/* { dg-do run } */ +/* { dg-options "-O2 -mavx512f" } */ +/* { dg-require-effective-target avx512f } */ + +#include "avx512f-check.h" + +#include "avx512f-helper.h" + +#define SIZE (128 / 32) +#include "avx512f-mask-type.h" + +static void +calc_max (float *r, float *s1, float *s2) +{ + r[0] = s1[0] > s2[0] ? s1[0] : s2[0]; + int i; + for (i = 1; i < SIZE; i++) + { + r[i] = s1[i]; + } +} + +void +avx512f_test (void) +{ + int i, sign; + union128 res1, res2, res3, src1, src2; + MASK_TYPE mask = MASK_VALUE; + float res_ref[SIZE]; + + sign = -1; + for (i = 0; i < SIZE; i++) + { + src1.a[i] = 1.5 + 34.67 * i * sign; + src2.a[i] = -22.17 * i * sign + 1.0; + sign = sign * -1; + } + for (i = 0; i < SIZE; i++) + res2.a[i] = DEFAULT_VALUE; + + res1.x = _mm_max_round_ss (src1.x, src2.x, _MM_FROUND_NO_EXC); + res2.x = _mm_mask_max_round_ss (res2.x, mask, src1.x, src2.x, _MM_FROUND_NO_EXC); + res3.x = _mm_maskz_max_round_ss (mask, src1.x, src2.x, _MM_FROUND_NO_EXC); + + calc_max (res_ref, src1.a, src2.a); + + if (check_union128 (res1, res_ref)) + abort(); + + MASK_MERGE () (res_ref, mask, 1); + if (check_union128 (res2, res_ref)) + abort (); + + MASK_ZERO () (res_ref, mask, 1); + if (check_union128 (res3, res_ref)) + abort (); +} + diff --git a/gcc/testsuite/gcc.target/i386/avx512f-vminsd-1.c b/gcc/testsuite/gcc.target/i386/avx512f-vminsd-1.c index 29c0948..22ada87 100644 --- a/gcc/testsuite/gcc.target/i386/avx512f-vminsd-1.c +++ b/gcc/testsuite/gcc.target/i386/avx512f-vminsd-1.c @@ -1,13 +1,18 @@ /* { dg-do compile } */ /* { dg-options "-mavx512f -O2" } */ /* { dg-final { scan-assembler-times "vminsd\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vminsd\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vminsd\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */ #include -volatile __m128d x1, x2; +volatile __m128d x1, x2, x3; +volatile __mmask8 m; void extern avx512f_test (void) { x1 = _mm_min_round_sd (x1, x2, _MM_FROUND_NO_EXC); + x1 = _mm_mask_min_round_sd (x1, m, x2, x3, _MM_FROUND_NO_EXC); + x1 = _mm_maskz_min_round_sd (m, x1, x2, _MM_FROUND_NO_EXC); } diff --git a/gcc/testsuite/gcc.target/i386/avx512f-vminsd-2.c b/gcc/testsuite/gcc.target/i386/avx512f-vminsd-2.c new file mode 100644 index 0000000..79a051f --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/avx512f-vminsd-2.c @@ -0,0 +1,54 @@ +/* { dg-do run } */ +/* { dg-options "-O2 -mavx512f" } */ +/* { dg-require-effective-target avx512f } */ + +#include "avx512f-check.h" + +#include "avx512f-helper.h" + +#define SIZE (128 / 64) +#include "avx512f-mask-type.h" + +static void +calc_min (double *r, double *s1, double *s2) +{ + r[0] = s1[0] < s2[0] ? s1[0] : s2[0]; + r[1] = s1[1]; +} + +void +avx512f_test (void) +{ + int i, sign; + union128d res1, res2, res3, src1, src2; + MASK_TYPE mask = MASK_VALUE; + double res_ref[SIZE]; + + sign = -1; + for (i = 0; i < SIZE; i++) + { + src1.a[i] = 1.5 + 34.67 * i * sign; + src2.a[i] = -22.17 * i * sign + 1.0; + sign = sign * -1; + } + for (i = 0; i < SIZE; i++) + res2.a[i] = DEFAULT_VALUE; + + res1.x = _mm_min_round_sd (src1.x, src2.x, _MM_FROUND_NO_EXC); + res2.x = _mm_mask_min_round_sd (res2.x, mask, src1.x, src2.x, _MM_FROUND_NO_EXC); + res3.x = _mm_maskz_min_round_sd (mask, src1.x, src2.x, _MM_FROUND_NO_EXC); + + calc_min (res_ref, src1.a, src2.a); + + if (check_union128d (res1, res_ref)) + abort(); + + MASK_MERGE (d) (res_ref, mask, 1); + if (check_union128d (res2, res_ref)) + abort (); + + MASK_ZERO (d) (res_ref, mask, 1); + if (check_union128d (res3, res_ref)) + abort (); +} + diff --git a/gcc/testsuite/gcc.target/i386/avx512f-vminss-1.c b/gcc/testsuite/gcc.target/i386/avx512f-vminss-1.c index 62a4d38..5703a88 100644 --- a/gcc/testsuite/gcc.target/i386/avx512f-vminss-1.c +++ b/gcc/testsuite/gcc.target/i386/avx512f-vminss-1.c @@ -1,13 +1,18 @@ /* { dg-do compile } */ /* { dg-options "-mavx512f -O2" } */ /* { dg-final { scan-assembler-times "vminss\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vminss\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */ +/* { dg-final { scan-assembler-times "vminss\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */ #include -volatile __m128 x1, x2; +volatile __m128 x1, x2, x3; +volatile __mmask8 m; void extern avx512f_test (void) { x1 = _mm_min_round_ss (x1, x2, _MM_FROUND_NO_EXC); + x1 = _mm_mask_min_round_ss (x1, m, x2, x3, _MM_FROUND_NO_EXC); + x1 = _mm_maskz_min_round_ss (m, x1, x2, _MM_FROUND_NO_EXC); } diff --git a/gcc/testsuite/gcc.target/i386/avx512f-vminss-2.c b/gcc/testsuite/gcc.target/i386/avx512f-vminss-2.c new file mode 100644 index 0000000..ad1fcb9 --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/avx512f-vminss-2.c @@ -0,0 +1,58 @@ +/* { dg-do run } */ +/* { dg-options "-O2 -mavx512f" } */ +/* { dg-require-effective-target avx512f } */ + +#include "avx512f-check.h" + +#include "avx512f-helper.h" + +#define SIZE (128 / 32) +#include "avx512f-mask-type.h" + +static void +calc_min (float *r, float *s1, float *s2) +{ + r[0] = s1[0] < s2[0] ? s1[0] : s2[0]; + int i; + for (i = 1; i < SIZE; i++) + { + r[i] = s1[i]; + } +} + +void +avx512f_test (void) +{ + int i, sign; + union128 res1, res2, res3, src1, src2; + MASK_TYPE mask = MASK_VALUE; + float res_ref[SIZE]; + + sign = -1; + for (i = 0; i < SIZE; i++) + { + src1.a[i] = 1.5 + 34.67 * i * sign; + src2.a[i] = -22.17 * i * sign + 1.0; + sign = sign * -1; + } + for (i = 0; i < SIZE; i++) + res2.a[i] = DEFAULT_VALUE; + + res1.x = _mm_min_round_ss (src1.x, src2.x, _MM_FROUND_NO_EXC); + res2.x = _mm_mask_min_round_ss (res2.x, mask, src1.x, src2.x, _MM_FROUND_NO_EXC); + res3.x = _mm_maskz_min_round_ss (mask, src1.x, src2.x, _MM_FROUND_NO_EXC); + + calc_min (res_ref, src1.a, src2.a); + + if (check_union128 (res1, res_ref)) + abort(); + + MASK_MERGE () (res_ref, mask, 1); + if (check_union128 (res2, res_ref)) + abort (); + + MASK_ZERO () (res_ref, mask, 1); + if (check_union128 (res3, res_ref)) + abort (); +} + diff --git a/gcc/testsuite/gcc.target/i386/sse-13.c b/gcc/testsuite/gcc.target/i386/sse-13.c index 0541129..79879d7 100644 --- a/gcc/testsuite/gcc.target/i386/sse-13.c +++ b/gcc/testsuite/gcc.target/i386/sse-13.c @@ -262,11 +262,15 @@ #define __builtin_ia32_maxpd512_mask(A, B, C, D, E) __builtin_ia32_maxpd512_mask(A, B, C, D, 8) #define __builtin_ia32_maxps512_mask(A, B, C, D, E) __builtin_ia32_maxps512_mask(A, B, C, D, 8) #define __builtin_ia32_maxsd_round(A, B, C) __builtin_ia32_maxsd_round(A, B, 4) +#define __builtin_ia32_maxsd_mask_round(A, B, C, D, E) __builtin_ia32_maxsd_mask_round(A, B, C, D, 4) #define __builtin_ia32_maxss_round(A, B, C) __builtin_ia32_maxss_round(A, B, 4) +#define __builtin_ia32_maxss_mask_round(A, B, C, D, E) __builtin_ia32_maxss_mask_round(A, B, C, D, 4) #define __builtin_ia32_minpd512_mask(A, B, C, D, E) __builtin_ia32_minpd512_mask(A, B, C, D, 8) #define __builtin_ia32_minps512_mask(A, B, C, D, E) __builtin_ia32_minps512_mask(A, B, C, D, 8) #define __builtin_ia32_minsd_round(A, B, C) __builtin_ia32_minsd_round(A, B, 4) +#define __builtin_ia32_minsd_mask_round(A, B, C, D, E) __builtin_ia32_minsd_mask_round(A, B, C, D, 4) #define __builtin_ia32_minss_round(A, B, C) __builtin_ia32_minss_round(A, B, 4) +#define __builtin_ia32_minss_mask_round(A, B, C, D, E) __builtin_ia32_minss_mask_round(A, B, C, D, 4) #define __builtin_ia32_mulpd512_mask(A, B, C, D, E) __builtin_ia32_mulpd512_mask(A, B, C, D, 8) #define __builtin_ia32_mulps512_mask(A, B, C, D, E) __builtin_ia32_mulps512_mask(A, B, C, D, 8) #define __builtin_ia32_mulsd_round(A, B, C) __builtin_ia32_mulsd_round(A, B, 8) diff --git a/gcc/testsuite/gcc.target/i386/sse-14.c b/gcc/testsuite/gcc.target/i386/sse-14.c index ce341a2..547314a 100644 --- a/gcc/testsuite/gcc.target/i386/sse-14.c +++ b/gcc/testsuite/gcc.target/i386/sse-14.c @@ -391,9 +391,13 @@ test_3 (_mm512_maskz_insertf64x4, __m512d, __mmask8, __m512d, __m256d, 1) test_3 (_mm512_maskz_inserti32x4, __m512i, __mmask16, __m512i, __m128i, 1) test_3 (_mm512_maskz_inserti64x4, __m512i, __mmask8, __m512i, __m256i, 1) test_3 (_mm512_maskz_max_round_pd, __m512d, __mmask8, __m512d, __m512d, 8) +test_3 (_mm_maskz_max_round_sd, __m128d, __mmask8, __m128d, __m128d, 8) test_3 (_mm512_maskz_max_round_ps, __m512, __mmask16, __m512, __m512, 8) +test_3 (_mm_maskz_max_round_ss, __m128, __mmask8, __m128, __m128, 8) test_3 (_mm512_maskz_min_round_pd, __m512d, __mmask8, __m512d, __m512d, 8) +test_3 (_mm_maskz_min_round_sd, __m128d, __mmask8, __m128d, __m128d, 8) test_3 (_mm512_maskz_min_round_ps, __m512, __mmask16, __m512, __m512, 8) +test_3 (_mm_maskz_min_round_ss, __m128, __mmask8, __m128, __m128, 8) test_3 (_mm512_maskz_mul_round_pd, __m512d, __mmask8, __m512d, __m512d, 9) test_3 (_mm_maskz_mul_round_sd, __m128d, __mmask8, __m128d, __m128d, 9) test_3 (_mm512_maskz_mul_round_ps, __m512, __mmask16, __m512, __m512, 9) @@ -483,9 +487,13 @@ test_4 (_mm512_mask_insertf64x4, __m512d, __m512d, __mmask8, __m512d, __m256d, 1 test_4 (_mm512_mask_inserti32x4, __m512i, __m512i, __mmask16, __m512i, __m128i, 1) test_4 (_mm512_mask_inserti64x4, __m512i, __m512i, __mmask8, __m512i, __m256i, 1) test_4 (_mm512_mask_max_round_pd, __m512d, __m512d, __mmask8, __m512d, __m512d, 8) +test_4 (_mm_mask_max_round_sd, __m128d, __m128d, __mmask8, __m128d, __m128d, 8) test_4 (_mm512_mask_max_round_ps, __m512, __m512, __mmask16, __m512, __m512, 8) +test_4 (_mm_mask_max_round_ss, __m128, __m128, __mmask8, __m128, __m128, 8) test_4 (_mm512_mask_min_round_pd, __m512d, __m512d, __mmask8, __m512d, __m512d, 8) +test_4 (_mm_mask_min_round_sd, __m128d, __m128d, __mmask8, __m128d, __m128d, 8) test_4 (_mm512_mask_min_round_ps, __m512, __m512, __mmask16, __m512, __m512, 8) +test_4 (_mm_mask_min_round_ss, __m128, __m128, __mmask8, __m128, __m128, 8) test_4 (_mm512_mask_mul_round_pd, __m512d, __m512d, __mmask8, __m512d, __m512d, 9) test_4 (_mm_mask_mul_round_sd, __m128d, __m128d, __mmask8, __m128d, __m128d, 9) test_4 (_mm512_mask_mul_round_ps, __m512, __m512, __mmask16, __m512, __m512, 9) diff --git a/gcc/testsuite/gcc.target/i386/sse-23.c b/gcc/testsuite/gcc.target/i386/sse-23.c index e2b4f89..96c663b 100644 --- a/gcc/testsuite/gcc.target/i386/sse-23.c +++ b/gcc/testsuite/gcc.target/i386/sse-23.c @@ -263,11 +263,15 @@ #define __builtin_ia32_maxpd512_mask(A, B, C, D, E) __builtin_ia32_maxpd512_mask(A, B, C, D, 8) #define __builtin_ia32_maxps512_mask(A, B, C, D, E) __builtin_ia32_maxps512_mask(A, B, C, D, 8) #define __builtin_ia32_maxsd_round(A, B, C) __builtin_ia32_maxsd_round(A, B, 4) +#define __builtin_ia32_maxsd_mask_round(A, B, C, D, E) __builtin_ia32_maxsd_mask_round(A, B, C, D, 4) #define __builtin_ia32_maxss_round(A, B, C) __builtin_ia32_maxss_round(A, B, 4) +#define __builtin_ia32_maxss_mask_round(A, B, C, D, E) __builtin_ia32_maxss_mask_round(A, B, C, D, 4) #define __builtin_ia32_minpd512_mask(A, B, C, D, E) __builtin_ia32_minpd512_mask(A, B, C, D, 8) #define __builtin_ia32_minps512_mask(A, B, C, D, E) __builtin_ia32_minps512_mask(A, B, C, D, 8) #define __builtin_ia32_minsd_round(A, B, C) __builtin_ia32_minsd_round(A, B, 4) +#define __builtin_ia32_minsd_mask_round(A, B, C, D, E) __builtin_ia32_minsd_mask_round(A, B, C, D, 4) #define __builtin_ia32_minss_round(A, B, C) __builtin_ia32_minss_round(A, B, 4) +#define __builtin_ia32_minss_mask_round(A, B, C, D, E) __builtin_ia32_minss_mask_round(A, B, C, D, 4) #define __builtin_ia32_mulpd512_mask(A, B, C, D, E) __builtin_ia32_mulpd512_mask(A, B, C, D, 8) #define __builtin_ia32_mulps512_mask(A, B, C, D, E) __builtin_ia32_mulps512_mask(A, B, C, D, 8) #define __builtin_ia32_mulsd_round(A, B, C) __builtin_ia32_mulsd_round(A, B, 8) diff --git a/gcc/testsuite/gcc.target/i386/testround-1.c b/gcc/testsuite/gcc.target/i386/testround-1.c index 829ce78..2c13381 100644 --- a/gcc/testsuite/gcc.target/i386/testround-1.c +++ b/gcc/testsuite/gcc.target/i386/testround-1.c @@ -187,12 +187,24 @@ test_round (void) m512 = _mm512_max_round_ps (m512, m512, 7); /* { dg-error "incorrect rounding operand" } */ m512 = _mm512_mask_max_round_ps (m512, mmask16, m512, m512, 7); /* { dg-error "incorrect rounding operand" } */ m512 = _mm512_maskz_max_round_ps (mmask16, m512, m512, 7); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_max_round_sd (m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_mask_max_round_sd (m128d, mmask8, m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_maskz_max_round_sd (mmask8, m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_max_round_ss (m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_mask_max_round_ss (m128, mmask8, m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_maskz_max_round_ss (mmask8, m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ m512d = _mm512_min_round_pd (m512d, m512d, 7); /* { dg-error "incorrect rounding operand" } */ m512d = _mm512_mask_min_round_pd (m512d, mmask8, m512d, m512d, 7); /* { dg-error "incorrect rounding operand" } */ m512d = _mm512_maskz_min_round_pd (mmask8, m512d, m512d, 7); /* { dg-error "incorrect rounding operand" } */ m512 = _mm512_min_round_ps (m512, m512, 7); /* { dg-error "incorrect rounding operand" } */ m512 = _mm512_mask_min_round_ps (m512, mmask16, m512, m512, 7); /* { dg-error "incorrect rounding operand" } */ m512 = _mm512_maskz_min_round_ps (mmask16, m512, m512, 7); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_min_round_sd (m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_mask_min_round_sd (m128d, mmask8, m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_maskz_min_round_sd (mmask8, m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_min_round_ss (m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_mask_min_round_ss (m128, mmask8, m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_maskz_min_round_ss (mmask8, m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ m256i = _mm512_cvtt_roundpd_epi32 (m512d, 7); /* { dg-error "incorrect rounding operand" } */ m256i = _mm512_mask_cvtt_roundpd_epi32 (m256i, mmask8, m512d, 7); /* { dg-error "incorrect rounding operand" } */ @@ -450,12 +462,24 @@ test_sae_only (void) m512 = _mm512_max_round_ps (m512, m512, 3); /* { dg-error "incorrect rounding operand" } */ m512 = _mm512_mask_max_round_ps (m512, mmask16, m512, m512, 3); /* { dg-error "incorrect rounding operand" } */ m512 = _mm512_maskz_max_round_ps (mmask16, m512, m512, 3); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_max_round_sd (m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_mask_max_round_sd (m128d, mmask8, m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_maskz_max_round_sd (mmask8, m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_max_round_ss (m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_mask_max_round_ss (m128, mmask8, m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_maskz_max_round_ss (mmask8, m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ m512d = _mm512_min_round_pd (m512d, m512d, 3); /* { dg-error "incorrect rounding operand" } */ m512d = _mm512_mask_min_round_pd (m512d, mmask8, m512d, m512d, 3); /* { dg-error "incorrect rounding operand" } */ m512d = _mm512_maskz_min_round_pd (mmask8, m512d, m512d, 3); /* { dg-error "incorrect rounding operand" } */ m512 = _mm512_min_round_ps (m512, m512, 3); /* { dg-error "incorrect rounding operand" } */ m512 = _mm512_mask_min_round_ps (m512, mmask16, m512, m512, 3); /* { dg-error "incorrect rounding operand" } */ m512 = _mm512_maskz_min_round_ps (mmask16, m512, m512, 3); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_min_round_sd (m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_mask_min_round_sd (m128d, mmask8, m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128d = _mm_maskz_min_round_sd (mmask8, m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_min_round_ss (m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_mask_min_round_ss (m128, mmask8, m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ + m128 = _mm_maskz_min_round_ss (mmask8, m128, m128, 7); /* { dg-error "incorrect rounding operand" } */ m256i = _mm512_cvtt_roundpd_epi32 (m512d, 3); /* { dg-error "incorrect rounding operand" } */ m256i = _mm512_mask_cvtt_roundpd_epi32 (m256i, mmask8, m512d, 3); /* { dg-error "incorrect rounding operand" } */