/* Prototype for the BLAS ?gemm subroutine, a pointer to which can be
passed to us by the front-end, in which case we call it for large
matrices. */
typedef void (*blas_call)(const char *, const char *, const int *, const int *,
const int *, const GFC_INTEGER_8 *, const GFC_INTEGER_8 *,
const int *, const GFC_INTEGER_8 *, const int *,
const GFC_INTEGER_8 *, GFC_INTEGER_8 *, const int *,
int, int);
#if defined(HAVE_AVX) && defined(HAVE_FMA3) && defined(HAVE_AVX128)
void
matmul_i8_avx128_fma3 (gfc_array_i8 * const restrict retarray,
gfc_array_i8 * const restrict a, gfc_array_i8 * const restrict b, int try_blas,
int blas_limit, blas_call gemm) __attribute__((__target__("avx,fma")));
internal_proto(matmul_i8_avx128_fma3);
void
matmul_i8_avx128_fma3 (gfc_array_i8 * const restrict retarray,
gfc_array_i8 * const restrict a, gfc_array_i8 * const restrict b, int try_blas,
int blas_limit, blas_call gemm)
{
const GFC_INTEGER_8 * restrict abase;
const GFC_INTEGER_8 * restrict bbase;
GFC_INTEGER_8 * restrict dest;
index_type rxstride, rystride, axstride, aystride, bxstride, bystride;
index_type x, y, n, count, xcount, ycount;
assert (GFC_DESCRIPTOR_RANK (a) == 2
|| GFC_DESCRIPTOR_RANK (b) == 2);
/* PR target/36936 */
/* { dg-do compile } */
/* { dg-require-effective-target ilp32 } */
/* { dg-options "-O2 -march=i386" } */
/* { dg-final { scan-assembler "cmov" } } */
extern int foo (int) __attribute__((__target__("arch=i686")));
int
foo (int x)
{
if (x < 0)
x = 1;
return x;
}
/* PR target/69969 */
/* { dg-do compile } */
/* { dg-skip-if "do not override -mcpu" { powerpc*-*-* } { "-mcpu=*" } { "-mcpu=power8" } } */
/* { dg-options "-mcpu=power8" } */
int bar (int x) { return x; }
__attribute__((__target__("no-vsx"))) int foo (int x) { return x; } /* { dg-bogus "-mallow-movmisalign requires -mvsx" } */
/* { dg-do run } */
/* { dg-options "-O2 -ffast-math -ftree-vectorize -mavx512f" } */
/* { dg-require-effective-target avx512f } */
/* { dg-skip-if "no M_PI" { vxworks_kernel } } */
#define __NO_MATH_INLINES
#include <math.h>
#include "avx512f-check.h"
#define NUM 64
static void
__attribute__((__target__("fpmath=sse")))
init_src (float *src)
{
int i, sign = 1;
float f = rand ();
for (i = 0; i < NUM; i++)
{
src[i] = (i + 1) * f * M_PI * sign;
if (i < (NUM / 2))
{
if ((i % 6) == 0)
f = f * src[i];
}
else if (i == (NUM / 2))
f = rand ();
else if ((i % 6) == 0)
f = 1 / (f * (i + 1) * src[i] * M_PI * sign);
sign = -sign;
/* Test whether all of the 32-bit function specific options are accepted
without error. */
/* { dg-do compile } */
/* { dg-require-effective-target ilp32 } */
extern void test_abm (void) __attribute__((__target__("abm")));
extern void test_aes (void) __attribute__((__target__("aes")));
extern void test_bmi (void) __attribute__((__target__("bmi")));
extern void test_mmx (void) __attribute__((__target__("mmx")));
extern void test_pclmul (void) __attribute__((__target__("pclmul")));
extern void test_popcnt (void) __attribute__((__target__("popcnt")));
extern void test_recip (void) __attribute__((__target__("recip")));
extern void test_sse (void) __attribute__((__target__("sse")));
extern void test_sse2 (void) __attribute__((__target__("sse2")));
extern void test_sse3 (void) __attribute__((__target__("sse3")));
extern void test_sse4 (void) __attribute__((__target__("sse4")));
extern void test_sse4_1 (void) __attribute__((__target__("sse4.1")));
extern void test_sse4_2 (void) __attribute__((__target__("sse4.2")));
extern void test_sse4a (void) __attribute__((__target__("sse4a")));
extern void test_fma4 (void) __attribute__((__target__("fma4")));
extern void test_ssse3 (void) __attribute__((__target__("ssse3")));
extern void test_tbm (void) __attribute__((__target__("tbm")));
extern void test_no_abm (void) __attribute__((__target__("no-abm")));
extern void test_no_aes (void) __attribute__((__target__("no-aes")));
extern void test_no_bmi (void) __attribute__((__target__("no-bmi")));
extern void test_no_mmx (void) __attribute__((__target__("no-mmx")));
extern void test_no_pclmul (void) __attribute__((__target__("no-pclmul")));
extern void test_no_popcnt (void) __attribute__((__target__("no-popcnt")));
extern void test_no_recip (void) __attribute__((__target__("no-recip")));
extern void test_no_sse (void) __attribute__((__target__("no-sse")));
for (size_t index = 0; index < haystack.size(); ++index) {
if (s.contains(haystack[index])) {
return index;
}
}
return StringPiece::npos;
}
#if FOLLY_HAVE_EMMINTRIN_H && __GNUC_PREREQ(4, 6)
template <bool HAYSTACK_ALIGNED>
inline size_t scanHaystackBlock(const StringPiece& haystack,
const StringPiece& needles,
int64_t idx)
// inline is okay because it's only called from other sse4.2 functions
__attribute__ ((__target__("sse4.2")));
// Scans a 16-byte block of haystack (starting at blockStartIdx) to find first
// needle. If HAYSTACK_ALIGNED, then haystack must be 16byte aligned.
// If !HAYSTACK_ALIGNED, then caller must ensure that it is safe to load the
// block.
template <bool HAYSTACK_ALIGNED>
inline size_t scanHaystackBlock(const StringPiece& haystack,
const StringPiece& needles,
int64_t blockStartIdx) {
DCHECK_GT(needles.size(), 16); // should handled by *needles16() method
DCHECK(blockStartIdx + 16 <= haystack.size() ||
(PAGE_FOR(haystack.data() + blockStartIdx) ==
PAGE_FOR(haystack.data() + blockStartIdx + 15)));
__v16qi arr1;
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with Kvazaar. If not, see <http://www.gnu.org/licenses/>.
****************************************************************************/
/*
* \file
*/
#include "../strategyselector.h"
#include "../picture.h"
#include <immintrin.h>
#include <assert.h>
#ifdef __GNUC__
__attribute__ ((__target__ ("sse2,sse4.1")))
#endif
static unsigned reg_sad_sse41(const pixel * const data1, const pixel * const data2,
const int width, const int height, const unsigned stride1, const unsigned stride2)
{
int y, x;
unsigned sad = 0;
__m128i sse_inc = _mm_setzero_si128 ();
long long int sse_inc_array[2];
for (y = 0; y < height; ++y) {
for (x = 0; x <= width-16; x+=16) {
const __m128i a = _mm_loadu_si128((__m128i const*) &data1[y * stride1 + x]);
const __m128i b = _mm_loadu_si128((__m128i const*) &data2[y * stride2 + x]);
sse_inc = _mm_add_epi32(sse_inc, _mm_sad_epu8(a,b));
}
/* Test whether all of the 64-bit function specific options are accepted
without error. */
/* { dg-do compile } */
/* { dg-require-effective-target lp64 } */
extern void test_abm (void) __attribute__((__target__("abm")));
extern void test_aes (void) __attribute__((__target__("aes")));
extern void test_mmx (void) __attribute__((__target__("mmx")));
extern void test_pclmul (void) __attribute__((__target__("pclmul")));
extern void test_popcnt (void) __attribute__((__target__("popcnt")));
extern void test_recip (void) __attribute__((__target__("recip")));
extern void test_sse (void) __attribute__((__target__("sse")));
extern void test_sse2 (void) __attribute__((__target__("sse2")));
extern void test_sse3 (void) __attribute__((__target__("sse3")));
extern void test_sse4 (void) __attribute__((__target__("sse4")));
extern void test_sse4_1 (void) __attribute__((__target__("sse4.1")));
extern void test_sse4_2 (void) __attribute__((__target__("sse4.2")));
extern void test_sse4a (void) __attribute__((__target__("sse4a")));
extern void test_fma4 (void) __attribute__((__target__("fma4")));
extern void test_ssse3 (void) __attribute__((__target__("ssse3")));
extern void test_no_abm (void) __attribute__((__target__("no-abm")));
extern void test_no_aes (void) __attribute__((__target__("no-aes")));
extern void test_no_mmx (void) __attribute__((__target__("no-mmx")));
extern void test_no_pclmul (void) __attribute__((__target__("no-pclmul")));
extern void test_no_popcnt (void) __attribute__((__target__("no-popcnt")));
extern void test_no_recip (void) __attribute__((__target__("no-recip")));
extern void test_no_sse (void) __attribute__((__target__("no-sse")));
extern void test_no_sse2 (void) __attribute__((__target__("no-sse2")));
extern void test_no_sse3 (void) __attribute__((__target__("no-sse3")));
extern void test_no_sse4 (void) __attribute__((__target__("no-sse4")));
// RUN: %clang_cc1 -fsyntax-only -ffreestanding %s -verify
// RUN: %clang_cc1 -fsyntax-only -ffreestanding -fno-lax-vector-conversions %s -verify
// RUN: %clang_cc1 -fsyntax-only -ffreestanding -x c++ %s -verify
// expected-no-diagnostics
#if defined(i386) || defined(__x86_64__)
// Include the metaheader that includes all x86 intrinsic headers.
#include <x86intrin.h>
void __attribute__((__target__("mmx"))) mm_empty_wrap(void) {
_mm_empty();
}
__m128 __attribute__((__target__("sse"))) mm_add_ss_wrap(__m128 a, __m128 b) {
return _mm_add_ss(a, b);
}
__m128d __attribute__((__target__("sse2"))) mm_sqrt_sd_wrap(__m128d a, __m128d b) {
return _mm_sqrt_sd(a, b);
}
void __attribute__((__target__("sse3"))) mm_mwait_wrap(int a) {
_mm_mwait(0, 0);
}
__m64 __attribute__((__target__("ssse3"))) mm_abs_pi8_wrap(__m64 a) {
return _mm_abs_pi8(a);
}
__m128i __attribute__((__target__("sse4.1"))) mm_minpos_epu16_wrap(__m128i v) {
/* Test whether all of the 64-bit function specific options are accepted
without error. */
/* { dg-do compile { target { ! ia32 } } } */
#include "funcspec-56.inc"
extern void test_arch_foo (void) __attribute__((__target__("arch=foo"))); /* { dg-error "bad value" } */
extern void test_tune_foo (void) __attribute__((__target__("tune=foo"))); /* { dg-error "bad value" } */
/* { dg-do compile { target { powerpc*-*-* && ilp32 } } } */
/* { dg-skip-if "" { powerpc*-*-darwin* } } */
/* { dg-require-effective-target powerpc_vsx_ok } */
/* { dg-skip-if "do not override -mcpu" { powerpc*-*-* } { "-mcpu=*" } { "-mcpu=power5" } } */
/* { dg-options "-O2 -ffast-math -mcpu=power5 -mabi=no-altivec" } */
/* { dg-final { scan-assembler-times "fabs" 3 } } */
/* { dg-final { scan-assembler-times "fnabs" 3 } } */
/* { dg-final { scan-assembler-times "fsel" 3 } } */
/* { dg-final { scan-assembler-times "fcpsgn" 4 } } */
/* { dg-final { scan-assembler-not "xscpsgndp" } } */
/* Like ppc-target-1.c, but do not enable the altivec abi on 32-bit, so the
power7 code should generate fcpsgn and not xscpsgndp. */
double normal1 (double, double);
double power5 (double, double) __attribute__((__target__("cpu=power5")));
double power6 (double, double) __attribute__((__target__("cpu=power6")));
double power6x (double, double) __attribute__((__target__("cpu=power6x")));
double power7 (double, double) __attribute__((__target__("cpu=power7")));
double power7n (double, double) __attribute__((__target__("cpu=power7,no-vsx")));
double normal2 (double, double);
/* fabs/fnabs/fsel */
double normal1 (double a, double b)
{
return __builtin_copysign (a, b);
}
/* fabs/fnabs/fsel */
double power5 (double a, double b)
{
// RUN: %clang_cc1 -fsyntax-only -ffreestanding %s -verify
// RUN: %clang_cc1 -fsyntax-only -ffreestanding -x c++ %s -verify
// expected-no-diagnostics
#if defined(i386) || defined(__x86_64__)
#include <pmmintrin.h>
int __attribute__((__target__(("sse3")))) foo(int a) {
_mm_mwait(0, 0);
return 4;
}
#endif
