c34e2cf380
for skylake kernels. This is the same method as used in [sd]asum. _mm_set1_epi64x was commented out for zasum, but has the advantage of avoiding possible undefined behaviour (using an uninitialized variable), optimized out by NVHPC and icx. The new code works fine with those compilers. For GCC 12.3 the generated code is identical; no matter what method you use, the compiler optimizes the code into a compile-time constant, there is no performance benefit using mm_cmpeq_epi8 since the corresponding instruction (VPCMPEQB) isn't actually generated!
345 lines
12 KiB
C
345 lines
12 KiB
C
/* need a new enough GCC for avx512 support */
|
|
#ifdef __NVCOMPILER
|
|
#define NVCOMPVERS ( __NVCOMPILER_MAJOR__ * 100 + __NVCOMPILER_MINOR__ )
|
|
#endif
|
|
#if ((( defined(__GNUC__) && __GNUC__ > 6 && defined(__AVX512CD__)) || (defined(__clang__) && __clang_major__ >= 9)) || (defined(__NVCOMPILER) && NVCOMPVERS >= 2203))
|
|
|
|
#if (!(defined(__NVCOMPILER) && NVCOMPVERS < 2203))
|
|
|
|
#define HAVE_ZASUM_KERNEL 1
|
|
|
|
#include <immintrin.h>
|
|
|
|
#include <stdint.h>
|
|
|
|
static FLOAT zasum_kernel(BLASLONG n, FLOAT *x)
|
|
{
|
|
FLOAT *x1 = x;
|
|
FLOAT sumf=0.0;
|
|
BLASLONG n2 = n + n;
|
|
|
|
|
|
if (n2 < 32) {
|
|
__m128d accum_10, accum_11, accum_12, accum_13;
|
|
__m128d abs_mask1;
|
|
|
|
accum_10 = _mm_setzero_pd();
|
|
accum_11 = _mm_setzero_pd();
|
|
accum_12 = _mm_setzero_pd();
|
|
accum_13 = _mm_setzero_pd();
|
|
|
|
abs_mask1 = (__m128d)_mm_set1_epi64x(0x7fffffffffffffff);
|
|
|
|
_mm_prefetch(&x1[0], _MM_HINT_T0);
|
|
if (n2 >= 16){
|
|
__m128d x00 = _mm_loadu_pd(&x1[ 0]);
|
|
__m128d x01 = _mm_loadu_pd(&x1[ 2]);
|
|
__m128d x02 = _mm_loadu_pd(&x1[ 4]);
|
|
__m128d x03 = _mm_loadu_pd(&x1[ 6]);
|
|
|
|
_mm_prefetch(&x1[8], _MM_HINT_T0);
|
|
__m128d x04 = _mm_loadu_pd(&x1[ 8]);
|
|
__m128d x05 = _mm_loadu_pd(&x1[10]);
|
|
__m128d x06 = _mm_loadu_pd(&x1[12]);
|
|
__m128d x07 = _mm_loadu_pd(&x1[14]);
|
|
|
|
x00 = _mm_and_pd(x00, abs_mask1);
|
|
x01 = _mm_and_pd(x01, abs_mask1);
|
|
x02 = _mm_and_pd(x02, abs_mask1);
|
|
x03 = _mm_and_pd(x03, abs_mask1);
|
|
|
|
accum_10 = _mm_add_pd(accum_10, x00);
|
|
accum_11 = _mm_add_pd(accum_11, x01);
|
|
accum_12 = _mm_add_pd(accum_12, x02);
|
|
accum_13 = _mm_add_pd(accum_13, x03);
|
|
|
|
x04 = _mm_and_pd(x04, abs_mask1);
|
|
x05 = _mm_and_pd(x05, abs_mask1);
|
|
x06 = _mm_and_pd(x06, abs_mask1);
|
|
x07 = _mm_and_pd(x07, abs_mask1);
|
|
|
|
accum_10 = _mm_add_pd(accum_10, x04);
|
|
accum_11 = _mm_add_pd(accum_11, x05);
|
|
accum_12 = _mm_add_pd(accum_12, x06);
|
|
accum_13 = _mm_add_pd(accum_13, x07);
|
|
|
|
x1 += 16;
|
|
n2 -= 16;
|
|
}
|
|
|
|
if (n2 >= 8) {
|
|
__m128d x00 = _mm_loadu_pd(&x1[ 0]);
|
|
__m128d x01 = _mm_loadu_pd(&x1[ 2]);
|
|
__m128d x02 = _mm_loadu_pd(&x1[ 4]);
|
|
__m128d x03 = _mm_loadu_pd(&x1[ 6]);
|
|
|
|
x00 = _mm_and_pd(x00, abs_mask1);
|
|
x01 = _mm_and_pd(x01, abs_mask1);
|
|
x02 = _mm_and_pd(x02, abs_mask1);
|
|
x03 = _mm_and_pd(x03, abs_mask1);
|
|
accum_10 = _mm_add_pd(accum_10, x00);
|
|
accum_11 = _mm_add_pd(accum_11, x01);
|
|
accum_12 = _mm_add_pd(accum_12, x02);
|
|
accum_13 = _mm_add_pd(accum_13, x03);
|
|
|
|
n2 -= 8;
|
|
x1 += 8;
|
|
}
|
|
|
|
if (n2 >= 4) {
|
|
__m128d x00 = _mm_loadu_pd(&x1[ 0]);
|
|
__m128d x01 = _mm_loadu_pd(&x1[ 2]);
|
|
x00 = _mm_and_pd(x00, abs_mask1);
|
|
x01 = _mm_and_pd(x01, abs_mask1);
|
|
accum_10 = _mm_add_pd(accum_10, x00);
|
|
accum_11 = _mm_add_pd(accum_11, x01);
|
|
|
|
n2 -= 4;
|
|
x1 += 4;
|
|
}
|
|
|
|
if (n2) {
|
|
__m128d x00 = _mm_loadu_pd(&x1[ 0]);
|
|
x00 = _mm_and_pd(x00, abs_mask1);
|
|
accum_10 = _mm_add_pd(accum_10, x00);
|
|
}
|
|
|
|
accum_10 = _mm_add_pd(accum_10, accum_11);
|
|
accum_12 = _mm_add_pd(accum_12, accum_13);
|
|
accum_10 = _mm_add_pd(accum_10, accum_12);
|
|
|
|
accum_10 = _mm_hadd_pd(accum_10, accum_10);
|
|
|
|
sumf = accum_10[0];
|
|
}
|
|
else {
|
|
__m512d accum_0, accum_1, accum_2, accum_3;
|
|
__m512d x00, x01, x02, x03, x04, x05, x06, x07;
|
|
__m512d abs_mask = (__m512d)_mm512_set1_epi64(0x7fffffffffffffff);
|
|
|
|
accum_0 = _mm512_setzero_pd();
|
|
accum_1 = _mm512_setzero_pd();
|
|
accum_2 = _mm512_setzero_pd();
|
|
accum_3 = _mm512_setzero_pd();
|
|
|
|
// alignment has side-effect when the size of input array is not large enough
|
|
if (n2 < 128) {
|
|
if (n2 >= 64) {
|
|
x00 = _mm512_loadu_pd(&x1[ 0]);
|
|
x01 = _mm512_loadu_pd(&x1[ 8]);
|
|
x02 = _mm512_loadu_pd(&x1[16]);
|
|
x03 = _mm512_loadu_pd(&x1[24]);
|
|
x04 = _mm512_loadu_pd(&x1[32]);
|
|
x05 = _mm512_loadu_pd(&x1[40]);
|
|
x06 = _mm512_loadu_pd(&x1[48]);
|
|
x07 = _mm512_loadu_pd(&x1[56]);
|
|
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
x01 = _mm512_and_pd(x01, abs_mask);
|
|
x02 = _mm512_and_pd(x02, abs_mask);
|
|
x03 = _mm512_and_pd(x03, abs_mask);
|
|
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
accum_1 = _mm512_add_pd(accum_1, x01);
|
|
accum_2 = _mm512_add_pd(accum_2, x02);
|
|
accum_3 = _mm512_add_pd(accum_3, x03);
|
|
|
|
x04 = _mm512_and_pd(x04, abs_mask);
|
|
x05 = _mm512_and_pd(x05, abs_mask);
|
|
x06 = _mm512_and_pd(x06, abs_mask);
|
|
x07 = _mm512_and_pd(x07, abs_mask);
|
|
|
|
accum_0 = _mm512_add_pd(accum_0, x04);
|
|
accum_1 = _mm512_add_pd(accum_1, x05);
|
|
accum_2 = _mm512_add_pd(accum_2, x06);
|
|
accum_3 = _mm512_add_pd(accum_3, x07);
|
|
|
|
n2 -= 64;
|
|
x1 += 64;
|
|
}
|
|
|
|
if (n2 >= 32) {
|
|
x00 = _mm512_loadu_pd(&x1[ 0]);
|
|
x01 = _mm512_loadu_pd(&x1[ 8]);
|
|
x02 = _mm512_loadu_pd(&x1[16]);
|
|
x03 = _mm512_loadu_pd(&x1[24]);
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
x01 = _mm512_and_pd(x01, abs_mask);
|
|
x02 = _mm512_and_pd(x02, abs_mask);
|
|
x03 = _mm512_and_pd(x03, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
accum_1 = _mm512_add_pd(accum_1, x01);
|
|
accum_2 = _mm512_add_pd(accum_2, x02);
|
|
accum_3 = _mm512_add_pd(accum_3, x03);
|
|
|
|
n2 -= 32;
|
|
x1 += 32;
|
|
}
|
|
|
|
if (n2 >= 16) {
|
|
x00 = _mm512_loadu_pd(&x1[ 0]);
|
|
x01 = _mm512_loadu_pd(&x1[ 8]);
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
x01 = _mm512_and_pd(x01, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
accum_1 = _mm512_add_pd(accum_1, x01);
|
|
|
|
n2 -= 16;
|
|
x1 += 16;
|
|
}
|
|
|
|
if (n2 >= 8) {
|
|
x00 = _mm512_loadu_pd(&x1[ 0]);
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
|
|
n2 -= 8;
|
|
x1 += 8;
|
|
}
|
|
|
|
if (n2) {
|
|
unsigned char tail_mask8 = (((unsigned char) 0xff) >> (8 - n2));
|
|
x00 = _mm512_maskz_loadu_pd(*((__mmask8*) &tail_mask8), &x1[ 0]);
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
}
|
|
accum_0 = _mm512_add_pd(accum_0, accum_1);
|
|
accum_2 = _mm512_add_pd(accum_2, accum_3);
|
|
accum_0 = _mm512_add_pd(accum_0, accum_2);
|
|
sumf = _mm512_reduce_add_pd(accum_0);
|
|
}
|
|
// n2 >= 128, doing alignment
|
|
else {
|
|
|
|
int align_header = ((64 - ((uintptr_t)x1 & (uintptr_t)0x3f)) >> 3) & 0x7;
|
|
|
|
if (0 != align_header) {
|
|
unsigned char align_mask8 = (((unsigned char)0xff) >> (8 - align_header));
|
|
x00 = _mm512_maskz_loadu_pd(*((__mmask8*) &align_mask8), &x1[0]);
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
|
|
n2 -= align_header;
|
|
x1 += align_header;
|
|
}
|
|
|
|
x00 = _mm512_load_pd(&x1[ 0]);
|
|
x01 = _mm512_load_pd(&x1[ 8]);
|
|
x02 = _mm512_load_pd(&x1[16]);
|
|
x03 = _mm512_load_pd(&x1[24]);
|
|
x04 = _mm512_load_pd(&x1[32]);
|
|
x05 = _mm512_load_pd(&x1[40]);
|
|
x06 = _mm512_load_pd(&x1[48]);
|
|
x07 = _mm512_load_pd(&x1[56]);
|
|
|
|
n2 -= 64;
|
|
x1 += 64;
|
|
|
|
while (n2 >= 64) {
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
x01 = _mm512_and_pd(x01, abs_mask);
|
|
x02 = _mm512_and_pd(x02, abs_mask);
|
|
x03 = _mm512_and_pd(x03, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
x00 = _mm512_load_pd(&x1[ 0]);
|
|
accum_1 = _mm512_add_pd(accum_1, x01);
|
|
x01 = _mm512_load_pd(&x1[ 8]);
|
|
accum_2 = _mm512_add_pd(accum_2, x02);
|
|
x02 = _mm512_load_pd(&x1[16]);
|
|
accum_3 = _mm512_add_pd(accum_3, x03);
|
|
x03 = _mm512_load_pd(&x1[24]);
|
|
|
|
x04 = _mm512_and_pd(x04, abs_mask);
|
|
x05 = _mm512_and_pd(x05, abs_mask);
|
|
x06 = _mm512_and_pd(x06, abs_mask);
|
|
x07 = _mm512_and_pd(x07, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x04);
|
|
x04 = _mm512_load_pd(&x1[32]);
|
|
accum_1 = _mm512_add_pd(accum_1, x05);
|
|
x05 = _mm512_load_pd(&x1[40]);
|
|
accum_2 = _mm512_add_pd(accum_2, x06);
|
|
x06 = _mm512_load_pd(&x1[48]);
|
|
accum_3 = _mm512_add_pd(accum_3, x07);
|
|
x07 = _mm512_load_pd(&x1[56]);
|
|
|
|
n2 -= 64;
|
|
x1 += 64;
|
|
}
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
x01 = _mm512_and_pd(x01, abs_mask);
|
|
x02 = _mm512_and_pd(x02, abs_mask);
|
|
x03 = _mm512_and_pd(x03, abs_mask);
|
|
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
accum_1 = _mm512_add_pd(accum_1, x01);
|
|
accum_2 = _mm512_add_pd(accum_2, x02);
|
|
accum_3 = _mm512_add_pd(accum_3, x03);
|
|
|
|
x04 = _mm512_and_pd(x04, abs_mask);
|
|
x05 = _mm512_and_pd(x05, abs_mask);
|
|
x06 = _mm512_and_pd(x06, abs_mask);
|
|
x07 = _mm512_and_pd(x07, abs_mask);
|
|
|
|
accum_0 = _mm512_add_pd(accum_0, x04);
|
|
accum_1 = _mm512_add_pd(accum_1, x05);
|
|
accum_2 = _mm512_add_pd(accum_2, x06);
|
|
accum_3 = _mm512_add_pd(accum_3, x07);
|
|
|
|
if (n2 >= 32) {
|
|
x00 = _mm512_load_pd(&x1[ 0]);
|
|
x01 = _mm512_load_pd(&x1[ 8]);
|
|
x02 = _mm512_load_pd(&x1[16]);
|
|
x03 = _mm512_load_pd(&x1[24]);
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
x01 = _mm512_and_pd(x01, abs_mask);
|
|
x02 = _mm512_and_pd(x02, abs_mask);
|
|
x03 = _mm512_and_pd(x03, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
accum_1 = _mm512_add_pd(accum_1, x01);
|
|
accum_2 = _mm512_add_pd(accum_2, x02);
|
|
accum_3 = _mm512_add_pd(accum_3, x03);
|
|
|
|
n2 -= 32;
|
|
x1 += 32;
|
|
}
|
|
|
|
if (n2 >= 16) {
|
|
x00 = _mm512_load_pd(&x1[ 0]);
|
|
x01 = _mm512_load_pd(&x1[ 8]);
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
x01 = _mm512_and_pd(x01, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
accum_1 = _mm512_add_pd(accum_1, x01);
|
|
|
|
n2 -= 16;
|
|
x1 += 16;
|
|
}
|
|
|
|
if (n2 >= 8) {
|
|
x00 = _mm512_load_pd(&x1[ 0]);
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
|
|
n2 -= 8;
|
|
x1 += 8;
|
|
}
|
|
|
|
if (n2) {
|
|
unsigned char tail_mask8 = (((unsigned char) 0xff) >> (8 - n2));
|
|
x00 = _mm512_maskz_load_pd(*((__mmask8*) &tail_mask8), &x1[ 0]);
|
|
x00 = _mm512_and_pd(x00, abs_mask);
|
|
accum_0 = _mm512_add_pd(accum_0, x00);
|
|
}
|
|
|
|
accum_0 = _mm512_add_pd(accum_0, accum_1);
|
|
accum_2 = _mm512_add_pd(accum_2, accum_3);
|
|
accum_0 = _mm512_add_pd(accum_0, accum_2);
|
|
sumf = _mm512_reduce_add_pd(accum_0);
|
|
}
|
|
}
|
|
|
|
return sumf;
|
|
}
|
|
#endif
|
|
#endif
|