71 const dim3 nblcks(((*m)+1024 - 1)/ 1024, 1, 1);
86 const dim3 nblcks(((*m)+1024 - 1)/ 1024, 1, 1);
101 const dim3 nblcks(((*m)+1024 - 1)/ 1024, 1, 1);
115 const dim3 nblcks(((*m)+1024 - 1)/ 1024, 1, 1);
130 const dim3 nblcks(((*mask_size) + 1024 - 1) / 1024, 1, 1);
150 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
163 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
177 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
190 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
204 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
218 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
232 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
248 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
263 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
278 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
293 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
309 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
327 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
343 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
360 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
377 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
394 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
410 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
423 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
437 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
451 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
464 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
478 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
493 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
507 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
521 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
536 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
551 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
563 void *v1,
void *v2,
void *v3,
int *n,
567 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
580 void *v1,
void *v2,
void *v3,
581 void *w1,
void *w2,
void *
w3,
585 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
638 if (
sizeof(
real) ==
sizeof(
float)) {
643 else if (
sizeof(
real) ==
sizeof(
double)) {
671 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
672 const int nb = ((*n) + 1024 - 1)/ 1024;
699 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
700 const int nb = ((*n) + 1024 - 1)/ 1024;
729 const int nt = 1024/
pow2;
732 const int nb = ((*n) + nt - 1)/nt;
758 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
759 const int nb = ((*n) + 1024 - 1)/ 1024;
787 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
788 const int nb = ((*n) + 1024 - 1)/ 1024;
815 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
816 const int nb = ((*n) + 1024 - 1)/ 1024;
843 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
861 const dim3 nblcks(((*n) + 1024 - 1) / 1024, 1, 1);
875 const dim3 nblcks(((*n) + 1024 - 1) / 1024, 1, 1);
889 const dim3 nblcks(((*n) + 1024 - 1) / 1024, 1, 1);
903 const dim3 nblcks(((*n) + 1024 - 1) / 1024, 1, 1);
917 const dim3 nblcks(((*n) + 1024 - 1) / 1024, 1, 1);
931 const dim3 nblcks(((*n) + 1024 - 1) / 1024, 1, 1);
945 const dim3 nblcks(((*n) + 1024 - 1) / 1024, 1, 1);
958 const dim3 nblcks(((*n) + 1024 - 1) / 1024, 1, 1);
973 const dim3 nblcks(((*n)+1024 - 1)/ 1024, 1, 1);
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ w
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ u
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ v
__global__ void const T *__restrict__ x
__global__ void const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ w3
__global__ void dirichlet_apply_scalar_kernel(const int *__restrict__ msk, T *__restrict__ x, const T g, const int m)
void device_mpi_allreduce(void *buf_d, void *buf, int count, int nbytes, int op)
void device_nccl_allreduce(void *sbuf_d, void *rbuf_d, int count, int nbytes, int op, void *stream)
void cuda_pwmax_sca3(void *a, void *b, real *c, int *n, cudaStream_t stream)
void cuda_addcol4(void *a, void *b, void *c, void *d, int *n, cudaStream_t strm)
void cuda_add2s2(void *a, void *b, real *c1, int *n, cudaStream_t strm)
real cuda_glsubnorm2(void *a, void *b, int *n, cudaStream_t stream)
void cuda_vdot3(void *dot, void *u1, void *u2, void *u3, void *v1, void *v2, void *v3, int *n, cudaStream_t strm)
void cuda_add3s2(void *a, void *b, void *c, real *c1, real *c2, int *n, cudaStream_t strm)
void cuda_global_reduce_add(real *bufred, void *bufred_d, int n, const cudaStream_t stream)
real cuda_glsc2(void *a, void *b, int *n, cudaStream_t stream)
void cuda_addcol3(void *a, void *b, void *c, int *n, cudaStream_t strm)
void cuda_pwmin_vec2(void *a, void *b, int *n, cudaStream_t stream)
void cuda_pwmax_vec3(void *a, void *b, void *c, int *n, cudaStream_t stream)
void cuda_cmult2(void *a, void *b, real *c, int *n, cudaStream_t strm)
void cuda_add2s1(void *a, void *b, real *c1, int *n, cudaStream_t strm)
void cuda_col3(void *a, void *b, void *c, int *n, cudaStream_t strm)
void cuda_add2s2_many(void *x, void **p, void *alpha, int *j, int *n, cudaStream_t strm)
void cuda_sub3(void *a, void *b, void *c, int *n, cudaStream_t strm)
void cuda_copy(void *a, void *b, int *n, cudaStream_t strm)
void cuda_cfill_mask(void *a, real *c, int *size, int *mask, int *mask_size, cudaStream_t strm)
void cuda_invcol2(void *a, void *b, int *n, cudaStream_t strm)
void cuda_pwmax_sca2(void *a, real *c, int *n, cudaStream_t stream)
void cuda_masked_copy(void *a, void *b, void *mask, int *n, int *m, cudaStream_t strm)
void cuda_col2(void *a, void *b, int *n, cudaStream_t strm)
void cuda_masked_scatter_copy(void *a, void *b, void *mask, int *n, int *m, cudaStream_t strm)
void cuda_add2(void *a, void *b, int *n, cudaStream_t strm)
void cuda_redbuf_check_alloc(int nb)
void cuda_cmult(void *a, real *c, int *n, cudaStream_t strm)
void cuda_add4(void *a, void *b, void *c, void *d, int *n, cudaStream_t strm)
void cuda_sub2(void *a, void *b, int *n, cudaStream_t strm)
void cuda_invcol3(void *a, void *b, void *c, int *n, cudaStream_t strm)
void cuda_vcross(void *u1, void *u2, void *u3, void *v1, void *v2, void *v3, void *w1, void *w2, void *w3, int *n, cudaStream_t strm)
void cuda_masked_gather_copy(void *a, void *b, void *mask, int *n, int *m, cudaStream_t strm)
real cuda_glsum(void *a, int *n, cudaStream_t stream)
real cuda_glsc3(void *a, void *b, void *c, int *n, cudaStream_t stream)
real cuda_vlsc3(void *u, void *v, void *w, int *n, cudaStream_t stream)
void cuda_pwmin_sca3(void *a, void *b, real *c, int *n, cudaStream_t stream)
void cuda_cadd2(void *a, void *b, real *c, int *n, cudaStream_t strm)
void cuda_addsqr2s2(void *a, void *b, real *c1, int *n, cudaStream_t strm)
void cuda_add3(void *a, void *b, void *c, int *n, cudaStream_t strm)
void cuda_addcol3s2(void *a, void *b, void *c, real *s, int *n, cudaStream_t strm)
void cuda_pwmin_sca2(void *a, real *c, int *n, cudaStream_t stream)
void cuda_cfill(void *a, real *c, int *n, cudaStream_t strm)
void cuda_glsc3_many(real *h, void *w, void *v, void *mult, int *j, int *n, cudaStream_t stream)
void cuda_absval(void *a, int *n, cudaStream_t stream)
void cuda_rzero(void *a, int *n, cudaStream_t strm)
void cuda_cdiv2(void *a, void *b, real *c, int *n, cudaStream_t strm)
void cuda_cdiv(void *a, real *c, int *n, cudaStream_t strm)
void cuda_iadd(void *a, int *c, int *n, cudaStream_t stream)
void cuda_pwmax_vec2(void *a, void *b, int *n, cudaStream_t stream)
void cuda_masked_atomic_reduction(void *a, void *b, void *mask, int *n, int *m, cudaStream_t strm)
void cuda_add5s4(void *a, void *b, void *c, void *d, void *e, real *c1, real *c2, real *c3, real *c4, int *n, cudaStream_t strm)
void cuda_invcol1(void *a, int *n, cudaStream_t strm)
void cuda_radd(void *a, real *c, int *n, cudaStream_t strm)
void cuda_add4s3(void *a, void *b, void *c, void *d, real *c1, real *c2, real *c3, int *n, cudaStream_t strm)
void cuda_subcol3(void *a, void *b, void *c, int *n, cudaStream_t strm)
void cuda_pwmin_vec3(void *a, void *b, void *c, int *n, cudaStream_t stream)
Object for handling masks in Neko.