Cyclops Tensor Framework
parallel arithmetic on multidimensional arrays
dgtog_redist.h
Go to the documentation of this file.
1 
2 #include "dgtog_calc_cnt.h"
3 
4 namespace CTF_int {
10  double dgtog_est_time(int64_t tot_sz, int np);
11 
12  void dgtog_reshuffle(int const * sym,
13  int const * edge_len,
14  distribution const & old_dist,
15  distribution const & new_dist,
16  char ** ptr_tsr_data,
17  char ** ptr_tsr_new_data,
18  algstrct const * sr,
19  CommData ord_glb_comm);
20 
21  void redist_bucket_r0(int * const * bucket_offset,
22  int64_t * const * data_offset,
23  int * const * ivmax_pre,
24  int rep_phase0,
25  int rep_idx0,
26  int virt_dim0,
27  bool data_to_buckets,
28  char * __restrict__ data,
29  char ** __restrict__ buckets,
30  int64_t * counts,
31  algstrct const * sr,
32  int64_t data_off,
33  int bucket_off,
34  int prev_idx);
35 
36 }
double dgtog_est_time(int64_t tot_sz, int np)
estimates execution time, given this processor sends a receives tot_sz across np procs ...
void dgtog_reshuffle(int const *sym, int const *edge_len, distribution const &old_dist, distribution const &new_dist, char **ptr_tsr_data, char **ptr_tsr_new_data, algstrct const *sr, CommData ord_glb_comm)
void redist_bucket_r0(int *const *bucket_offset, int64_t *const *data_offset, int *const *ivmax_pre, int rep_phase0, int rep_idx0, int virt_dim0, bool data_to_buckets, char *__restrict__ data, char **__restrict__ buckets, int64_t *counts, algstrct const *sr, int64_t data_off, int bucket_off, int prev_idx)
Definition: dgtog_bucket.h:90
def np(self)
Definition: core.pyx:315