floraachy
/
OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Calculation of chunk range limits was ignoring num_cpu 

bug introduced by me in #1262 - should fix #1388
This commit is contained in:
Martin Kroeker 2017-12-09 18:44:34 +01:00 committed by GitHub
parent 281a2b952f
commit e90a147493
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
driver/level2/gbmv_thread.c
View File

@ -230,10 +230,10 @@ int CNAME(BLASLONG m, BLASLONG n, BLASLONG ku, BLASLONG kl, FLOAT *alpha, FLOAT
#ifndef TRANSA
range_m[num_cpu] = num_cpu * ((m + 15) & ~15);
if (range_m[num_cpu] > m) range_m[num_cpu] = m;
if (range_m[num_cpu] > m * num_cpu) range_m[num_cpu] = m * num_cpu;
#else
range_m[num_cpu] = num_cpu * ((n + 15) & ~15);
if (range_m[num_cpu] > n) range_m[num_cpu] = n;
if (range_m[num_cpu] > n * num_cpu) range_m[num_cpu] = n * num_cpu;
#endif
queue[num_cpu].mode = mode;