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

Fix z/ctrmv stack allocation on AMD bulldozer and barcelona target 

* Hopefully, because this was found by error and trial (dark magic)
* Ref #786
This commit is contained in:
Jerome Robert 2016-06-07 16:11:09 +02:00
parent f04af36ad0
commit d346c533b1
1 changed files with 2 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
interface/ztrmv.c
View File

@ -243,6 +243,8 @@ void CNAME(enum CBLAS_ORDER order, enum CBLAS_UPLO Uplo,
#endif #endif
{ {
buffer_size = ((n - 1) / DTB_ENTRIES) * 2 * DTB_ENTRIES + 32 / sizeof(FLOAT); buffer_size = ((n - 1) / DTB_ENTRIES) * 2 * DTB_ENTRIES + 32 / sizeof(FLOAT);
// It seems to be required for some K8 or Barcelona CPU
buffer_size += 8;
if(incx != 1) if(incx != 1)
buffer_size += n * 2; buffer_size += n * 2;
} }