amd/blis
1
0
Fork 0
mirror of https://github.com/amd/blis.git synced 2026-05-12 18:15:37 +00:00
Code Issues Packages Projects Releases Wiki Activity

Aligning rntm_t struct to 64 byte address to address performance issues.

Browse Source 
Non aligned rntm_t struct can potentially have its first/last cache
line shared with other objects in memory. This could affect performance
depending on how much the shared cache lines are used. rntm_t struct is
aligned to 64 bytes to workaround this issue.

Change-Id: Id0956fca771be062ada9f81e8cd75ac1f290fd8e
This commit is contained in:
mkadavil
2023-11-30 06:33:41 +05:30
committed by MithunMohan KadavilMadanaMohanan
parent 4c77ef4953
commit 89bb999afa
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
frame/include/bli_type_defs.h
View File

@@ -1586,7 +1586,7 @@ typedef struct cntx_s
// NOTE: The order of these fields must be kept consistent with the definition
// of the BLIS_RNTM_INITIALIZER macro in bli_rntm.h.
typedef struct rntm_s
typedef struct __attribute__((aligned(64))) rntm_s
{
// "External" fields: these may be queried by the end-user.
bool auto_factor;