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

Merge pull request #2282 from martin-frbg/issue2281 

Optimize RPCC function on ARM64
This commit is contained in:
Martin Kroeker 2019-10-25 09:56:30 +02:00 committed by GitHub
parent 4ae1610f37 b687fba5bc
commit 48f5a89f92
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 11 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
common_arm64.h
View File

@ -78,7 +78,18 @@ static void __inline blas_lock(volatile BLASULONG *address){
#define BLAS_LOCK_DEFINED #define BLAS_LOCK_DEFINED
#if !defined(OS_DARWIN) && !defined (OS_ANDROID)
static __inline BLASULONG rpcc(void){
BLASULONG ret = 0;
__asm__ __volatile__ ("isb; mrs %0,cntvct_el0":"=r"(ret));
return ret;
}
#define RPCC_DEFINED
#define RPCC64BIT
#endif
static inline int blas_quickdivide(blasint x, blasint y){ static inline int blas_quickdivide(blasint x, blasint y){
return x / y; return x / y;