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

Amended description of GEMM_MULTITHREAD_THRESHOLD 

to reflect #742 making it track floating point operations rather than matrix size
This commit is contained in:
Martin Kroeker 2018-06-15 09:20:17 +02:00 committed by GitHub
parent a217049dc6
commit b9d1ba2e3a
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 5 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
Makefile.rule
View File

@ -164,9 +164,11 @@ NO_AFFINITY = 1
# CONSISTENT_FPCSR = 1 # CONSISTENT_FPCSR = 1
# If any gemm arguement m, n or k is less or equal this threshold, gemm will be execute # If any gemm arguement m, n or k is less or equal this threshold, gemm will be execute
# with single thread. You can use this flag to avoid the overhead of multi-threading # with single thread. (Actually in recent versions this is a factor proportional to the
# in small matrix sizes. The default value is 4, but values as high as 50 have been # number of floating point operations necessary for the given problem size, no longer
# reported to be optimal for certain workloads (50 is the recommended value for Julia). # an individual dimension). You can use this setting to avoid the overhead of multi-
# threading in small matrix sizes. The default value is 4, but values as high as 50 have
# been reported to be optimal for certain workloads (50 is the recommended value for Julia).
# GEMM_MULTITHREAD_THRESHOLD = 4 # GEMM_MULTITHREAD_THRESHOLD = 4
# If you need santy check by comparing reference BLAS. It'll be very # If you need santy check by comparing reference BLAS. It'll be very