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

Add build-time option for OMP scheduler; document MULTITHREAD_THRESHOLD range (#1620) 

* Allow choosing the OpenMP scheduler and add range hint for GEMM_MULTITHREAD_THRESHOLD
* 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 11:25:05 +02:00 committed by GitHub
parent 12603b7dbb
commit 47bf0dba8f
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 18 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
Makefile.rule
View File

@ -60,6 +60,14 @@ VERSION = 0.3.1.dev
# This flag is always set for POWER8. Don't modify the flag
# USE_OPENMP = 1
# The OpenMP scheduler to use - by default this is "static" and you
# will normally not want to change this unless you know that your main
# workload will involve tasks that have highly unbalanced running times
# for individual threads. Changing away from "static" may also adversely
# affect memory access locality in NUMA systems. Setting to "runtime" will
# allow you to select the scheduler from the environment variable OMP_SCHEDULE
# CCOMMON_OPT += -DOMP_SCHED=dynamic
# You can define maximum number of threads. Basically it should be
# less than actual number of cores. If you don't specify one, it's
# automatically detected by the the script.
@ -156,8 +164,11 @@ NO_AFFINITY = 1
# CONSISTENT_FPCSR = 1
# 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
# in small matrix sizes. The default value is 4.
# with single thread. (Actually in recent versions this is a factor proportional to the
# number of floating point operations necessary for the given problem size, no longer
# 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
# If you need santy check by comparing reference BLAS. It'll be very

6
driver/others/blas_server_omp.c
View File

@ -48,6 +48,10 @@
#else
#ifndef OMP_SCHED
#define OMP_SCHED static
#endif
int blas_server_avail = 0;
static void * blas_thread_buffer[MAX_PARALLEL_NUMBER][MAX_CPU_NUMBER];
@ -331,7 +335,7 @@ int exec_blas(BLASLONG num, blas_queue_t *queue){
break;
}
#pragma omp parallel for schedule(static)
#pragma omp parallel for schedule(OMP_SCHED)
for (i = 0; i < num; i ++) {
#ifndef USE_SIMPLE_THREADED_LEVEL3