From: Martin Kroeker <martin@ruby.chemie.uni-freiburg.de>
Date: Fri, 15 Jun 2018 09:25:05 +0000 (+0200)
Subject: Add build-time option for OMP scheduler; document MULTITHREAD_THRESHOLD range (#1620)
X-Git-Tag: v0.3.1~1^2~17
X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=47bf0dba8f7a9cbd559e2f9cabe0bf2c7d3ee7a8;p=platform%2Fupstream%2Fopenblas.git

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
---

diff --git a/Makefile.rule b/Makefile.rule
index 5c03d01..649aabe 100644
--- a/Makefile.rule
+++ b/Makefile.rule
@@ -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
diff --git a/driver/others/blas_server_omp.c b/driver/others/blas_server_omp.c
index fccdb43..4255852 100644
--- a/driver/others/blas_server_omp.c
+++ b/driver/others/blas_server_omp.c
@@ -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