OpenMP in Small Bites/NUMA - Revision history

Marc-andre-hermanns-bc32@rwth-aachen.de: Tweak page sorting and title

2020-12-04T15:42:14Z

Tweak page sorting and title

2020-12-04T14:53:55Z

Add DISPLAYTITLE

2020-12-01T09:01:31Z

2020-11-30T19:18:51Z

2020-11-30T19:18:29Z

2020-11-24T08:43:35Z

Adapt aspect ration of video embedding

2020-11-19T11:20:20Z

Remove unwanted initial line breaks

2020-11-18T09:28:27Z

Use new syllabus template

2020-11-02T11:05:50Z

Set correct video link

2020-11-02T10:09:49Z

@@ Line 1: / Line 1: @@
-[[Category:Tutorials]]<nowiki />
+[[Category:Tutorials|Non-Uniform Memory Access (NUMA) Effects in OpenMP]]<nowiki />
 {{DISPLAYTITLE:Non-Uniform Memory Access (NUMA) Effects in OpenMP}}<nowiki />
 {{Syllabus OpenMP in Small Bites}}<nowiki />

@@ Line 1: / Line 1: @@
 [[Category:Tutorials]]<nowiki />
 {{Syllabus OpenMP in Small Bites}}<nowiki />
 __TOC__

← Older revision		Revision as of 09:01, 1 December 2020
Line 8:		Line 8:

	<youtube width="600" height="340" right>MhlM-GiS1EM</youtube>		<youtube width="600" height="340" right>MhlM-GiS1EM</youtube>
		+
		+	([[Media:hpc.nrw_11_NUMA.pdf \| Slides as pdf]])

	=== Quiz === <!--T:5-->		=== Quiz === <!--T:5-->

@@ Line 3: / Line 3: @@
 __TOC__
-This video shows how a non-uniform memory access (NUMA) architecture ifluences the performance of OpenMP programs. It explains how distribute data and threads across NUMA domains and how to avoid uncontrolled data or thread migration.
+This video shows how a non-uniform memory access (NUMA) architecture influences the performance of OpenMP programs. It explains how distribute data and threads across NUMA domains and how to avoid uncontrolled data or thread migration.
 === Video === <!--T:5-->

@@ Line 2: / Line 2: @@
 {{Syllabus OpenMP in Small Bites}}<nowiki />
 __TOC__
 === Video === <!--T:5-->

@@ Line 5: / Line 5: @@
 === Video === <!--T:5-->
-<youtube width="600" height="400" right>MhlM-GiS1EM</youtube>
+<youtube width="600" height="340" right>MhlM-GiS1EM</youtube>
 === Quiz === <!--T:5-->

← Older revision		Revision as of 10:09, 2 November 2020
Line 45:		Line 45:
	</quiz>		</quiz>
	{{hidden end}}		{{hidden end}}
−
−	Why is it important to initialize your data in parallel when executing on a NUMA architecture? (Initializing the data in parallel distributes the date amoung the different sockets. When accessing the data in a hotspot region in the same pattern you avoid remote memory accesses.)
−	Why is it important to bind the threads? (Otherwise the operating system might decide to migrate the thread from one core to anothers. This results in a performance penalty due to the context switch and a potential remote data access on NUMA architectures.)
−	Given a NUMA architecture with to two sockets with six cores each: How can you place the threads of an OpenMP program running with 4 threads among both sockets and bind them to a core? (Set OMP_PROC_BIND=spread and OMP_PLACES=cores )