Abstract
It is generally agreed that memory requirements should be taken into account in the scheduling of parallel jobs. However, so far the work on combined processor and memory scheduling has not been based on detailed information and measurements. To rectify this problem, we present an analysis of memory usage by a production workload on a large parallel machine, the 1024-node CM-5 installed at Los Alamos National Lab. Our main observations are - The distribution of memory requests has strong discrete components, i e. some sizes are much more popular than others. - Many jobs use a relatively small fraction of the memory available on each node, so there is some room for time slicing among several memory-resident jobs. - Larger jobs (using more nodes) tend to use more memory, but it is difficult to characterize the scaling of per-processor memory usage.
| Original language | English |
|---|---|
| Title of host publication | Job Scheduling Strategies for Parallel Processing - IPPS 1997 Workshop, Proceedings |
| Editors | Dror G. Feitelson, Larry Rudolph |
| Publisher | Springer Verlag |
| Pages | 78-94 |
| Number of pages | 17 |
| ISBN (Print) | 3540635742, 9783540635741 |
| DOIs | |
| State | Published - 1997 |
| Event | 3rd Workshop on Job Scheduling Strategies for Parallel Processing, IPPS 1997 - Geneva, Switzerland Duration: 5 Apr 1997 → 5 Apr 1997 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 1291 |
| ISSN (Print) | 0302-9743 |
| ISSN (Electronic) | 1611-3349 |
Conference
| Conference | 3rd Workshop on Job Scheduling Strategies for Parallel Processing, IPPS 1997 |
|---|---|
| Country/Territory | Switzerland |
| City | Geneva |
| Period | 5/04/97 → 5/04/97 |
Bibliographical note
Publisher Copyright:© Springer-Verlag Berlin Heidelberg 1997.