Argonne National Laboratory

Performance and Scalability Evaluation of 'Big Memory' on Blue Gene Linux

TitlePerformance and Scalability Evaluation of 'Big Memory' on Blue Gene Linux
Publication TypeJournal Article
Year of Publication2011
AuthorsYoshii, K, Iskra, K, Naik, H, Beckman, PH, Broekema, PC
JournalInternational Journal of High Performance Computing Applications
Start Page148-160
Date Published05/2011
Other NumbersANL/MCS-P1690-1109

We address memory performance issues observed in Blue Gene Linux, and discuss the design and implementation of ""Big Memory," an alternative, transparent memory space introduced to eliminate the memory performance issues. We evaluate the performance of Big Memory using custom memory benchmarks, NAS Parallel Benchmarks, and the Parallel Ocean Program, at a scale of up to 4096 nodes. We find that Big Memory is successful in resolving performance issues we had encountered. For the ocean simulation program, we even find that Linux provides better scalability than does the lightweight compute node kernel designed solely for high-performance applications. Originally intended exclusively for compute node tasks, our new memory subsystem turns out to dramatically improve the performance of certain I/O node applications as well. We demonstrate this performance using the central processor of the LOw Frequency ARray radio telescope as an example.