User-level dynamic page migration for multiprogrammed shared-memory multiprocessors
Dimitrios S. Nikolopoulos, Theodore S. Papatheodorou, Constantine D. Polychronopoulos et al. · Proceedings of the International Conference on Parallel Processing · 2002
This paper presents algorithms for improving the performance of parallel programs on multiprogrammed shared-memory NUMA multiprocessors, via the use of user-level dynamic page migration. The idea that drives the algorithms is that a page migration engine can perform accurate and timely page migrations in a multiprogrammed system if it can correlate page reference information with scheduling information obtained from the operating system. The necessary page migrations can be performed as a response to scheduling events that break the implicit association between threads and their memory affinity sets. We present two algorithms that use feedback from the kernel scheduler to aggressively migrate pages upon thread migrations. The first algorithm exploits the iterative nature of parallel programs, while the second targets generic codes without making assumptions on their structure. Performance evaluation on an SGI Origin2000 shows that our page migration algorithms provide substantial improvements in throughput of up to 264% compared to the native IRIX 6.5.5 page placement and migration schemes.
Cited by 50 works. doi:10.1109/icpp.2000.876083 Open-access PDF
Authors
- Dimitrios S. Nikolopoulos
- Theodore S. Papatheodorou
- Constantine D. Polychronopoulos
- Jesús Labarta
- Eduard Ayguadé
References
- The SGI Origin
- Process control and scheduling issues for multiprogrammed shared-memory multiprocessors
- Process control and scheduling issues for multiprogrammed shared-memory multiprocessors
- Operating system support for improving data locality on CC-NUMA compute servers
- The OpenMP Implementation of NAS Parallel Benchmarks and its Performance
- Scheduling and page migration for multiprocessor compute servers
- Scheduling and page migration for multiprocessor compute servers
- WildFire: a scalable path for SMPs
- Flexible use of memory for replication/migration in cache-coherent DSM multiprocessors
- Performance experiences on Sun's Wildfire prototype