Modeling the communication performance of the IBM SP2

Abandah, Gheith A.; Davidson, Edward S.

doi:10.1109/ipps.1996.508066

Cited by 30 publications

(27 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The general idea of prediction from partial measurement is adopted here. Abandah and Davidson [32], and Boyd et al [33] proposed hierarchical modeling methods for parallel machines that is kindred in spirit to our work, and was effective on machines in the early and mid 90's.…”

Section: Hpcmentioning

confidence: 79%

A genetic algorithms approach to modeling the performance of memory-bound computations

Tikir

Carrington

Strohmaier

et al. 2007

Proceedings of the 2007 ACM/IEEE Conference on Supercomputing

View full text Add to dashboard Cite

Benchmarks that measure memory bandwidth, such as STREAM, Apex-MAPS and MultiMAPS, are increasingly popular due to the "Von Neumann" bottleneck of modern processors which causes many calculations to be memory-bound. We present a scheme for predicting the performance of HPC applications based on the results of such benchmarks. A Genetic Algorithm approach is used to "learn" bandwidth as a function of cache hit rates per machine with MultiMAPS as the fitness test. The specific results are 56 individual performance predictions including 3 full-scale parallel applications run on 5 different modern HPC architectures, with various CPU counts and inputs, predicted within 10% average difference with respect to independently verified runtimes.

show abstract

Section: Hpcmentioning

confidence: 79%

A genetic algorithms approach to modeling the performance of memory-bound computations

Tikir

Carrington

Strohmaier

et al. 2007

Proceedings of the 2007 ACM/IEEE Conference on Supercomputing

View full text Add to dashboard Cite

show abstract

“…The maximum measured in any single scheduling step is also much less than a second. While these numbers are significantly higher than for EASY (where simulations complete in a few seconds), they are still very low in absolute terms, especially relative to other overheads on parallel machines, where loading a parallel job for execution may take several minutes [1].…”

Section: Running Timementioning

confidence: 86%

Backfilling with lookahead to optimize the packing of parallel jobs

Shmueli

Feitelson

2005

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

The utilization of parallel computers depends on how jobs are packed together: if the jobs are not packed tightly, resources are lost due to fragmentation. The problem is that the goal of high utilization may conflict with goals of fairness or even progress for all jobs. The common solution is to use backfilling, which combines a reservation for the first job in the interest of progress with packing of later jobs to fill in holes and increase utilization. However, backfilling considers the queued jobs one at a time, and thus might miss better packing opportunities. We propose the use of dynamic programming to find the best packing possible given the current composition of the queue, thus maximizing the utilization on every scheduling step. Simulations of this algorithm, called LOS (Lookahead Optimizing Scheduler), using trace files from several IBM SP parallel systems, show that LOS indeed improves utilization, and thereby reduces the mean response time and mean slowdown of all jobs. Moreover, it is actually possible to limit the lookahead depth to about 50 jobs and still achieve essentially the same results. Finally, we experimented with selecting among alternative sets of jobs that achieve the same utilization. Surprising results indicate that choosing the set at the head of the queue does not necessarily guarantee best performance. Instead, repeatedly selecting the set with the maximal overall expected slowdown boosts performance when compared to all other alternatives checked.

show abstract

“…Multistage interconnection networks (MINs) are often proposed in high performance environments such as multiprocessor systems [1], high-bandwidth communication networks (e.g. ATM or Gigabit/10G Ethernet), and are candidates for application in distributed real-time systems.…”

Section: Introductionmentioning

confidence: 99%