Characterization of scientific workloads on systems with multi-core processors

DMCA

Characterization of scientific workloads on systems with multi-core processors (2006)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Sadaf R. Alam , Richard F. Barrett , Jeffery A. Kuehn , Philip C. Roth , Jeffrey S. Vetter

Venue:	In IISWC
Citations:	24 - 0 self

Citations

653	Fast parallel algorithms for short-range molecular dynamics - Plimpton - 1995 (Show Context) Citation Context ...pteron multi-core processors. Based on our observations of the micro-benchmark and kernel results, we experimented with a set of representative applications from bio-molecular and climate simulations =-=[7, 9, 10]-=-. We confirmed that the memory and task placement configurations that result in an optimal performance for scientific kernels provide 10-20% performance improvement for full application runs. This val...
132	AMBER, a package of computer programs for applying molecular mechanics, normal mode analysis, molecular dynamics and free energy calculations to simulate the structure and energetic properties of molecules. - Pearlman, Case, et al. - 1995 (Show Context) Citation Context ...pteron multi-core processors. Based on our observations of the micro-benchmark and kernel results, we experimented with a set of representative applications from bio-molecular and climate simulations =-=[7, 9, 10]-=-. We confirmed that the memory and task placement configurations that result in an optimal performance for scientific kernels provide 10-20% performance improvement for full application runs. This val...
38	Molecular dynamics simulations in biology. - Karplus, GA - 1990 (Show Context) Citation Context ...rmance for almost all test cases. 4 Applications 4.1 Molecular Dynamics Simulations Molecular dynamics (MD) simulations enable the study of complex, dynamic processes that occur in biological systems =-=[8]-=-. MD methods are now used routinely to investigate the structure, dynamics, functions, and thermodynamics of biological molecules and their complexes. The types of biological activity that have been i...
32	a computer program for applying molecular mechanics, normal mode analysis, molecular dynamics and free energy calculations to elucidate the structures and energies of molecules - Pearlman, Case, et al. - 1995 (Show Context) Citation Context ...pteron multi-core processors. Based on our observations of the micro-benchmark and kernel results, we experimented with a set of representative applications from bio-molecular and climate simulations =-=[7, 9, 10]-=-. We confirmed that the memory and task placement configurations that result in an optimal performance for scientific kernels provide 10-20% performance improvement for full application runs. This val...
9	et al., “Practical performance portability - Jones, Worley - 2004 (Show Context) Citation Context ...pteron multi-core processors. Based on our observations of the micro-benchmark and kernel results, we experimented with a set of representative applications from bio-molecular and climate simulations =-=[7, 9, 10]-=-. We confirmed that the memory and task placement configurations that result in an optimal performance for scientific kernels provide 10-20% performance improvement for full application runs. This val...
7	Barszcz et al., “The NAS Parallel Benchmarks (94 - Bailey, E - 1994 (Show Context) Citation Context ...cientific calculations. To evaluate computation characteristics of multi-core processors, we measured the performance of several low-level micro-benchmarks and a subset of the NAS Parallel Benchmarks =-=[4]-=- running on test systems with multi-core Opteron processors. We observed the expected effect that performance remained comparable to that of single-core processors when memory accesses were satisfied ...
6	Design Considerations for Shared Memory - Ekman, Mucci - 2005 (Show Context) Citation Context ...mmon link to memory. To evaluate the communication performance of multi-core processors, we investigated the intra-node communication behavior of several MPI benchmarks on our multi-core test systems =-=[6, 11]-=-. Although a hybrid programming model that uses OpenMP for parallelism within a node and MPI for parallelism among nodes is often proposed as the best way to use systems with multi-core processors, th...
5	Boville et al - Blackmon, B (Show Context) Citation Context ...ons that perform FFT operations. The performance impacts are similar to what we observed in AMBER. 4.2 Parallel Ocean Program (POP) POP is the ocean component of Community Climate System Model (CCSM) =-=[5, 7]-=-. The code is based on a finitedifference formulation of the three-dimensional flow equations on a shifted polar grid. In its high-resolution configuration, 1/10-degree horizontal resolution, the code...
1	http://www.cs.virginia.edu/stream - benchmarks (Show Context) Citation Context ...he ongoing early systems evaluation effort at the Oak Ridge National Laboratory. 3.1 STREAM: Memory Latency & Bandwidth STREAM is a benchmark used for determining a system’s maximum memory throughput =-=[14]-=-. The LMbench benchmark suite includes an implementation of the STREAM benchmark. Figure 2 and Figure 3 show bandwidth scaling plots based on the LMbench3 STREAM-triad benchmark. Global bandwidth - 2 ...