LAIPE Benchmark — parallel direct solver

        Most parallel libraries never disclose speedup. LAIPE is different. LAIPE can yield highly efficient speedup. Equation Soltuion provide benchmarks to show users a speedup.

        The benchmark applies a parallel multi-entry direct solver to solve a sparse band system [A]{X}={B}, of order 200000-by-200000 with a half bandwidth 8. Timing result can show speedup with respect to number of processors.

IMPLEMENTATION

        The benchmark detects how many processors or cores (hereinafter "computing elements") are available. Use one computing element to execute the benchmark, and report elapsed time. Then, use 2 computing elements to execute the benchmark, and so on. For example,

CPUs	elapsed time	speedup	efficiency
1	18.33	1.00	100.0%
2	9.29	1.97	98.7%
3	6.29	2.91	97.1%
4	4.80	3.81	95.5%

The timing result shows speedup and efficiency with respect to number of computing elements. The benchmark can be downloaded below.

DOWNLOAD

        Select a platform, and download the benchmark. For a measurement of elapsed time, don't run other programs on the computer when testing.