Excelsior JET Benchmarks

SciMark 2.0a

From SciMark Home Page:

SciMark is a composite Java benchmark measuring the performance of numerical codes occurring in scientific and engineering applications. It consists of five computational kernels: FFT, Gauss-Seidel relaxation, Sparse matrix-multiply, Monte Carlo integration, and dense LU factorization.

These kernels are chosen to provide an indication of how well the underlying JVM/JITs perform on applications utilizing these types of algorithms. The problems sizes are purposely chosen to be small in order to isolate the effects of memory hierarchy and focus on internal JVM/JIT and CPU issues. A larger version of the benchmark (SciMark 2.0 LARGE) addresses performance of the memory subsystem with out-of-cache problem sizes.

SYSTEM CONFIGURATION

CPU:		Intel Pentium IV with HyperThreading running at 2.40 GHz
RAM:		512 MB
OS:		Windows 2000 Professional
Video:		RADEON 9200 SE 64MB at 1024x768x16777216c

RESULTS

Longer bar means better performance.

	JET	HSC	HSS	BEA		JET	HSC	HSS	BEA
Composite Score	172.29	153.39	319.07	248.36		112%	100%	208%	162%
FFT	54.07	70.13	183.60	192.52		77%	100%	262%	275%
LU	430.43	220.64	567.47	333.50		195%	100%	257%	151%
Monte Carlo	17.39	10.18	14.41	75.02		171%	100%	142%	737%
SOR	195.11	342.63	592.91	364.82		57%	100%	173%	106%
Sparse matmult	164.46	123.37	236.93	275.94		133%	100%	192%	224%
Average:						124%	100%	206%	276%

SYSTEM CONFIGURATION

CPU:		AMD Athlon XP running at 1.0 GHz
RAM:		512 MB
OS:		Windows 2000 Advanced Server
Video:		S3 Virge DX/GX at 1024x768x16777216c

RESULTS

Longer bar means better performance.

	JET	HSC	HSS	BEA		JET	HSC	HSS	BEA
Composite Score	157.87	124.08	157.23	170.28		127%	100%	127%	137%
FFT	65.51	52.06	119.27	107.35		126%	100%	229%	206%
LU	358.07	185.91	314.30	238.78		193%	100%	169%	128%
Monte Carlo	30.90	22.14	37.03	48.52		140%	100%	167%	219%
SOR	168.44	245.66	217.10	233.63		69%	100%	88%	95%
Sparse matmult	166.42	114.61	98.46	223.14		145%	100%	86%	195%
Average:						133%	100%	144%	163%