Hello All I'm looking for people with access to multiprocessor machines who would be willing to run some benchmarks. The application in question is a numerically intensive Quantum Molecular Dynamics code. It currently runs on multiprocessor Linux boxes (Intel based), SGI/Cray T3E's and Origin 2000's and IBM power machines running AIX 4.3. The code is highly parallel and has been run on a 1024 processor T3E and a 128 processor O2K. The benchmarks would be split into three categories with the following hardware requirements. Small: 1-4 processors with a minimum of 128 Megs/RAM Medium: 1-32 processors with a minimum of 1 Gigs/RAM Large: 8-128 processors with a minimum of 4 Gigs/RAM Software requirements are a C compiler and blas and lapack libraries. I am looking for this information to get some baseline performance figures for a paper to be published in a computational physics journal. I would like to include some of the benchmark results in the paper. The subject of the paper will primarily deal with computational issues -- in particular the efficient parallelization of the algoriths used in the code on different computer architectures. I currently have access to T3E's, Origin 2000's, and Intel based Linux. (Dual Pentium-Pro). I would especially like to get some results for a quad Xeon and Linux Alpha/Sparc machines. (Alpha and Sparc would require some porting though). If anyone is willing to run some benchmarks please contact me at. [EMAIL PROTECTED] Thanks Emil Briggs P.S. The code is GPL but is to specialized to be of much general use. However some of the implementations of the numerical algorithms are potentially useful in a general sense (e.g. parallel multigrid solvers) and I eventually hope to package them as libraries. - Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/ To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
