On Mon, 2004-06-28 at 14:32, Doug Shubert wrote: > Hello Joe, > > Your numbers look very linear to ours. > I would expect a 45%-50% performance > increase based on CPU speed (240 vs 246). > > Other factors could be in memory speeds and OS tuning. > Our Opteron is running DDR Reg ECC 266 memory. > > If you have a chance to run memtest86+, I would be interested > in the memory speed of a 246 system. Ok. Are you using a particular memtest? I found 3.1 was needed for Opteron. I am using DDR 333 memory. Lets also check the layout of your ram. Is it all piled on one processing node? If so, you want to set it up as interleaved. I get streams of about 2GB/s per CPU with streams triad. I am in the middle of some other stuff now, so I'll try to get the memtest86 data later tonight (unless you meant a different memtest...) > > Thanks for the link, I will test your NCBI patch too. > Doug > > >Our most > >recent BBS v1 run for a 246 gave: > > > >1,1,1083 > >2,2,1086 > >2,1,1093 > >3,2,2168 > >3,3,2168 > >3,1,2174 > >3,4,2174 > >4,1,559 > >5,1,1099 > >5,2,1107 > >6,1,556 > > > > > > > >> > >>BBS times run in 2GB ram disk for a 240 > >> > >>1,1,1641 > >>2,1,1647 > >>2,2,1685 > >>3,4,3289 > >>3,1,3296 > >>3,3,3332 > >>3,2,3335 > >>4,1,863 > >>5,1,1683 > >>5,2,1690 > >>6,1,848 > >> > >> > >> > >> > > _______________________________________________ > Bioclusters maillist - Bioclusters@bioinformatics.org > https://bioinformatics.org/mailman/listinfo/bioclusters -- Joseph Landman, Ph.D Scalable Informatics LLC, email: landman@scalableinformatics.com web : http://scalableinformatics.com phone: +1 734 612 4615