mash@mips.UUCP (03/23/87)
Does anybody out there have any measured numbers for these? i.e., I'm especially interested in the following: Whetstone: DP & SP FORTRAN 80386 (with FP) HP-840 new MASSCOMPs IBM systems, anything current over 6,000 DP KWhets VAX 8700, VAX/VMS [I've got Pascal numbers, want FORTRAN] VAX 8XXX, VAX/VMS [I've got 8600 & 8650, anything else would be of interest, say over 4,000 DP KWhets] Linpack: (DP & SP) x (FORTRAN & Coded BLAS) 80386 (with FP) HP-840 new MASSCOMPs [I'm just finishing up next MIPS Performance Brief, and numbers would be helpful. This time I've included numbers for as many machines as possible.] If you have publishable numbers, please mail them to me, incl. configuration, compiler, etc; I'll include them in our next Brief, and summarize that for the net. Thanks! -- -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> UUCP: {decvax,ucbvax,ihnp4}!decwrl!mips!mash, DDD: 408-720-1700, x253 USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086
clif@intelca.UUCP (03/28/87)
> Does anybody out there have any measured numbers for these? i.e., > I'm especially interested in the following: > > Whetstone: DP & SP FORTRAN > 80386 (with FP) > HP-840 > new MASSCOMPs > IBM systems, anything current over 6,000 DP KWhets > VAX 8700, VAX/VMS [I've got Pascal numbers, want FORTRAN] > VAX 8XXX, VAX/VMS [I've got 8600 & 8650, anything else > would be of interest, say over 4,000 DP KWhets] > Linpack: (DP & SP) x (FORTRAN & Coded BLAS) > 80386 (with FP) > HP-840 > new MASSCOMPs > > -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> > UUCP: {decvax,ucbvax,ihnp4}!decwrl!mips!mash, DDD: 408-720-1700, x253 > USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086 John (and anyone else who is interested) Here are the 386 numbers Benchmark Environment The results were obtained with the Green Hill's FORTRAN compiler version 1.8.2G. The system we used to run was an Intel Multibus I system running Unix System V Release 3.0. The CPU board was a 386/24 MultiBus I which has a 64 Kbyte direct-mapped write-through cache and 2-3 wait states for cache misses. The floating point processors were the Intel 80387-16 and -20 MHz and the Weitek 1167 chip set (also at 16 and 20 MHz) Benchmark 16 MHz 20 MHz 387 1167 387 1167 Whetstone (FORTRAN 77) Single 1510K 3310K 1860K 4170K Double 1380K 2070K 1730K 2590K Linpack (FORTRAN) Rolled Single 182K 494K 225K 662K Rolled Double 151K 306K 187K 383K Unrolled Single 193K 541K 238K 669K Unrolled Double 158K 323K 195K 405K There is still some room for improvements in the compiler but these numbers should be fairly representative of the performance of our processors. The version of the compiler which generated these numbers is stil very new (less than two weeks) so there is the possibility that you may have recieved out of date results. -- Clif Purkiser, Intel, Santa Clara, Ca. {pur-ee,hplabs,amd,scgvaxd,dual,idi,omsvax}!intelca!clif These views are my own property. However anyone who wants them can have them for a nominal fee.
wunder@hpcea.UUCP (04/02/87)
I can quote numbers for the HP900/840 from the Performance Brief (Part Number 5953-8783, 6/86). Whetstones: B1 (single) 2795 B1D (double) 1927 Linpack (coded BLAS): Single 753 Double 574 These are pretty old numbers. I've seen graphs of benchmarks vs. compiler release, and things keep getting faster. The Perf. Brief claims 10,000 Dhystones (1.0), and I know we do 12,000 (1.1) now. You should call up the sales office in Palo Alto (Neely Sales Region) and see if you can get some current figures. If I find anything more current which I can publish, I'll send it along. Disclaimer from page one of the Performance Brief: Hewlett-Packard does not warrant or represent that the performance data stated in this document will be acheived by any particular customer application. Walter Underwood a Hewlett-packard employee not involved with benchmarking