pjt@druhi.ATT.COM (Paul James Taysom) (05/06/90)
I have not yet seen any SPEC marks for the Intel 80x86 processor family. I'm particulary interested in the integer subset of the SPEC marks for the '486. Thanks, Paul James Taysom AT&T Bell Labs (303)538-4702
mash@mips.COM (John Mashey) (05/07/90)
In article <4776@druhi.ATT.COM> pjt@druhi.ATT.COM (Paul James Taysom) writes: >I have not yet seen any SPEC marks for the Intel 80x86 processor family. >I'm particulary interested in the integer subset of the SPEC marks >for the '486. SOmething like this was posted before: From Intel's "i486(tm) Microprocessor Performance, Release 1.0, Feb '90, Intel Corporation", p.27 25MHz i486, 128K write-back cache; C: Metaware High C; FORTRAN: SVS; I forget the versions. Code-seconds-ratio: gcc 107 13.8 espresso 186 12.2 spice2g6 2731 8.8 doduc 324 5.8 nasa7 3456 5.8 li 370 16.8 eqntott 100 11.0 matrix300 477 9.5 fpppp 431 7.0 tomcatv 619 4.3 The geometric mean of the 4 integer SPECmarks (gcc, espresso, li &eqntott) comes out at 13.3, vs. 19.7 for MIPS-25 (in M/2000). (Thus, the faster RISCs are about 1.5X faster on integer, and something like 2.5X faster on FP; as usual, FP varies more, and ratios betweeen systems vary more.) "For access to source programs & data files used in this study, please contact Intel 386(TM)DX/i486 Microprocessor Technical Marketing group at (408) 765-5767". -- -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> UUCP: mash@mips.com OR {ames,decwrl,prls,pyramid}!mips!mash DDD: 408-524-7015 or 408-720-1700 USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086