wasnsr@nmt.edu (T.O.R.S.O.) (03/28/91)
I jusr ran the mips program recently posted to comp.binaries.ibm.pc on my ACMA 386 20MHz, and the values it gave don't seem correct. Benchmark Perf Ibm pc Ibm AT Compaq Actual relative to -> 4.7Mhz 8Mhz 386 MIPS ------------------------------------------------------ General Instruc 8.40 2.44 1.23 : 1.39 Integer Instruc 18.49 2.89 1.27 : 3.11 Memory to Memory 7.51 2.31 1.29 : 1.79 Regis to Regis 22.00 2.86 1.20 : 3.96 Regis to Memory 7.90 2.37 1.31 : 2.43 Overall Perfor 10.62 2.56 1.26 : 2.54 Now, following the numbers, I am assuming that the lower the number the better. If this is so, am I wrong in assuming that a 386 20MHz should beat an Ibm AT (8MHz) in every category. For my machine, the NORTON SI 5.0 is 12.4 (CPU index), and for SI 4.5, the rating is 21.0. Is there something wrong with my machine, am I reading the numbers wrong, or is the MIPS program in error? Any help would be appreciated! Thanks in advance, wayne, wasnsr@jupiter.nmt.edu
antonyc@nntp-server.caltech.edu (Antony Chan) (03/28/91)
wasnsr@nmt.edu (T.O.R.S.O.) writes: > I jusr ran the mips program recently posted to comp.binaries.ibm.pc on my ACMA 386 20MHz, and the values it gave don't seem correct. > Benchmark Perf Ibm pc Ibm AT Compaq Actual > relative to -> 4.7Mhz 8Mhz 386 MIPS > ------------------------------------------------------ > General Instruc 8.40 2.44 1.23 : 1.39 > Integer Instruc 18.49 2.89 1.27 : 3.11 > Memory to Memory 7.51 2.31 1.29 : 1.79 > Regis to Regis 22.00 2.86 1.20 : 3.96 > Regis to Memory 7.90 2.37 1.31 : 2.43 > Overall Perfor 10.62 2.56 1.26 : 2.54 > Now, following the numbers, I am assuming that the lower the number the better. If this is so, am I wrong in assuming that a 386 20MHz should beat an Ibm AT (8MHz) in every category. i'm not an expert on this, but the way i read these numbers is that they tell you how many times faster than the computer at the top of the list is, i.e. in the category of general instuructions, your machine is 8.40 times as fast as a pc, 2.44 times as fast as an at, and 1.23 times as fast as a compaq 386, and you have an actual mips rating of 1.39. so bigger numbers are better.
davidsen@sixhub.UUCP (Wm E. Davidsen Jr) (03/31/91)
In article <1991Mar28.045507.3905@nmt.edu> wasnsr@nmt.edu (T.O.R.S.O.) writes: | Benchmark Perf Ibm pc Ibm AT Compaq Actual | relative to -> 4.7Mhz 8Mhz 386 MIPS | ------------------------------------------------------ | General Instruc 8.40 2.44 1.23 : 1.39 | Now, following the numbers, I am assuming that the lower the | number the better. If this is so, am I wrong in assuming that a | 386 20MHz should beat an Ibm AT (8MHz) in every category. If you read the headers, the number is the ration of *your* system to the computer named. So it 8.40 times faster then the PC, 1.23 times faster then the Compaq original 386. Since that was a 386-16 I'm not surprised. A 386 is not N times faster in every category, so the performance is check in a number of areas. The original Compaq didn't have cache, either, my GV-386 would beat it by 19% on some benchmarks just beacuse of this. -- bill davidsen - davidsen@sixhub.uucp (uunet!crdgw1!sixhub!davidsen) sysop *IX BBS and Public Access UNIX moderator of comp.binaries.ibm.pc and 80386 mailing list "Stupidity, like virtue, is its own reward" -me