randy@petfe.UUCP (Randy Banton) (05/13/85)
Has anyone run the BYTE benchmarks on a 68020 based system with UNIX System V yet?? (Motorola guys, are you listening?) I have a benchmark paper from the Intel literature group which claims a 6Mhz 80286 is 1.38 a 10Mhz 68010. The 286 machines were the Intel 286/310 and IBM PC/AT. The 68010 machines were a Sun 2/120 and a Masscomp workstation. Next they determined that a 10Mhz 80286 (0 wait states) is 2.85 times the same 10Mhz 68010 machines. Note these times are all measured on real machines (as opposed to paper calculations). The 10Mhz 80286 is also rated as equal to the 16Mhz 68020. The 68020 assumptions were zero wait states and that a 16Mhz 68020 was 2.84 times a 10Mhz 68010 (i.e. no real 020 system). They finally extrapolate that a 12.5Mhz 80286 (0 wait states) is 1.27 times a 16Mhz 68020 (0 wait states). Without arguing the merits of benchmarks, has anyone run any of the "common" benchmarks a 16Mhz (or 12.5Mhz) 68020 system? (For those who haven't seen it, Intel had a two page color ad about the results mentioned above. It was is Electronic News, I believe May 6, 1985. The report I mention is called "iAPX 286 High Performance Benchmark Study Report" and is dated April 1985.) Randy
henry@utzoo.UUCP (Henry Spencer) (05/16/85)
It sure would be nice if Intel would normalize their benchmark reports
to memory access times rather than clock rates, since different chips
vary so much in the amount of work they do in one clock cycle. But I
guess their chips wouldn't show up so well then... :-)
--
Henry Spencer @ U of Toronto Zoology
{allegra,ihnp4,linus,decvax}!utzoo!henrycarter@masscomp.UUCP (Jeff Carter) (06/08/85)
> > I have a benchmark paper from the Intel literature group which > claims a 6Mhz 80286 is 1.38 a 10Mhz 68010. The 286 machines > were the Intel 286/310 and IBM PC/AT. The 68010 machines were > a Sun 2/120 and a Masscomp workstation. > > Next they determined that a 10Mhz 80286 (0 wait states) is 2.85 > times the same 10Mhz 68010 machines. Note these times are all measured > on real machines (as opposed to paper calculations). The 10Mhz >>>>>See the comments below as to the 'Real Machine' bit. <<<<<< > 80286 is also rated as equal to the 16Mhz 68020. The 68020 > assumptions were zero wait states and that a 16Mhz 68020 was 2.84 > times a 10Mhz 68010 (i.e. no real 020 system). > > They finally extrapolate that a 12.5Mhz 80286 (0 wait states) > is 1.27 times a 16Mhz 68020 (0 wait states). > > I believe May 6, 1985. The report I mention is called "iAPX 286 > High Performance Benchmark Study Report" and is dated April 1985.) When I saw the ad mentioned above, I was slightly concerned at the claims therein, since it is based on benchmarks on a MASSCOMP MC-500. The reference quoted in the Intel ad refers to benchmarks run 18 to 24 months ago on the 68010 systems and only recently on the 286-based systems. I think it is particularly enlighting to review the benchmark results obtained on a MASSCOMP vs. the iAPX-286 based machines. Results from a 10MHz MASSCOMP MC-500 running RTU (MASSCOMP's Real-time UNIX) Version 2.2A (current production release, not pre-release) run in May 1985 give the following results: 68010:286 68010:286 bmark MC-500 6MHz 286(avg) relative perf. (As Reported by Intel) ----- ------ ------------- ------------ -------------- pipes 1.5s 2.10s 1.4 0.646 scall 5.3s 6.95s 1.31 0.942 sieve 2.3s 2.85s 1.24 1.075 dwrite 1.2s 1.15s 0.96 0.657 dread 1.6s 1.3s 0.81 0.265 shell 1.5s 3.20s 2.13 0.914 loop 6.3s 12.0s 1.90 1.805 ------------------------------------ ------ avg perf. 68010/286 1.39 0.72 Note the inversion of performance levels if current benchmarks are used on both machines! If we use the method presented in the benchmark report to extrapolate to 68020 and 12.5Mhz 286 performance, using this new baseline, the results are a bit different than in Intel's ad. Processor Rel. Perf. _________ _________ VAX 750 0.67 AT&T 7300 0.71 6MHz 286 1.00 VAX 780 1.08 MASSCOMP 1.39 <- base for 68020 estimation 8MHz 286 1.65 10MHz 286 2.07 12.5MHz 286 2.62 (est.) 16.7MHz 68020 3.95 (est.) This is a slightly different result than presented in the ad. Please don't misinterpret this 'response'. I don't think this selection of benchmarks is particularly good for the purpose stated, of comparing processor speed and extrapolating to non-existent hardware. The suite includes too many disk benchmarks to have real meaning for this purpose. These benchmarks are really UNIX implementation benchmarks, as can be seen if the 3 68010 processors are compared to each other. (Use the results above and Intel's benchmark report, or contact me directly if you want more details on this suite.) US Mail: Jeff Carter, Masscomp, 1 Technology Park, Westford MA 01886 Tel: (617)692-6200 x226 Net: ...!{decvax|ihnp4|harpo}!masscomp!carter
johnt@tekecs.UUCP (John Theus) (06/13/85)
In article <716@masscomp.UUCP> carter@masscomp.UUCP (Jeff Carter) writes: >> >> I have a benchmark paper from the Intel literature group which >> claims a 6Mhz 80286 is 1.38 a 10Mhz 68010. The 286 machines >> were the Intel 286/310 and IBM PC/AT. The 68010 machines were >> a Sun 2/120 and a Masscomp workstation. >> >> I believe May 6, 1985. The report I mention is called "iAPX 286 >> High Performance Benchmark Study Report" and is dated April 1985.) > >When I saw the ad mentioned above, I was slightly concerned at the >claims therein, since it is based on benchmarks on a MASSCOMP >MC-500. The reference quoted in the Intel ad >refers to benchmarks run 18 to 24 months ago on the >68010 systems and only recently on the 286-based systems. I just saw this Intel ad in the June 13 issue of EDN. Following the ad's data comparing the 68010, 68020, and 286 is the following quote: "So there you have it. The 286. Absolutely the most powerful microprocessor architecture in the world." Just about says it all. One wonders why Intel is bothering with the 386. John Theus Tektronix 6200 Series Development tektronix!tekecs!johnt