keller@uicsl.UUCP (04/19/84)
#N:uicsl:7000059:000:1262 uicsl!keller Apr 18 16:54:00 1984 (dangling by a single thread) The March 1984 issue of IEEE Spectrum is devoted to personal computers. There is an interesting paper on the TI Professional Computer that describes the speech processing option in medium detail. In the paper on HP's 68000 machines is a description of a project with MIT to make systems for the introductory programming class that is taught using the Scheme dialect of LISP. Each machine has 4M bytes of RAM and links to a 66M byte file server. They say that this isn't a commercial product. There is an article on the LISA but not the Macintosh. The Intel people put in a paper about the 80286 that has some timing comparisons with the 8086, 80186 and VAX 11/780. The timings are relative to an 8MHz 8086 with no wait states (They're all running at 8MHz.) Assembly EDN* Dig.Filter Sieve Berkeley Pascal 8086 1.0 1.0 1.0 1.0 1.0 80186 1.7 1.2 2.1 1.8 2.6 80286 3.0 2.1 4.0 3.7 5.0 2.6 VAX 11/780 * EDN refers to another set of benchmarks It is easy to see that the 80286 is FAST. They mention the 80386 which will be 2 to 3 times faster yet. You can run faster than an 8MHz clock if you can afford fast RAM. Watch out CRAY! Anyone got figures for 68000, 68020, Z8000, NS16032, HPxxxxx, RISC Chips, etc? -Shaun
dan@idis.UUCP (Dan Strick) (04/22/84)
Interesting. Intel finds that the 80286 is twice as fast as the VAX 11/780. I find something else. I got my hands on a real 5 MHz 80286 system last year and ran some simple cpu benchmarks. I found that the 80286 system ran between 3 to 6 times slower than the VAX 11/780. I suspect the difference is that Intel is reporting exceptionally favorable high level language benchmarks which may not be worth the paper they are printed on. I saw ads in at least one magazine about 2 years ago that reported the same kind of statistics (i.e. 8 MHz 80286 blows away competition) when the 8 MHz version of the 80286 was apparently a figment of the marketing department's imagination but all the other machines mentioned in the advertisement were real. Dan Strick University of Pittsburgh [decvax|mcnc]!idis!dan
ee163aca@sdccs7.UUCP (04/22/84)
[] Ugh! Intel plays games with the benchmarks again. I seem to recall a really glowing benchmark on the 8088 when it first came out. It made it look like it blew the doors off an 68000 :-). Not that I think that a 80286 is a real slow-poke, but remember Intel writes the rules for the benchmarks - they know the processor's strengths & weaknesses. So be sceptical when they say it blows a VAX 780's doors off ( who uses Berkley Pascal anyway? ). Wait till the *real* machine comes out. Sometimes I think Intel plays numbers games to hide their brain-damaged archi- tectures. Please, no flames ... Especially from Intel! Paul van de Graaf U.C. San Diego sdcsvax!sdccs7!ee163aca
binder@dosadi.DEC (Do not adjust your set...) (04/26/84)
>> Sometimes I think Intel plays numbers games to hide their brain-damaged >> architectures. > Sometimes? Nope, must be all the time. I was part of a group doing a design for a hi- availability machine based on the 432. We had lived through all the lies in the 432 specs, and then the degradations of the spec, and we actually had a proto partly running. And then Intel jerked the 432. No w the guy on the other side of the wall is doing something using the 80186. I wish him the best of luck - he'll need it. 286? Eek, I'm gun-shy. Don't walk up to me and say "Intel" - I won't be responsible for the reaction! Cheers, Dick Binder decvax!decwrl!rhea!dosadi!binder
keller@uicsl.UUCP (04/28/84)
#R:uicsl:7000059:uicsl:7000066:000:171
uicsl!keller Apr 28 12:51:00 1984
Oops!
Sorry, the IEEE journal with the INTEL paper
is NOT SPECTRUM--it is PROCEEDINGS.
So what about comparisons with other micros, esp. 68xxxx and RISC?
-Shaun