doon@unsvax.UUCP (Harry W. Reed) (08/06/87)
Hi, Now that NSC's latest and greatest Series 32000 chip is about to debut I have a question. What comes after the '532? What new periph's ? What new co-processors etc? Can anybody at NSC say anything? Cheers, Harry Reed (a devout 32000 fan)
grenley@nsc.nsc.com (George Grenley) (08/07/87)
In article <167@unsvax.UUCP> doon@unsvax.uucp (Harry W. Reed) writes: > Now that NSC's latest and greatest Series 32000 chip is about to >debut I have a question. What comes after the '532? What new periph's ? >What new co-processors etc? Can anybody at NSC say anything? Yes, many of us are capable of speech, although we are still working on speech recognition... 8-) Complete this sequence: 032,332,532,? (6? 7? 8?) But seriously, folks, new cpu designs take YEARS. In other words, CPU projects overlap. But, it is really too soon for me to say anything. As far as new peripherals, there is a higher performance math chip, the '381. There are also other peripherals, but again, I cannot say much about 'em. Go design in the '532. Now. We need the bucks. Besides that, it's the fastest general purpose high-integration CPU you can (almost) buy. Our marketing people sez they'll sample in Q1 - I suspect early Q1. Watch this space for details... GET BUSY!!!
david@sun.uucp (David DiGiacomo) (08/10/87)
In article <4538@nsc.nsc.com> grenley@nsc.UUCP (George Grenley) writes: >Go design in the '532. Now. We need the bucks. Besides that, it's >the fastest general purpose high-integration CPU you can (almost) buy. >Our marketing people sez they'll sample in Q1 - I suspect early Q1. Is it really faster than the 68030? The vague numbers I have seen seem to indicate otherwise, and the 68030 seems to be somewhat closer to sampling. Serious question: What advantages does the 32532 have when compared to the 68030, 80386, and MIPS R2000? (I left out the Sun/Fujitsu SPARC since it doesn't have an on-chip MMU.) -- David DiGiacomo, Sun Microsystems, Mt. View, CA sun!david david@sun.com
mash@mips.UUCP (John Mashey) (08/11/87)
In article <25238@sun.uucp> david@sun.uucp (David DiGiacomo) writes: > ...it == 532.... >Is it really faster than the 68030? The vague numbers I have seen seem to >indicate otherwise, and the 68030 seems to be somewhat closer to sampling. In order to make a comparison, are there any useful numbers available for the 68030? I've seeen recent Moto claims that the 030 is 5.5mips at 16MHz, and 7mips at 20MHz; I'm not sure what flavor of mips those are. They don't exactly seem like VAX-mips: that would give about 2.9 cycles/vmip, which is almost twice as good as the current best 68020 number I know [ 4 vmips Sun-3/2000, 25MHz, no-wait-state 64K virtual writeback cache, 64-bit bus, i.e., aggressive, well-designed memory system, but still needs >6 cycles/vmip]. It's hard to see how the 1-cyle reduction in bus cycles, plus the on-chip data cache [whose net benefit is unclear] halve the cycles/mips. It seems that a 68030 is a reasonable way to build a cheaper lower-speed system [i.e., a cacheless 3-mips design], in that it has an MMU, and may get some mileage from the caches. It's not instantly obvious why a 20MHz 030 is much faster than the 4mips Sun-3 [4.5mips, maybe]. Presumably clock rates will improve. Were those the nubmers you ahd in mind? Does anybody have any other estimates that have some numbers attached? -- -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> UUCP: {decvax,ucbvax,ihnp4}!decwrl!mips!mash OR mash@mips.com DDD: 408-991-0253 or 408-720-1700, x253 USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086
srm@violet.berkeley.edu (Richard Mateosian) (08/20/87)
>Complete this sequence: 032,332,532,? (6? 7? 8?)
132 and 232 were actually contemplated products. They just never came out.
432 was deliberately skipped, since certain folks thought that by the time
it (now the 532) came out, people would not yet have forgotten a disaster
that once bore the number 432.
Richard Mateosian 2919 Forest Avenue 415/540-7745
srm@violet.Berkeley.EDU Berkeley, CA a daver@nsc.nsc.com (Dave Raulino) (08/21/87)
In article <4814@jade.BERKELEY.EDU> srm@violet.berkeley.edu (Richard Mateosian) writes: >>Complete this sequence: 032,332,532,? (6? 7? 8?) > >132 and 232 were actually contemplated products. They just never came out. > >432 was deliberately skipped, since certain folks thought that by the time >it (now the 532) came out, people would not yet have forgotten a disaster >that once bore the number 432. > Actually, the number sequence is based on the assumption that a for a "pure" processor, the lead digit is an odd number (I'm not going to argue whether 0 is odd/even). The 132 was an 032 set up for dual processing, 332 is the 2nd generation, 532 the third, etc. (I'll leave it as an exercise for the reader to determine the name of the next member of the family). By the way, I'm in marketing, so even if this is not true, it now is because I said so. So there... 8-) And anyway, about that 432 comment, I don't know if that would matter much. I mean, Audi used to build a Fox, now VW does. (Okay, they're sort of the same company). Champagne wishes and Caviar dreams. -- Thank you for your continued support -- Dave R.
greg@xios.XIOS.UUCP (Greg Franks) (08/27/87)
In article <585@winchester.UUCP> mash@winchester.UUCP (John Mashey) writes: > >6 cycles/vmip]. It's hard to see how the 1-cyle reduction in bus cycles, >plus the on-chip data cache [whose net benefit is unclear] halve the >cycles/mips. I seem to recall that the 68030 can access the code and data caches simultaneously (Harvard architecture?). Perhaps this feature enables the speed up. -- Greg Franks XIOS Systems Corporation, 1600 Carling Avenue, (613) 725-5411 Ottawa, Ontario, Canada, K1Z 8R8 seismo!mnetor!dciem!nrcaer!xios!greg "Vermont ain't flat!"