Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!ukc!dcl-cs!aber-cs!athene!pcg From: pcg@cs.aber.ac.uk (Piercarlo Grandi) Newsgroups: comp.sys.m88k Subject: Re: Tektronix shutdown & move away from 88k's?? Message-ID: Date: 16 Nov 90 16:37:10 GMT References: <1990Oct25.183519.19324@iconsys.uucp> <42487@mips.mips.COM> <1990Nov14.142702.19478@unx.sas.com> Sender: aro@aber-cs.UUCP Organization: Coleg Prifysgol Cymru Lines: 64 Nntp-Posting-Host: odin In-reply-to: massengi@unx.sas.com's message of 14 Nov 90 14:27:02 GMT On 14 Nov 90 14:27:02 GMT, massengi@unx.sas.com (Darrell Massengill) said: massengi> In article <42487@mips.mips.COM> mash@mips.COM (John Mashey) writes: [ ... who has chosen which RISC ... ] massengi> Yes, but remember it "ain't" over 'til its over. I don't know massengi> if anyone is really "committed". For example, Electronic News massengi> (Oct 29) had 2 articles that indicated that people "committed" massengi> to MIPS were moving away. Not really, not really. Let me defend MIPS for a change. massengi> It indicated that Stardent which has MIPS boxes, was going to massengi> move to the '860 chip for the "next-generation" systems. (page massengi> 13). Only as vector coprocessors; they still keep MIPSes as their main engine. The 860 makes for a poor general purpose chip, but an excellent quasi-vector one. Meiko also sell systems that have SPARCs as main engines, transputers for supporting parallel programming, and 860s as vector processors hanging off the transputers. massengi> And the article on the new DEC RISC architecture indicated massengi> that DEC will be moving away from MIPS (at least long term). Well what is know is that given fabrication difficulties for ECL and advances in CMOS geometries DEC have decided not to have any product based on the R6000 ECL MIPS; they will use some unspecified new implementation of the MIPS architecture in CMOS. This is not moving away from MIPS. It is also known that the VMS people find that the VAX architecture is not competitive with the MIPS one in price/performance, so they are moving away from the VAX architecture, towards an as-yet unannounced VMS oriented RISC architecture of their own. The (comparatively small) Ultrix side of DEC seems to remain loyal to MIPS. I would bet that they ain't going to use something like MIPS for the VMS oriented RISC, more probably something like streamlined VAX. massengi> I don't think you can count anyone out yet. I think that massengi> there are likely to be shifting commitments for some number of massengi> years, especially as the architectural limitations of some of massengi> the chips are encountered. Well, as far as workstations are concerned, AMD, HP-PA, ARM[23], and Clipper are out of the running. For workstations it is pretty much SPARC vs. everybody else, and that everybody else is MIPS and 88K (The IBM America is the dark horse that may yet have the last laugh). 88K has to contend with the fact that not only it is strictly proprietary, it is strictly proprietary to a company that owns the 68K architecture, a situation not dissimilar to DEC and the VAX. Motorola may be betting either way, like DEC is doing (68k/VMS I win; 88k/MIPS I don't lose). If they get the Apple account (and every indication seems to be that they have) the 88k will be for most observers as well positioned as the MIPS architecture. As to me, I am most happy with open architectures, that can be implemented or second sourced by many parties. -- Piercarlo Grandi | ARPA: pcg%uk.ac.aber.cs@nsfnet-relay.ac.uk Dept of CS, UCW Aberystwyth | UUCP: ...!mcsun!ukc!aber-cs!pcg Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk