Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!wuarchive!brutus.cs.uiuc.edu!usc!bloom-beacon!spdcc!merk!alliant!lewitt From: lewitt@Alliant.COM (Martin E. Lewitt) Newsgroups: comp.arch Subject: Re: VLIW Architecture Keywords: VLIW Message-ID: <3460@alliant.Alliant.COM> Date: 10 Oct 89 11:12:10 GMT References: <251FCB3F.12366@maccs.dcss.mcmaster.ca> <1050@m3.mfci.UUCP> <13050@pur-ee.UUCP> <1630@l.cc.purdue.edu> <1989Oct5.025841.2046@esegue.segue.boston.ma.us> <3449@alliant.Alliant.COM> <1071@m3.mfci.UUCP> Reply-To: lewitt@alliant.Alliant.COM (Martin E. Lewitt) Organization: Alliant Computer Systems, Littleton, MA Lines: 63 In article <1071@m3.mfci.UUCP> colwell@mfci.UUCP (Robert Colwell) writes: >In article <3449@alliant.Alliant.COM> lewitt@Alliant.COM (Martin Lewitt) writes: >>The Multiflow machine was no mystery to FPS sales analysts. It was the >>FPS dream machine: UNIX, virtual memory, better compilers, etc. We had >>asked FPS to give us these features right from the beginning of the 164 >>back in 1981. By the time, Multi-flow delivered, it was too, late, >>Convex, Alliant and FPS's own ECL machine, 264 were already on the scene, >>and the high end workstations arrived in short order. > >Too late for whom or for what? What a weird paragraph. I agree, what a weird paragraph! By "too late", I meant that Multiflow was too late to exploit the wide open market opportunities that the FPS analysts saw when they were specifying the machine they wanted FPS to build, i.e., a wide instruction machine supporting virtual memory and UNIX. FPS's analysts saw these requirements of the market place in plenty of time to preempt the later arrivals of Convex and Alliant. I was with Alliant when the Multi-flow Trace 7 first started shipping. While we didn't fear the Trace 7 in most application areas, the Trace 7 did turn out to be a tough competitor in certain areas, e.g., the integer (non-floating point) ECAD codes. However, after some initial competitive success for the Trace 7, I noticed that Alliant was no longer losing this market to the Multi-flow, but to the SUN-4 Sparcstation instead. While this was bad news for Alliant, it had to be terrible news for Multi-flow. A key area of competitive strength against the product that Alliant was shipping at that time, was being undercut by cheap integer mips from SUN. So this is another sense in which I meant "too late". *** stuff deleted *** >marketing hype, to me. And if your references to having a workable compiler >were legit, then why did Convex's first machine beat the 164, which had more >mflops under the hood? Convex's first machine the C1-XL had more megaflops, 40 single and 20 double peak megaflops vs the FPS's 11 peak megaflops. Even so the Convex rarely beat the FPS-164 on benchmarks unless the codes were highly vectorizable. The XL had poor scalar performance and the FPS-164 often beat it by a factor of two. Good scalar performance is a strength of these wide instruction machines. The C1-XP had much improved scalar performance but still faced tough competition from the 164. Of course, performance is only one purchase criterion and FPS could lose sales because of the lack of virtual memory, lack of a multi-user operating system, reliability concerns, etc. >> 4) The first commercially available machine to successfully exploit >> parallel processors automatically using "dusty deck", serial >> FORTRAN, the Alliant FX8, (by 4 years and counting). > >"And counting"? What does that mean? No competitors seem to have replicated Alliant's success in automatically producing code for a MIMD machine. So Alliant's lead is 4 years and growing. By success, I mean that this is the standard way most customers use the machine. Most Cray jobs still only use one CPU. *** more stuff deleted *** >Bob Colwell ..!uunet!mfci!colwell -- Phone: (206) 931-8364 Martin E. Lewitt My opinions are Domain: lewitt@alliant.COM 2945 Scenic Dr. SE my own, not my UUCP: {linus|mit-eddie}!alliant!lewitt Auburn, WA 98002 employer's.