Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!gatech!hubcap!fpst From: pratt@cs.stanford.edu (Vaughan Pratt) Newsgroups: comp.parallel Subject: Re: New i860 parallel machine Message-ID: <1991Apr18.205449.7627@hubcap.clemson.edu> Date: 18 Apr 91 18:43:26 GMT Sender: fpst@hubcap.clemson.edu (Steve Stevenson) Organization: Clemson University Lines: 51 Approved: parallel@hubcap.clemson.edu From: Steven Ericsson Zenith Date: Wed, 17 Apr 91 22:39:08 +0100 Subject: New i860 parallel machine Newsgroups: comp.parallel Building a parallel supercomputer today out of off the shelf parts has got to be real easy (given the time and money). That brings back memories. This was exactly our attitude at Stanford about workstations in 1980. We spent 11/80 to 8/81 building the Sun workstation out of off-the-shelf parts in our "garage", namely Margaret Jacks Hall 428, then Andy Bechtolsheim's office, previously Forest Baskett's, now mine. Sun the company was formed at the end of 2/82, and the ink turned black in 6/82. But use of off-the-shelf parts was only one of several key ingredients. Another was Unix, of which C was an integral part. Sun shipped off-the-shelf Unisoft Unix throughout 1982 prior to its in-house Lyon-Shannon port of Berkeley Unix. The GT860 needs the parallel supercomputer equivalent of Unix and C. This is not an off-the-shelf item today. But fixing this is more than just finding another Dennis Ritchie. The biggest obstacle is that we don't even know what concurrency is. People think they know now, but in hindsight in 2020 they will be able to see that they really did not know in 1991. In 1891 physicists had no inkling of what physics would look like in 1926. And the great majority had no inkling that they had no inkling. In 1991 computer scientists are in exactly the same situation with regard to concurrency. Understanding the nature of concurrency is a problem as important yet unsolved today as understanding the nature of matter was a century ago. (I predict that ideas will soon start to flow from the former to the latter, which is still far from wrapped up, but that's another story.) Although both sides of the Atlantic are attacking this question, in my view Europe is ahead of the US in this very important area. The largest electronic cache of US-generated papers on European-style concurrency resides on a disk in, by coincidence, the above-mentioned Sun garage. Its contents are listed in pub/README available by anonymous ftp from Boole.Stanford.EDU. Vaughan Pratt -- =========================== MODERATOR ============================== Steve Stevenson {steve,fpst}@hubcap.clemson.edu Department of Computer Science, comp.parallel Clemson University, Clemson, SC 29634-1906 (803)656-5880.mabell