peters@cubsvax.UUCP (Peter S. Shenkin) (02/26/86)
Here are some replies I got to my query about array processor discussions on the net. Several people suggested net.arch, but I've been monitoring that news group for the past month, and nothing has appeared on the subject (though there have been other interesting postings). I did get some interesting mail on array processors, and this appears below, followed with some comments of my own. >From phri!roy Thu Jan 30 03:51:34 1986 >Organization: Public Health Research Inst. (NY, NY) > > Peter, as far as I know, there really isn't a "good" place to talk >about AP's on the net. I suppose net.arch is about as good as any. At any >rate, *I'm* interested in talking about them. We just got a grant from NSF >with which we will be buying about a dozen Sun workstations and some sort >of AP. I've been kicking around some algorithms with Peter Sellers at >Rockefeller for doing sequence homology searches that look like they would >work well on an AP. > > We originally were planning on a Floating Point Systems 5205. I >think Dayton Clark at Courant has one of these. There is a software >package available for running these under Unix that looks to be very nice. >The biggest problem with the FPS is that it is expensive and big. > > Now we're looking at the ZIP-3232 from Mercury. At first glance, >it seems to be about the same performance as the FPS, at about half the >price. Eric Schwartz at Bellevue has one of these, and seems to be happy >with it. Are you familiar with either of these products? As I mentioned in my original posting, we have a Star ST-100, which we're hosting with a microvax running VMS. This is a good arrangement. (Since microvax has a q-bus and the Star needs a unibus, we have an Able microverter to convert formats. This works well.) But I thought people on the net should know that Star is working on drivers that will allow the ST-100 to be hosted from UNIX; it's being done for a specific customer, and may or may not be released as an announced commercial package, and it's specific to Berkeley 4.2 running on a VAX, but if someone wants to buy a Star and is running 4.2 on a VAX I'm sure you'll be able to get it. Talk to Star if interested. Peter S. Shenkin Columbia Univ. Biology Dept., NY, NY 10027 {philabs,rna}!cubsvax!peters cubsvax!peters@columbia.ARPA