gnu@hoptoad.uucp (John Gilmore) (04/01/88)
dfk@duke.cs.duke.edu (David Kotz) wrote: > Each processor: > single 300-pin CMOS chip has > 50 Mips fixed-point unit > 100 Mflop float unit > 128 (32 bit) registers > interface to switch (50 Mbytes/s) > and two 200 Mbyte/s channels to > 4M of data RAM (=> 128 Gbytes total) > 1M of instruction RAM > Processors are packed 8 to a board (actually 16, all are replicated). I wouldn't mind a workstation with this 50 MIP chip in it -- maybe a laptop, since it only takes 1/16th of a board to implement the whole thing, according to this spec. What's the catch, does a single CPU/RAM draw more than 15 amps at 120 volts? 'Twould seem that rather than build one TF-1 with 32,000 chips, you would do better selling 32,000 workstations (~ half of all the Suns in existence) that ran 16x as fast as Sun-3's. -- {pyramid,pacbell,amdahl,sun,ihnp4}!hoptoad!gnu gnu@toad.com "Don't fuck with the name space!" -- Hugh Daniel
malcolm@spar.SPAR.SLB.COM (Malcolm Slaney) (04/02/88)
In article <4297@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes: >'Twould seem that rather than build one TF-1 with 32,000 chips, you >would do better selling 32,000 workstations (~ half of all the Suns >in existence) that ran 16x as fast as Sun-3's. Well, a single 50MIP processor sure sounds nice but I remember some IBM speech people describing how they thought real time connected word natural speech recognition would consume a good portion of the TF-1 (or whatever it is called.) I suspect that natural (connected word, speaker independent) speech recognition done cheaply will be a significantly bigger market than Sun has ever dreamed about. Speech recognition has always been a problem where brute force can make for big wins. I think the numbers Fred Jelink (sp?) presented at the DARPA Speech Recognition workshop were based on the current "best" algorithms with the desired vocabulary (>5000 words) and a complete english grammar. I'm sure that lots of other people can come up with uses for a machine with 32k processors. This sounds to me like a Connection machine done right :-). Cheers. Malcolm
brooks@lll-crg.llnl.gov (Eugene D. Brooks III) (04/02/88)
>In article <4297@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes: >'Twould seem that rather than build one TF-1 with 32,000 chips, you >would do better selling 32,000 workstations (~ half of all the Suns >in existence) that ran 16x as fast as Sun-3's. I would put serious doubts on how soon we would need 32000 of these buggers in one box, but no matter how fast your processor is, two of them in a shared memory configuration done right is twice as fast. We would have little problem using a few tens or so effectively on "real dirty physics codes" at the moment and doubling the number each year as more is learned makes quite a lot of sense. I would agree that 32000 workstations would generate a higher profit, but that may not be what the TF-1 is all about.
jk3k+@andrew.cmu.edu (Joe Keane) (04/05/88)
In article <30@spar.SPAR.SLB.COM>, malcolm@spar.SPAR.SLB.COM (Malcolm Slaney) writes: > I'm sure that lots of other people can come up with uses for a machine with > 32k processors. How about one hell of a game-playing (read `chess') machine. You don't even need game-specific hardware. Of course it'd make it all that much faster... --Joe
david@daisy.UUCP (David Schachter) (04/06/88)
What word processor does it run? Is it PC-DOS compatible? Does it run Flight Simulator?
gwu@clyde.ATT.COM (George Wu) (04/07/88)
How about a marketing answer instead of all the technical replies people
have been posting? Packaging single-processor workstations means competing
with Sun, Apollo, DEC, IBM, and the rest of the world. And since all those guys
already have a big jump, plus lotsa bucks to throw into research, it makes
things that much more competative. And just maybe, the workstation market is
starting to saturate. I for one don't think there will be quite so many makers
of workstations further down the road.
On the other hand, who out there commercially builds massively parallel
machines? Obviously, there's the Connection Machine. I'm sure there are more,
thought they just don't pop into my mind, but the competition just isn't as
fierce.
--
George J Wu
UUCP: {ihnp4,ulysses,cbosgd,allegra}!clyde!gwu
ARPA: gwu%clyde.att.com@rutgers.edu or gwu@faraday.ece.cmu.edu