From: Eugene Miya on 14 May 2007 13:34 In article <46454fff$1(a)news.meer.net>, Greg Lindahl <lindahl(a)pbm.com> wrote: >Sounds like ordinary microcode on a pipelined processor. The FPS >AP120B array processor (which didn't come along until the early 80s) I had one at JPL in Fall 1977. >let you microcode it directly. But since it's just microcode, it's not >exactly a novel idea, it's just unusual today to think of directly >programming it that way. I think there are FPS ads in the Sept. 77 Sci Am which touted it as a 7600 class processor on your PDP-11 (not bad, but I/O then becomes your bottle neck). We could use one for the Museum. -- ------------ And now a word from our sponsor ------------------ Want to have instant messaging, and chat rooms, and discussion groups for your local users or business, you need dbabble! -- See http://netwinsite.com/sponsor/sponsor_dbabble.htm ----
From: Eugene Miya on 14 May 2007 13:36 >> (Rau and Fisher say "The earliest VLIW processors built were the >> so-called attached array processors ..." [IBM 2938/3838, AP-120B], >> "Instruction-level parallel processing: History, overview, and >> perspective," J. Supercomputing, 1993.) In article <T1c1i.10046$yM2.3407(a)bgtnsc04-news.ops.worldnet.att.net>, Stephen Fuld <S.Fuld(a)PleaseRemove.att.net> wrote: >I'm sure that depends upon the definition of VLIW. What about the Univac ISP Stephen? While it was after the APs in the 70s, would you consider it and ILP/VLIW machine? --
From: Eugene Miya on 14 May 2007 13:43 >>> VLIW isn't just wide microcode with multiple functional units. It's >>> also the compiler techniques with trace scheduling and speculative >>> execution that let it keep large numbers of units busy. Eric wrote: >>That doesn't explain how I can look at a piece of hardware and >>definitively state that it is or is not VLIW. I think it should >>be possible to do that without looking at the software. I have memories of seeing my first Trace board.... In article <f28b1i$ujk$1(a)gal.iecc.com>, John L <johnl(a)iecc.com> wrote: >Well, like I said, try programming it by hand, and if you can't, it >must be VLIW. So gee Eric, does that mean we are going to dust off the Multiflow and grab Josh to reboot it? 8^) --
From: Eugene Miya on 14 May 2007 13:49 In article <1179113537.089454.4260(a)n59g2000hsh.googlegroups.com>, Quadibloc <jsavard(a)ecn.ab.ca> wrote: >So I may be mistaken in some of my thinking about this topic. > >But I would think that there is a very simple definition of VLIW. >Instructions explicitly code for superscalar operation. So go back to Fisher's original ELI. Or ILP. >Note that the Cyberplus from Control Data had multiple functional >units which could all be busy at once, but did not have a large number >of fast data memories that could be used as sources and destinations >in the same instruction, so it functioned as a *dataflow* >architecture. Well many firms kept multiple functional units busy. The AFP/Cyberplus was far from the first dataflow machine. I can think of 3 firms which used that word such as what which later became Alliant, and I think that other one which was Dana/Ardent/Stardent,.... You then have to define who does and does not constitute a functional unit. >The term VLIW has since been applied to such things as a DSP from >Texas Instruments that fetches eight instructions at once, with these ELI was so much simpler..... -- ------------ And now a word from our sponsor ------------------ Do your users want the best web-email gateway? Don't let your customers drift off to free webmail services install your own web gateway! -- See http://netwinsite.com/sponsor/sponsor_webmail.htm ----
From: Eugene Miya on 14 May 2007 14:11
In article <f22rat$o79$1(a)hubcap.clemson.edu>, Mark Smotherman <mark(a)clemson.edu> wrote: >I wonder if there are other VLIW designs prior to the definition >of the term in the 1980's. I am aware of: >- IBM SSEC (1948) - two instructions in a "line of sequence", >- various horizontal microprogramming approaches (e.g., the > 1953 paper suggesting horizontal microcode by Wilkes and > Stringer; some suggest Turing's ACE, ca. 1946) >- Elliott 152 (1950) and Elliott 153 (1954) - the 154 had a >- array processors, including > IBM 2938 Array Processor (1969) > IBM 3838 Array Processor (1974) > FPS AP-120B (1975) >- Culler patent (1973) - "Data processor with parallel operations >- Pomerene patent (1981) - "Machine for multiple instruction >- Rau's Polycyclic Architecture project at TRW/ESL (1981) > and then his work at Cydrome (1983-1988) >- Fisher's ELI-512 design (1983) > and then his work at Multiflow (1984-1990) >Anyone know of other VLIW-like designs pre-1980? > >(Rau and Fisher say "The earliest VLIW processors built were the >so-called attached array processors ..." [IBM 2938/3838, AP-120B], >"Instruction-level parallel processing: History, overview, and >perspective," J. Supercomputing, 1993.) Did you also get their book from the same year? We got the Cydra 5 and the Trace at the Museum but they are likely never to be placed on display, because the public can't discriminate between one box and another at that level. I still have to look for samples of the various FPS boxes (the hypercube as well as the 164 and 264, and trying to explain this to the somewhat clueless, Cray-biased, curators is my next headache on them). Mark, you would know better, but I've even heard the Stretch (7030) as a precursor VLIW. I think most claims except Fisher's are somewhat overrated and in some cases after the fact. You might also looks of Senzig's VAMP paper (remote chance). >P.M. Melliar-Smith, "A design for a fast computer >for scientific calculations," in 1969 AFIPS FJCC, pp. 201-208. >. (He's writing in reaction to >execution resources "squandered" and "wasted" by a Tomasulo-like >E-box coupled with a one-instruction-decode-per-cycle I-box.) Yeah, I'd go with the quotes. A lot of that kind of thinking back then. -- ------------ And now a word from our sponsor ------------------ Want to have instant messaging, and chat rooms, and discussion groups for your local users or business, you need dbabble! -- See http://netwinsite.com/sponsor/sponsor_dbabble.htm ---- |