VLIW pre-history
|
From: Alex Colvin on 14 May 2007 14:22 >Well, like I said, try programming it by hand, and if you can't, it >must be VLIW. no, it's just me. -- mac the na�f
From: Alex Colvin on 14 May 2007 14:40 >Well, like I said, try programming it by hand, and if you can't, it >must be VLIW. maybe it's just me ;-) Historically, it seems that microcodes, like CISC, assumes a skilled programmer able to manage a limited number of resources. User code compiled to an instruction set interpreted by carefully-crafted microcode. VLIW, like RISC, assumed a compiler algorithm able to manage a much larger set of resources. User code compiles directly to VLIW code. Once it was understood how to compile to VLIW, the need for an intermediate instruction set was gone. -- mac the na�f
From: Stephen Fuld on 14 May 2007 15:04 Eugene Miya wrote: >>> (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$yM2.3407(a)bgtnsc04-news.ops.worldnet.att.net>, > Stephen Fuld <S.Fuld(a)PleaseRemove.att.net> wrote: >> I'm sure that depends upon the definition of VLIW. > > What about the Univac ISP Stephen? While it was after the APs in the 70s, > would you consider it and ILP/VLIW machine? While I am aware of it, and actually talked to people at NSA who had one and had to deal with it, I don't know enough about it to be able to answer your question. Sorry. :-( -- - Stephen Fuld (e-mail address disguised to prevent spam)
From: Eugene Miya on 14 May 2007 20:15 >>>> (Rau and Fisher say "The earliest VLIW processors built were the >>>> so-called attached array processors ..." [IBM 2938/3838, AP-120B], >>>> "Instruction-level parallel processing: History, overview, and >>>> perspective," J. Supercomputing, 1993.) >> >> What about the Univac ISP ? >> would you consider it and ILP/VLIW machine? In article <fp22i.150255$VU4.78872(a)bgtnsc05-news.ops.worldnet.att.net>, Stephen Fuld <S.Fuld(a)PleaseRemove.att.net> wrote: >and had to deal with it, I don't know enough about it to be able to >answer your question. >Sorry. :-( No need to apologize. The other Steve formerly at Unisys I knew who knew something about it is unreachable in the bowels of MS. He was interesting in his own right independent of the ISP. It's just too bad we didn't have one at JPL. --
From: John L on 14 May 2007 20:01
>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 had one to play with. The I/O was an issue, but the difficulty of writing programs that kept all the functional units busy was even more of an issue. R's, John |