Why so little parallelism?
|
From: Eugene Miya on 3 Nov 2006 10:48 >> we do agree here. von Neumann In article <1162546273.195679.202060(a)b28g2000cwb.googlegroups.com>, BDH <bhauth(a)gmail.com> wrote: >So if I was smart enough to handle the position I want, I wouldn't be >on Usenet? Ouch! So do something. Accomplish something. Well, in the early days of Usenet, when far fewer people were here, in fact guys like Alan Smith (Mr. cache memory) and John Hennessy (now Stanford's President) were here (posted) or lurked. Others had assistants lurk, post and troll. They looked for talented people to hire. Discussion quality was high (it's typically higher in the comp.* groups). Then after a while as the net grew (I'll use a graphics example) some of these people had to tell their people NOT to reveal corporate info on the net. But its a two way street. Now many people can make a living off the net. I am lucky in this regard to a point. My job was to watch how some of the money was spent by the guys whose mail and news server you are using. So I have a pretty idea what they doing scanning your email and your posts (and mine, but I won't get a gmail acct.). That's fine with me. If you are young and sharp, build a machine, write some code, write a paper, etc. Do something to benefit the world or the net or something. --
From: Eugene Miya on 3 Nov 2006 10:55 In article <1162546968.282024.247210(a)i42g2000cwa.googlegroups.com>, BDH <bhauth(a)gmail.com> wrote: >> >Golomb rulers are already found with massively parallel computation. I > >They're found by exhaustive searching of a reduced search space. It's >easy to partition that for massive parallelism. Then I suggest go find the next set and do it. >> >Some people definitely have a higher opinion of him than I do though. >> von Neumann > >Hey, I have time... >...or maybe I don't. Probably don't. =( Write the next definitive book on game theory or some new topic. >I was thinking of APL, but thanks for that. I am not the APL expert you need. Maybe track down Bernecky. Or maybe not, he talks to me because I attempt to maintain my parallelism biblio. >Thinking Machines went under too, but we still have Microsoft. Sure, and Tom Blank, Marpar's architect works for them. >I suspect that in another 10 years people will still be assuming >incorrect things about compilers. Oh I suspect if you think that, then, they probably already do now. If you have a better idea for a parallelizing compiler: do it. Do better than Kuck or Kennedy. I know they would like to be bettered. But keep Amdahl in the brain. >I guess my question was, is there something about APL that is >incompatible with either someone educated a certain way or inherently >incompatible with the majority of programmers, or was it just >happenstance of history? My guess is all of the above. Compatibility works against you. You would do better not asking here as much as comp.lang.apl. The document/typesetting guys have similar problems and they are only beginning to learn that. The history of computing is largely about neglect and ignorance. >> Anyways. I'm outta here. >Bye. Briefly back. --
From: Eugene Miya on 3 Nov 2006 11:02 >> The Star -100 In article <1162547455.916567.38180(a)k70g2000cwa.googlegroups.com>, BDH <bhauth(a)gmail.com> wrote: >CDC, I guess it was run in a smart way. But some kinds of useful smart >need tempering to keep them away from insanity. I guess maybe they were >missing that. >Of course, usually companies get much less intelligent insanity. Well CDC began as a very focused company for a number of public and a number of secret reasons (as in: we may never know, because we have no need to know). The problem began when as it was growing it hired a bunch of IBM business guys who apparently decided to expand their market and they also took their money to attempt to diversity into other more commodity areas, to change the general world, etc. There were business decisions involved. They lost a number of their best people. They were doing hard technical things. There was a CDC lecture back in June, but I skipped off on an annual trip to Alaska to help some friends. What I heard was that it was a marketing love fest w/o any views of technical history. I have not viewed the tape. In fact when there was an Alto lecture also a few years back but in June I have not gone to view that Alto lecture (I just go eat salad with some of my PARC buddies [I'll see some of them next weekend anyways]). --
From: Del Cecchi on 3 Nov 2006 19:52 "Eugene Miya" <eugene(a)cse.ucsc.edu> wrote in message news:454b7638$1(a)darkstar... >>> The Star -100 > > In article <1162547455.916567.38180(a)k70g2000cwa.googlegroups.com>, > BDH <bhauth(a)gmail.com> wrote: >>CDC, I guess it was run in a smart way. But some kinds of useful smart >>need tempering to keep them away from insanity. I guess maybe they were >>missing that. >>Of course, usually companies get much less intelligent insanity. > > Well CDC began as a very focused company for a number of public and a > number of secret reasons (as in: we may never know, because we have > no need to know). The problem began when as it was growing it > hired a bunch of IBM business guys who apparently decided to expand > their market and they also took their money to attempt to diversity > into > other more commodity areas, to change the general world, etc. > There were business decisions involved. > They lost a number of their best people. > They were doing hard technical things. > > There was a CDC lecture back in June, but I skipped off on an annual > trip to > Alaska to help some friends. What I heard was that it was a marketing > love fest w/o any views of technical history. I have not viewed the > tape. > In fact when there was an Alto lecture also a few years back but in > June > I have not gone to view that Alto lecture (I just go eat salad with > some > of my PARC buddies [I'll see some of them next weekend anyways]). > They lost seymour cray. They weren't content in their niche. The market changed and they didn't adapt. Bill Norris got the big head disease. Pick your reason. del > --
From: BDH on 4 Nov 2006 02:06
> If you are young and sharp, build a machine, write some code, write a > paper, etc. Do something to benefit the world or the net or something. That's a good position, and I guess it's sad to say I don't care half as much today as I did. What, you have something in mind? Video compression? |