eugene@pioneer.arpa (Eugene Miya N.) (12/29/87)
Just a short message: Marty mentions parallel hardware and programmability. >In article <76700007@uiucdcsp> johnson@uiucdcsp.cs.uiuc.edu writes: >> >>Even though this is comp.arch, I will argue that the real problem in >>building parallel systems is finding the parallel algorithms to run on them. > >I agree. There is also another problem which has to do with >theoretical understanding of computational complexity. Algorithm > . . . >It seems to me that some useful work could be done in putting together >an algorithm complexity model which would be useful for dealing with a >multiple processor application. . . There is a recent book by Leah Jamieson et al on this. If you need a reference let me know. Before you blame hardware, software or algorithms, the folks at LLNL and LANL have a nice triangular symbol with all three topics (sort of like the "fire triangle") where you need all three components to achieve parallelism (or fire). I think Minsky is quoted somewhere as "thinking in threes." From the Rock of Ages Home for Retired Hackers: --eugene miya, NASA Ames Research Center, eugene@ames-aurora.ARPA "You trust the `reply' command with all those different mailers out there?" "Send mail, avoid follow-ups. If enough, I'll summarize." {uunet,hplabs,hao,ihnp4,decwrl,allegra,tektronix}!ames!aurora!eugene