Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!ogicse!dali!uakari.primate.wisc.edu!zaphod.mps.ohio-state.edu!rpi!uupsi!sunic!dkuug!freja.diku.dk!skinfaxe.diku.dk!thorinn From: thorinn@skinfaxe.diku.dk (Lars Henrik Mathiesen) Newsgroups: comp.arch Subject: Re: Yet Another Upgrade Anecdote Message-ID: <1990May6.204122.3793@diku.dk> Date: 6 May 90 20:41:22 GMT References: <43777@iuvax.cs.indiana.edu> <1161.263f4987@gp.govt.nz> Sender: news@diku.dk (The Netnews System) Organization: Department Of Computer Science, University Of Copenhagen Lines: 16 bzs@world.std.com (Barry Shein) writes: >The way I heard it at the time was that a group in England had sped up >a 750 somewhat (I doubt it was 4X, more like 1.25X) by removing noops >from the microcode (and, of course, the apocryphal "if you do this DEC >won't service the machine anymore", probably true.) We also looked into it at the time, but we didn't buy one. As I heard it, it involved a card which would sit on the CPU backplane (probably replacing the clock module) and watch the microcode words going by. Normally, the various clock subcycles are long enough for the longest delay path, but this card would detect cases where only short paths were needed by the microcode, and shorten the subcycles accordingly. -- Lars Mathiesen, DIKU, U of Copenhagen, Denmark [uunet!]mcvax!diku!thorinn Institute of Datalogy -- we're scientists, not engineers. thorinn@diku.dk