Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!rutgers!bellcore!texbell!uhnix1!nuchat!abbadon From: abbadon@nuchat.UUCP (David Neal) Newsgroups: comp.sys.amiga Subject: CMI Processor Accelerator Summary: Not unless you have $200 to spare Message-ID: <2684@nuchat.UUCP> Date: 19 Dec 88 15:44:57 GMT References: <6018@louie.udel.EDU> Reply-To: abbadon@nuchat.UUCP (David Neal) Organization: Houston Public Access Lines: 54 In article <6018@louie.udel.EDU> CHUCHEM%BGUEE.BITNET@cunyvm.cuny.edu writes: >> From: Julius A Cisek >> Message-ID: <1391@leah.Albany.Edu> >> >> [Julius asks about the CMI Accelerator] > > [Dave Chuchem replies] >I have had the CMI Accelerator for a couple of months. It is installed in a >standard A1000 with a Microbotics Starboard 2Meg add-on. > > [Text deleted] >CMI did not supply any advice or software to demonstrate any gains in >using their system. They did not supply any test programs at all. > >I could not notice any appreciable speed improvements in any casual >applications. > I too bought the CMI, but they did pack their disk full of handy stuff, including a mandelbrot program that uses the 68881 and some programs the graphically illustrate the speed differences. >Running some standard benchmark programs I found that: >1. "sieve" benchmark ran in 64 seconds instead of 68.9 seconds. >2. "savage" benchmark ran in 59.8 seconds instead of 75.9 seconds. > >Interesting, but nothing to get excited about! > On the 500, it's not night and day, but thing did (see below) obviously improve. > [Flames about the 68881 not helping any deleted] > >Dave Chuchem >chuchem@bguee.bitnet I don't have a 68881 and even if I did, it wouldn't help; I had my PA about 1 week when I shipped it back (somewhere in the second week of November I think) and haven't seen it since. This is no reflection on CMI, per se, as they blame an independant firm (since fired) for all the delays. Perhaps they do have someone else building their boards now, but these guys won't even commit to my having a board before year end. CMI's work, and they do help, but throw it away if it breaks. David Neal {killer!texbell! or uunet!}nuchat!abbadon The opinions expressed are strictly my own.