Path: utzoo!attcan!uunet!know!zaphod.mps.ohio-state.edu!usc!ucsd!ucbvax!mtxinu!unisoft!hoptoad!fidogate!f555.n161.z1.FIDONET.ORG!Sonny.Shrivastava From: Sonny.Shrivastava@f555.n161.z1.FIDONET.ORG (Sonny Shrivastava) Newsgroups: comp.sys.mac.misc Subject: Re: Surely A Iifx Blows An Amiga 3 Message-ID: <7877.271B9F2D@fidogate.FIDONET.ORG> Date: 13 Oct 90 09:49:14 GMT Sender: ufgate@fidogate.FIDONET.ORG (newsout1.26) Organization: FidoNet node 1:161/555 - MacCircles, Pleasanton CA Lines: 36 Ok, I got the results of the Speedometer tests. Unfortunately, the latest version of Speedometer that I have is version 2.0. I will post the results I got with that. The figures you quoted seem way off to me. Som of them make sense, but others are totally ridiculous, like the math results. Speedometer 2.0 compares its results to a stock Mac SE with a 20 MB hard disk. Here are those comparison numbers: CPU: 6.62 x speed of Mac SE Math: 13.97 Disk: 5.05 Overall: 7.78 Now... Kwhetstones/sec: 108.108 Dhrystones/sec: 5008.347 Sieve (Sec. for 10 times): .583 Savage Cum. Error: 2.29745e-0111 Savage Time: 22.950 Savage Iterations: 5000 Unfortunately, some of these results are incredibly low. That is because Speedometer 2.0 was not compiled and optimized for the 68882 FPU. In the documentation of 2.0, I found that the author wrote a version (which I don't have, natch) that is optimized for and takes full advantage of the 68881. This is slower than a 68882, mind you. With that, he got over 600 Kwhetstones/second, and the above "Math: 13.97" jumped to 118!! I will try to get my hands on a copy of Speedometer 2.5, so we can do an accurate comparison. I provided the above results because I had them, but don't use them or quote them, because they aren't accurate or representative. When I get 2.5, I'll post a message to you with the results. -- Sonny Shrivastava - via FidoNet node 1:125/777 UUCP: ...!uunet!hoptoad!fidogate!161!555!Sonny.Shrivastava INTERNET: Sonny.Shrivastava@f555.n161.z1.FIDONET.ORG