CompuServe Messages

#161583

    24-Apr-95 20:31:50
Sb: #161583
Fm: Chris Braid 72530,1202
To: Charles H. Klimushyn 74064,1466
Chuck: >So I was a little concerned if the 601 Turbo Boards would have similar >slowdowns due to the older architecture of our machines. Keep me posted on >your eventual results. The eventual results are: The Radius 24XK arrived, installed w/o any problems, and runs fine in 601 and 68x mode. I downloaded Speedometer 4.0 the other day and have been having fun inerpreting benchmarks etc. Overall it seems the Turbo 601 puts the IIci just about on par with a PPC 7100/66, except in video where it scores about 57% of a 7100's video speed (that's using the built-in video). I've also conpared it to a couple of (ugh) DOS machines, the first being a standard-issue 486/33 PC clone, and a 66 MHz Pentium. REsults as follows: CPU Dhrystones (integer) Whetstones (floating point) IPC 486/33 15172 5683 Pentium/66 68275 21083 PowerMac IIci 63796 32297 ..So the 601 is about where you'd expect it to be. A bit slower on integer and a full third faster on floating point. And it even gets the right answer when it does long division. In real-world terms, the two major applications for which I have native code binaries are Photoshop 3.0, and KPT Bryce. The 601 makes a huge difference in how you use these apps. In photoshop you get instant visual feedback as you set up the contrls for Unsharp Mask. This makes a major difference in usability. Bryce completes the same render in an hour that the IIci worked on for most of a day; I'm sure it's faster by a factor of twenty, though I haven't actually timed it. Here are some Video benchmark results: PPC 7100/66 100% Quadra 630 86% PowerMac IIci 57% Stock IIci 31% I suspect that the bottleneck is the video circuitry, and that a faster processor wouldn't make much difference. If you're interested I'll try to organize my results a bit better and post them here. Regards – Chris