alverson@decwrl.DEC.COM (Robert Alverson) (09/22/86)
Well, I finally sat down and looked hard at how Tomas Rokicki's life program works. After trying many bizarre schemes to try to make it faster, the best I could do was a minor optimization to save one memory reference. So now it takes 10 blits for a total of 37 operations. If someone can make it even faster, I'd like to know, too (certainly tell Tom). Rather than posting the changes now, I'll probably let the next version of blitter-life include them. BTW, I wonder if anyone has a CAD program that attempts to do logic decomposition. A two-level (PLA style) optimizer is useful for evaluating hand attempts at decomposition, but I'm curious if the computer can do better. Bob soon to be path: glacier!sonoma!rla