Exhaust{ive,ed}
September 15th, 2005
Programs tested: 24
Tests failed: 0/144 (0%)
A few hours ago I finally got both of my register allocators generating correct code. There are still inefficiencies, but it’s far too late to add further optimizations (to the allocation scheme, not to the instruction stream itself!) without risking the stability of the code. So, a few comments and it’s done. (Now for the 5–10 page lab report—oof.)