Reviewed by gpscientist from Tracy, CA on 12/28/12
Several years ago I upgraded to an i7, 8 threaded an electromagnetic code, & got a speed-up of 8. I was hoping to get a speed-up of something close to 4 more with the AMD FX-8350 (Black Ed)+Asus CROSSHAIR V FORMULA-Z combo using the new VEX vector operations that do 4 floating ops/command. When I turned on VEX in the compiler I got a 5% speed-up. I studied the resulting assembler code, made changes, & now have a 15% speed-up. If I figure out how to make a signigicant increase in speed, I'll update this posting (if I can).
I haven't done assembly coding for a long time, so I might be overlooking something.