Originally posted by: ViRGE
It's like talking to a brick, I swear. To realize a performance increase from a move from x86-32 to x86-64, either the application would have to be extremely bottlenecked by 64bit integer math, or highly register starved. There simply aren't any applications of the former, and the latter hasn't been a problem in ages due to well designed compilers. 64bit integers are extremely rare, so the whole thing is generally a break-even proposition; whatever performance improvement from the additional registers is usually offset by the performance penalty from longer instructions. You'll never find those kinds of performance improvements, so we don't even talk about them.Originally posted by: Aberforth
*stuff*
to me, it's like talking to a bunch of deluded Neanderthals, honestly. According to you, it's like -if it takes 10 seconds process 8 instructions it will probably take 20 seconds to process 16 instructions in 64bit, oh yea...sure, congrats.