GOM3RPLY3R :
hcl123 :
60-70% ? doesn't compute. Try it with a radeon, and it can be better for sure.
Nvidia on an AMD platform... ummm doesn't spell anything good, if not intentionally cripple, doubt nvidias does any tiny bit of *driver* optimization for its archirival ... sometimes what we can't see is what delude us.
Why SuperPI (kind of useless in any case) have disappeared from reviews !?
http://www.xtremesystems.org/forums/showthread.php?286448-The-Book-of-Bulldozer-Revelations-Episode-2-%28SuperPI-x87%29
The software / hardware interactions have more than meets the eye, or the benchmark lol (imagine a BD winning lol ->see that thread)
What you just said is complete bogus. There are
many people who use AMD CPUs and Nvidia GPUs, I have no idea what you are talking about. Please give me your reasoning as to why you say its false, and then tell my why you are trying to avoid the fault I pointed out.
I would like to note for you: Taking a poll of people who like baseball, and giving the poll to baseball fans, and that's it, wont show how many people in the world like that sport, its just a failed science experiment.
You need to test many people, not just one. If you took middle school / high school science classes you may understand.
Its the drivers dude. You go nvidia, you have to install their drivers right ?
You don't think that nvidia would use XOP/FMA4 in their drivers when on intel it would have to run by microcode(painful slow), do you ? .. There you have it, its only an example, and there are gazillion of issues for optimization, different from CPU to CPU, in any driver layer.
But its typical, ppl haven't yet realized, that many of this bickering and angry posts about hardware, what ppl in fact are bickering about are SOFTWARE differences in high percentage, much included all benchmarks. And you can't escape it, not only hardware without software is paper weight, its not really possible to completely isolate them.
So this should all be fun, knowing that ERROR MARGINS can even go quite above 10% sometimes, there are positive points for all tastes, its an approximation of potentialities, there is nothing
absolute on those results
(understand relative (a clew) vs absolute ? )