Hi David and all, Concerning FSB saturation with Quadcore, I read the following two days ago in the December issue of Maximum PC:
"Intel claims that analysis of the bus traffic under heavy loads using today's applications indicates that we're not even close to saturating the FSB.... Intel engineers do admit that it's possible to saturate the bus, but that you'd have to write code specifically to do that." Thus they decided to forgo a 1,333MHz bus for this generation. Now, how a CPU intensive apps like RS relate to "today's applications" is another question. On another note: Windows XP will obviously work fine with Quadcore because it works with two sockets, and isn't concerned with the number of cores per socket. The real super power will be when Intel releases the Yorkfield (quadcore) architecture at 45nm, a monolithic die, and puts two of those on a chip for an 8 core CPU! But that is likely 18 months out:( Regards, Brandon >> Yes , if that's not enough to make you drool ( 4 >> quad core CPU's ) then nothing is . Finally we >> have what Amiga developers wanted for all of us >> 10 years ago ... pure drop-dead power ! > >Well, I'm curious as to how much benefit quad-core brings when half the >problems with processing lie with bandwidth and getting hold of data to >process. Four cores sharing 10 GB/s or whatever the main RAM is might not >perform anywhere near as well as four separate processors on four separate >RAM pools. In any complex scene, the amount of data accessed per frame will >run into many GB and with lots of cache thrashing, and the processors will >likely spend a fair bit of their time waiting to get object data to work on. > >Well, the same problem applies to single core too, so I guess with four >cores you could get 4x the speed! > >David Coombes >[EMAIL PROTECTED] > >... >
