08-20-2006, 11:47 AM
Originally Posted by Marl Huss
There are two cores in each CPU, and 2 CPUs. Each CPU has 4MB L2 chache. Therefore two cores share the L2 Cache. Each CPU has its own FSB. That means there are two FSB.
That means minimal bottle necks, and an average (real life) bandwith of up to 8GB/S..
If you want more information on the technical side then go to:
and for more technical info
Thank you so much for the info. You saved me from making a big mistake.
The MacPro is a serous peice of junk. I dont say that lightly. Its not Apple fault either.
first off all. 8GB/S is the theoretical bandwidth of the Mac pro. its real world bandwidth is a shocking 4.292GBs.
(A pentium 4 can do better, No wonder Apple dont show the memory bandwidth of the Mac pro).
Please read just this page HERE
A core 2 duo will whip it.
I have a 3 year old Opteron set up it has a theoretical bandwidth of 12.8GB/s I can easily get 9500 GB/s in real world test.(I would get more if i got new Dual core opterons) This is where a seperate memory bus for each CPU and AMDs memory architecture just run rings around intels.
Seriously Im shocked. When will intel get its act together. Whats the use of have a blinding fast CPU if you dont have the memory bandwidth to utilise it properly? the latancy just sucks on xeons. A core 2 duo will give it a run in memory sensitive tasks and my three year old Opteron will obliterate it.
WOW that brought me down to earth. I was so exited about the thought of getting a new Mac Pro.
Guess I have to stick to a PC workstation. Shame on you intel.