Help! Duo Cores & Quad Cores: Memory Bus Hogs??

izanaki

Distinguished
Nov 8, 2006
8
0
18,510
Would it not be ideal for each core to have its own memory bus? Or is the current memory bus fast enough for quad core and dual cores?

I mean if you're rendering a video in HIGH DEF, Listening to MP3s, Burning a CD and surfing the internet... would the memory bus be so backupped by multiple cores running all that?

P.S. I'm not really tech savvy with this stuff in great detail but i would appreciate a better understanding of the memory bus and multiple cores.
 

AODBOB

Distinguished
Aug 10, 2006
73
0
18,630
If what you are talking about is the fact that four cores share 1333Mhz FSB, which means each core, if all four are all needing the FSB at the same time, would only get ~333Mhz then I was also wondering about that.

If anyone knows anything either general or specific about that please explain, because it seems like a huge, obvious pit for the Kentsfield. The benchmarks seem to say otherwise, but it still bothers me.
 

WR

Distinguished
Jul 18, 2006
603
0
18,980
So far there is no evidence of quad cores being noticeably hampered by FSB or memory bandwidth in ordinary user applications. Testing is commonly done with a top-range dual core chip using half the normal bus speed and RAM but double the stock multiplier. However, such testing has not been done for many professional and enterprise applications.

Listening to MP3's, burning CDs, and surfing the web all require very little CPU power and memory bandwidth. Encoding videos takes primarily CPU power, not memory, so if you could get 3 of the Kentsfield cores to encode video while you multitasked as aforementioned, it is not likely that your memory would be breaking a sweat - especially with the excellent caching system found in Conroe.