post #11 of 11
I posted a htread on this a while ago:
Originally Posted by ericeod View Post

I've been surprised to say the least to see so many posts about i7 (1366) running tri channel vs dual channel.

So just to help clear up this issue, here are some articles discussing the integrated memory controller, latency and ram speed, and of course, dual channel vs tri channel:

And what about memory speed and latency? Well, in real world applications there is little advantage:

Tech Report: Exploring the impact of memory speed on Core i7 performance, A look at channels, bus speeds, and latencies
by Geoff Gasior, December 1, 2008


Although three channels of DDR3 memory might seem excessive, the Core i7 really does seem to make good use of faster memory, at least in synthetic tests. But that was to be expected. The real question is whether those gains translate to real-world applications, and it's here that the results are more mixed. Certainly, in scientific computing tests like Euler3d and MyriMatch, which we already know to be sensitive to memory subsystem performance, faster memory can provide tangible performance perks. However, common desktop applications like those highlighted by WorldBench don't benefit much from higher memory clocks or tighter timings. Neither do most games, which at best show minor frame rate improvements that aren't significant enough for most folks to even notice, let alone appreciate.

Digital Daily: CPU Intel Core i7-920 (Bloomfield)
Aleksandr Mitrofanov, Dec 02, 2008


This controller is optimized for operation with DDR3 memory and supports 1-, 2-, and 3-channel access modes. In particular, in using the 3-channel access the bandwidth of DDR3-1066 is 25.6 GB/s, which meets the bandwidth of the QPI bus. Therefore, for the Socket LGA 1366 platform the faster memory is not yet needed for now.

Yes the 1366 CPU does support dual and tri channel via the QuickPath Interface (QPI) link and the on-board memory controller. And yes, you can run dual channel instead of tri channel with minimal impact.

And how does dual channel perform against tri channel? Here is a quick read:

Inside HW: Intel Core i7: Dual-Channel vs. Triple-Channel Memory Mode
Djordje Kovacevic, 20 March 2009


First test that we run was Everest and we were really shocked with results. It was clear that memory access was faster in case of dual channel memory mode while bandwidth had very close values. After Everest we were slightly less shocked with results that provided WinRAR test that is very sensitive to memory subsystem performances. It showed only 2-3% difference in performances in favor of dual channel mode. Better performance of dual channel mode showed also DivX test where our test showed 3% performance difference in favor of dual channel mode also. Then we started Blender that showed minor differences between system operating in dual and triple channel memory mode. Nuclear MC also showed minor differences between two memory modes tested here, while 3DMark06 showed almost identical results. We all know that Crysis can react positively in case of better memory performances so results weren’t that much of a surprise. World in Conflict also showed differences in performance levels of both memory modes. Both gaming tests favored system that operated in triple channel mode.

I hope this clears up all the misconception on OCN.