What is memory node interleaving?

[Spudwig]

Hello,
I have heard of "memory interleaving" which I believe is where the interleaving occurs between memory modules on different channels attached to the same memory controller.
Is "memory node interleaving" the same thing, or does it require two processors and two memory controller with a NUMA configuration?
The reason I ask is that I am not going to buy two processors (as my application is numerically intensive but not parallelised), but I still want to know if I can get the performance advantage of memory interleaving?

 

[Noworldorder]

You don't have to get two CPUs; The IMC is still going to interleave even single channel memory - it just does it in larger bursts, so you'll still get the bandwidth advantage. If I remember correctly, you cannot disable NUMA if two slots are populated - you won't POST if you try to.
Since the memory controller is off the chipset now, NUMA is set to enable with on-die IMC - I don't think there is a way around it.
You can check with the guys at Storageforum though, they would know better than we would.