Core i7 and VDimm

[chiLLZ]

Anyone using more then 1.65v VDimm with Core i7?

I want to overclock my memory more and maybe use 1.7 to 1.75v Vdimm but of course there are warnings that this could kill the i7 memory controller. Just wondering if its possible to do this successfully and if any one has done this and had their system stable over a period of time?

I did run 1.78v Vdimm briefly but not too sure how risky it is... Also after doing some searches on other forums some people say they have done this but its not clear for how long, maybe the question to ask is, has anyone actually killed an i7 doing this?

 

[Bomag]

I had a play round with my memory and found that

A increasing the speed pat 1600mhz had no benefit except in benchmarks. Even encoding HD material it had no benefit.

B. Loading the mem controller by upping the speed lowers the over clocking ability of the CPU (for me by 100mhz).

Since they started putting large stickers on MBs at the very last minute probably means that chances are high of killing the chip.

 

[chiLLZ]

In that case, should I rather be thinking of tightening the memory timings instead of more overclocking? I can run this memory at 1300 6-6-6-18. ddr 3 timings seem pretty loose.

also, is there a program that can benchmark memory by itself? 3d mark tests the whole system, I am just interested to test different memory speeds and timings with the ability to see direct results with a memory benchmarking proggy

i7 tripple chanel has so much bandwidth maybe it would be better to sacrifice a little for tighter timings to reduce latency.

thanks for your thoughts on this

 

[Meridian]

Sandra and its memory bandwidth benchmark is the usual tool. SuperPI is also very memory speed dependant. But as Bomag has correctly pointed out, it's just willy-waving: it will have no real-world effect.


M