Article on nForce2 timings
I found this article today.
And it seems that the memory timing on the nForce2 chipset is a bit puzzling.
It looks like the chipset has a sweetspot where the performance is high.
Look at these excerpts.
Here's the link
Article
Anyone having similar experiences?
And it seems that the memory timing on the nForce2 chipset is a bit puzzling.
It looks like the chipset has a sweetspot where the performance is high.
Look at these excerpts.
As you can see, there is a consistent trend in the bandwidth. Seems that when the tRAS latency hits 10 or above, there is a bandwidth spike. Yet when it hits a tRAS value of 12, bandwidth drops lower than when at 5. I'm not sure why this happens, but all my tests came back with very consistent results....lets see if they continue to keep this consistent.
I am convinced. I doubted it at first but after running this many tests (176 tests in total!) and producing similar, consistent results I am now running a latency spec of 2-2-2-11. I found that I score higher on 3DMark2001SE, and what further more amazes me, is that with these latencies and this RAM, I can achieve a maximum FSB of 219. It passes Memtest86 running for 12 hours+, and I was unable to get this high of FSB with timings such as 2-2-2-5 or 2-2-2-6. So in my eyes, it even gave me an edge on stability.
Here's the link
Article
Anyone having similar experiences?
0
Comments
I've been running my timings @ 2-2-2-11 for almost a month now since I read about it at the abit-usa forums here
this thread was begun by tha_noname and is very informative read for anyone who has the time!
thanks again
csimon
http://www.short-media.com/forum/showthread.php?s=&threadid=1787&highlight=timings+help
I went from 2-2-2-6 to 2-2-2-11 and my Sandra score was unchanged.