If running DDR2 at a slower FSB, can timings typically be tightened ?

[brjoon1021]

HI,

I have an ASRock motherboard that does not allow for much of a fast FSB. I have overclocked an e2160 to 290x9 and that is as fast as she goes. I have some Kingston ValueRAM, KVR667D2 1 GB sticks running dual channel which is able to run on a FSB of 333 at 5-5-5-?.

Well, since I can't go up on the FSB any more I was wondering if tightening the timings could get me a little more speed and also what might be a reasonable first attempt at timings for this RAM running at 290 FSB.

4-4-4-12 or could I even try 3-3-3-9 ? I am still used to the DDR days and trying to get 2-2-2-6! These numbers seem so slow to me.

So to recap, the 667 RAM is loafing at 580, so can I possibly tighten the timings ?

 

[n7]

If you can't use a higher ratio (there's nothing higher than 1:1 on that board?!), then sure, tighten them up.

I'd say 3-3-3 should be possible, but try 4-4-4 first, then 4-4-3, etc.

 

[brjoon1021]

Hi,

Yep, I am pretty sure that I can change the ratio from 1:1 to something else, but the overclocking guide in this forum says not to if I understand it correctly:

found here: http://forums.anandtech.com/me...=2057083&enterthread=y

"DRAM Frequency - This the speed your RAM will run. For now set it for double the FSB you plan to use to overclock (for example if you're running a 333 MHz FSB, set it for 667; if you're running a 290 MHz FSB, set it to 580, etc.) Remember the RAM is DDR2 so divide that 667 by 2 to get 333 or the ratio of the CPU-to-DRAM is 333:333 or 1:1.

Most good boards will offer 5 modes listed below. Assuming that you're using a 333 MHz FSB the ratios are:
FSB : DRAM
1:1 = 333 MHz : 667 MHz
4:5 = 333 MHz : 833 MHz
2:3 = 333 MHz : 1,000 MHz
3:5 = 333 MHz : 1,111 MHz
1:2 = 333 MHz : 1,333 MHz


Running in 1:1 mode is termed, "synchronous mode." If you use a higher frequency, you're running is so-called "asynchronous mode" which offers marginal speed advantages at the price of more heat and power consumption on a C2D/C2D Quad-based system for most users. At the end of this article, I linked small study I did that talks about the different memory dividers and shows experimentally why running 1:1 (synchronous) is really best on a C2Q/C2D machine; in my experiments, I only saw a 2-3 % boost running in a faster asynchronous mode. Doing so however required more NB vcore and ICH vcore which meant more heat.

 

[n7]

Yeah, i don't agree with that

Not saying his guide isn't good; it's superb.

But if you know what you're doing, & you're using a motherboard that actually allows for controlling more than just a few of the main timings, running a higher speed will be faster.

I'm not saying hugely so, but somewhat certainly.

P5B-D is a poor mobo for doing accurate comparisons (i own one), since it automatically changes important subtimings like tRD depending on different ratios, etc, which makes getting truely accurate results pretty much impossible.

Obviously, it'll be the same story on your AsRock, but in general, on pretty much all Intel-chipset mobos, you'll end up with better performance running RAM at a faster speed & looser timings than lower speed like 1:1 with tight timings.

Again, not saying it's a huge difference; it's isn't.

But it's there.

And you rarely have to increase vdimm or NB voltage all that much, if at all, unless you're going for insanely faster results.

My suggestion is to run your own benches to compare.

And then either tweak for higher speeds, or 1:1 w/ tighter timings

 

[myocardia]

Originally posted by: brjoon1021
Yep, I am pretty sure that I can change the ratio from 1:1 to something else, but the overclocking guide in this forum says not to if I understand it correctly:

"DRAM Frequency - This the speed your RAM will run. For now set it for double the FSB you plan to use to overclock (for example if you're running a 333 MHz FSB, set it for 667; if you're running a 290 MHz FSB, set it to 580, etc.) Remember the RAM is DDR2 so divide that 667 by 2 to get 333 or the ratio of the CPU-to-DRAM is 333:333 or 1:1.

You didn't understand correctly. Reread the bolded portion.

 

[DSF]

You could probably bring the timings down a little bit, but if you're expecting a noticeable speed increase you'll probably be disappointed.

 

[Mondoman]

Just be sure to run memtest86+ and Orthos (or equivalent) in "blend" mode to make sure you aren't getting memory errors.