Tinkering with my E2160 - a fewq Qs

NozzaC

NozzaC

NozzaC

Associate
Joined
14 Apr 2008
Posts
16
My OC has worked out fine but wondering about pushing it a bit further and looking for some advice.

It's an E2160 on a P35-DS3L with 800mhz ram.

I've set it up as 9x333, core voltage 1.29375V, RAM voltage at 2.1V (as rated by the manufacturer) and the SPD is set so that the RAM speed shows up equal to the rated speed 800 this is a multiplier of 2.4 (not entirely sure what this means to be honest). I manually set the RAM timings to the ratings from the mrfr (4-4-4-15) and set the MCH read delay (tRD) to 7 (was 11). Idle core temps 32C, northbridge the same. 100% load temps are 60C. All tested with a Blend test on Orthos and is stable.

I'm on stock Intel cooling. Temps were hotter but my new case has helped a lot.

My Q's are:

1. Should I push this anymore with a stock cooler?
2. If I up the FSB speed then the RAM speed shows up higher than 800 - should I accept this or lower the multiplier, which at all but a few speeds won't match 800 but will be lower. How important is this?
3. What would the effects of lowering the step to 8 whilst upping the FSB a lot more be?
4. If it needed to loosen the RAM timings to say 5-5-5-x - is that going to make much difference to the performance?
5. How much heat can the northbridge take?

Many thanks
 
I own an E2160, so maybe i can help ya out :).

1. i don't rate the stock cooler for doing well at anything higher than 3ghz. Get a better cooler - the Arctic Cooling Freezer 7 Pro does a wonderful job, IMO, but there are better coolers out there.

2.I'm not so sure about ram timings, tbh. didn't mess around with them on my OC, but i think yo ought to be ok OCing your ram a bit.

3. I'm no so sure, tbh. Some find it easier to do so, but i prefer to actually use the higher multiplier.

4. It shouldn't make any drastic difference, i think. But if your timings are stable at your OC, why not use that to your advantage?

5. No idea. if you're worried, get a small fan or something, but it ought to be ok. After all, it's not a massive OC you're doing.

Also, visit the E2140/2160 OCing thread in the CPU area, there's loads of info there that you can use. Welcome to the forums by the way :)
 
If your e2160 is M0 stepping (use CPUz to check) then it's max safe temp is 73ºC so u have a bit to go yet so i'd say try 3.1GHz and see how u do although u will see much reduced temps using an aftermarket cooler.

As to the FSB and RAM timings, it depends what RAM u have, if it is good RAM it will OC well. For instance, I have the GeIL ULL 800MHz ram and am running it at 960MHz without fault but had to slacken the timings from 4-4-4-12 to 5-5-5-15 to be able to achieve this. If u have good RAM for OCing then I would suggest using the 8x multi on the CPU and a higher FSB (the DS3 can take it) to achieve the same 3.0GHz overclock this would be an FSB of 375 and then if this is stable try using a multi on the ram of x2.4 - this would give u 900MHz on the RAM. If this causes problems then go back to the higher multi and keep ** RAM as close to 800MHz as u can get it. It's just a matter of fiddling around really and then running some benchmarks to see if the performance is better with higher MHz or tighter timings.
 
My RAM is Corsair 2GB DDR2 XMS2 DHX PC6400 (800) TwinX (2x1GB) CAS 4-4-4-12. It's supposed to be decent. It says 2.1v on the sticks.

What is the effect of underclocking the RAM, which seems to be indicated in various guides (they say to ensure the number produced by the SPD setting is lower than the rated RAM speed)?

Which benchmarking tools would be suitable to compare faster speeds+loose timings vs slower speeds+tight timings?

Could you possibly explain how those RAM divider numbers (2.0, 2.4 etc) relate to timing ratios I've read about?
 
That RAM should comfortably do 900MHz and maybe even still at 4-4-4-12. Check corsairs website for max/min voltages and up it if the system becomes unstable when OCing the RAM.

Underclocking the RAM would only be relevant when trying to find the full stable speed of the CPU, once found then turn to the RAM. Benchmarking I used was PCMark 05 but i am sure someone here can come up with a quicker more suitable test for RAM benching? I got bored and started playing games before i compiled any significant results :)

I dont know how the 2.0, 2.5, 3.0 convert to the rations 1:1, 3:2 etc but to find speed of the ram use the RAM divider number and multipy it by the FSB, that will give u the rated speed and use CPUz to verify. For stability testing i used the CPU and RAM stress test in Orthos at an OC i knew the CPU was stable at.

Hope someone else can give u some more specific advice.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom