From n00b to novice, overclocking my E2160

Inquisitor M · 27 Jul 2009 at 18:57

Ok, so after the palava about my poor little overheating E2160 a couple of weeks ago, I now have new ram and am getting on with some serious first time testing! Just wanted to post results so far and have some of you more experienced folk offer suggestions.

Intel Core 2 Duo E2160 @ 1.8GHz
Foxconn G33/31 motherboard (max 1333FSB/800DDR2)
Kingston HyperX 2x2Gb DDR8500+ @ 5-6-6-15 (motherboard can't register more than 15 tRAS)

Started at stock voltages for CPU and ram, dropped the ram divider to 667 and pushed up to 2.6GHz (1152FSB / 480x2 ram), past that it came up unstable.

Tried pushing up memory voltage to 2.2 as rated with no effect, dropped the FSB back down and bumped the ram divider to synchronous to test the same ram speed at lower CPU load. Passed 490MHz at 1.8v, got up to 548MHz at 2.0v, now pressing on at 2.2v to find it's maximum.

That much seems obvious, so next is dropping the divider back to 5:3 and pushing on with the CPU clocking by upping the Vcore, knowing the ram is well within it's limitations. That's where the questions start, since bumping the Vcore is where things get theoretically dicey. Obviously at first you bump FSB until it comes up unstable and then bump the Vcore by one grade to see if it becomes stable again.

Am I missing anything obvious in my proceedure?
How much should I expect to get out of the increased core voltage?
Is the return going to be a diminishing one? ( I assume so)
What is the maximum reasonable Vcore bump? (I've read 20% but wanted to ask again anyway for opinions)

Cheers in advance for any thoughts...

Scottage

pastymuncher · 27 Jul 2009 at 19:14

The cpu should be good for at least 3Ghz. I have had 3.6Ghz out of a E2140 and also a E2180. 1.45v is ok for a 65nm cpu. Have you tried increasing the NB voltage a tad?

Inquisitor M · 27 Jul 2009 at 19:29

Ahh yes, the other thing I ment to ask about but forgot...Northbridge is pretty much just RAM stability right? At this point I have reached a maximum 557MHz ram frequency at 2.2v, which far outstrips what I had acheived on an asyncronous divider, so I'm not too sure I'm worries about playing with the Northbridge voltage yet. I probably can't max out CPU and RAM at the same time, so as long as the ram keeps up witht he CPU on the asynchonous divider I'm happy. May try NBv tonight to see if I can get a few more cycles out of the ram for educational purposes.

-=Vect0r=- · 27 Jul 2009 at 22:09

These chips are known to suffer from FSB holes between 380-400 MHz. Also most tend to top out at around 420 MHz.

Leave the FSB

RAM ratio 1:1 for now to exclude RAM as a limiting factor. Try lowering the multi and experiment with higher FSB to check for FSB holes.

These 65nm chips can take a decent level of vcore and are safe up to 1.5v 24/7. Mine has been running at 1.6v for the last 6 months.

pastymuncher · 28 Jul 2009 at 13:34

Inquisitor M said:
Ahh yes, the other thing I ment to ask about but forgot...Northbridge is pretty much just RAM stability right? At this point I have reached a maximum 557MHz ram frequency at 2.2v, which far outstrips what I had acheived on an asyncronous divider, so I'm not too sure I'm worries about playing with the Northbridge voltage yet. I probably can't max out CPU and RAM at the same time, so as long as the ram keeps up witht he CPU on the asynchonous divider I'm happy. May try NBv tonight to see if I can get a few more cycles out of the ram for educational purposes.

NB voltage also allows the board to run a higher fsb.

Inquisitor M · 28 Jul 2009 at 20:42

Ahh, cheers for that, didn't realise.

Inquisitor M · 28 Jul 2009 at 21:28

Haven't seen any comments about how much boost is safe for the Northbridge, though I assume that's because they vary so much. Would be greatful if anyone had any insights they care to share though.

mrthingyx · 28 Jul 2009 at 22:44

Just try upping the NB voltage (as required) slowly and see what happens: each case is different.

Chances are it might not actually help, just as much as it might well help push your overclock that much further.

As the components in a computer vary so wildly, there is almost no reliable indicator of how a system will overclock. Trial and error (though methodical) is the important bit. Now get trying.

Inquisitor M · 29 Jul 2009 at 11:30

Couldn't have imagined that 0.06v on the NB would have gone from crashing within 2 mins unloaded to being stable at full load (IBT high load, 10 runs).

Testing continues...

Inquisitor M · 30 Jul 2009 at 12:16

Have now managed the 2.8GHz that seemed so far away before, and after resetting CPU and RAM voltages to stock it took 3 NBv bumps (0.1875v) to do it.

The E2160 isn't running any hotter now than it was at 2.3GHz, that chip is dynamite! Huge thanks to MrThingyX, It might have taken me ages to try that, if at all. I'll probably just see how far it goes on this voltage and leave it rather than bumping NB again, and just up the vDIMM a little to see what optimised latency's I can get out of this 1066 DDR2, since it's not even runing at it's stock speed right now.

Big thanks folks. Mission is declared a success