my e6400 is at 3ghz, possible to get higher?

Associate
Joined
12 Oct 2010
Posts
1,825
Location
Guildford
Ive got my e6400 running pretty damn cool at 3ghz, got an artic cooler etc, im worried that 3.2ghz might be pushing it, do i need to raise voltages or anything, or could that be achived on stock volts? it usually runs in the 20s-30s on idle right now, were these known to be able to hit 3.2? Want to get the most out of my aging c2d before i replace it with a new mobo/cpu next year some time.

thanks
 
Mine does 3.2 Ghz 400x8. Ive raised the voltage though because the bios kept defaulting back to 2.13 GHz for some reason every so often.

EDIT: duh! yes I do mean 400x8!
 
Last edited:
^^ I think you mean 400x8.

To op: the early 65nm e6000 conroe chips are nearly indestructable. 1.4/1.5v+ is common with good cooling so by all mean bump up the voltage a bit and see how far that will take you, just keep an eye on the temp. I think 3.2-3.4ghz is common for e6300/e6400. More if you're lucky but the fsb becomes a limit because of low multi on these chips.
 
^^ I think you mean 400x8.

To op: the early 65nm e6000 conroe chips are nearly indestructable. 1.4/1.5v+ is common with good cooling so by all mean bump up the voltage a bit and see how far that will take you, just keep an eye on the temp. I think 3.2-3.4ghz is common for e6300/e6400. More if you're lucky but the fsb becomes a limit because of low multi on these chips.

thanks very much, think ill go for 3.2/3.4, what vcore should i set it at, ive never done it with raised voltages before.

edit: got it at 1.35 and 3.2 ghz :) will try stressing it now, still 30c idle so think it should be fine.
 
Last edited:
pretty pleased with this, even when the cpu got to full load it was under 50c, im thinking i may be able to push it to 3.4ghz with a bit more of a vcore increase.



Uploaded with ImageShack.us


edit: running orthos right now with 1.4 vcore and 3.4ghz, temps at full load are hitting 60c which is a bit worrying, or is this ok?

seems stable enough though.
 
Last edited:
Nice.

60c is fine, if you really want to see how hot it gets load up IBT (Intel burn test). The temp is usually 10c hotter than prime/orthos. For an indication of stability put it under 5 runs with max mem usage.

For now don't increase any more voltage but rather gradually increase fsb until it's unstable then up the voltage one more notch. Also if you have ddr2 pc6400 ram beware that above fsb400 it will be running above its stock speed so if it gets unsatable keep in mind that it might be the ram. Although most ddr2 800 kits should be ok running slightly above stock.

Think about it this way that's an impressive 1.3 ghz oc already ;)
 
Tom 91x91

Any chance you can send me your settings...I am an OC newbie and struggling to get my E6420 on a Gigabyte GA-650SLI-DS4 with PC-6400 DDR2 to OC past 2.4ghz. This is on the FSB:RAM linked setting. Everything else seems to fail when I go manual.

I am sure it is all to do with FSB:RAM ratio

Any settings you used would be so helpful.
 
Nice.

60c is fine, if you really want to see how hot it gets load up IBT (Intel burn test). The temp is usually 10c hotter than prime/orthos. For an indication of stability put it under 5 runs with max mem usage.

Use the 'Custom' Stress level option instead of 'Maximum' and input the Free physical memory (ideally slightly lower) for more accurate testing which is shown in windows task manager. That way your GFlops during stress testing should be higher and it will stress the cpu more.

'Maximum' stress level makes use of both physical and virtual memory and this doesn't stress your cpu as much. Here is the article from which I got the info above:
http://www.overclock.net/intel-cpus/645392-how-run-linpack-stress-test-linx.html
:)
 
Last edited:
Very interesting read, thanks for the link. I've always just used the maximum option as it won't take any value above available ram anyway (I'm using the v2.4 tool by agentgod). When I want it to use more I just start and stop the test several times as it releases more free mem space each time you stop. I always make sure the physical ram usage in task manager is 3.8/3.9Gb+ and the Gflop value I get is always 51/52, which seems right considering that article puts the estimated Q9550 4Ghz value at 54-56.

I did notice the test would take forever to run a single pass when using less ram with the temp no where near max I just didn't know why.
 
Very interesting read, thanks for the link. I've always just used the maximum option as it won't take any value above available ram anyway (I'm using the v2.4 tool by agentgod). When I want it to use more I just start and stop the test several times as it releases more free mem space each time you stop. I always make sure the physical ram usage in task manager is 3.8/3.9Gb+ and the Gflop value I get is always 51/52, which seems right considering that article puts the estimated Q9550 4Ghz value at 54-56.

I did notice the test would take forever to run a single pass when using less ram with the temp no where near max I just didn't know why.

You seem to be getting right values. Mine is usually 6-8GFlops lower than expected even though I turn off anti-virus momentarily along with EIST and C1E which seem to affect the Gflops values heavily.

The amount of physical ram shouldn't really affect the GFlops values very much as stated in the article and which I also confirmed by way of testing. Normally with low physical ram values, each run or iteration should be completed much quicker.

The amount of physical ram is directly proportional to the time it takes for a iteration to complete and not the GFlops (speed). So higher the ram selected, the longer it will take for iteration to complete and the GFlops (speed) is a constant more or less.

But in your case it strange that for lower values it takes forever:confused:.
 
Sorry it doesn't happen all the time, just sometimes it's not running "max out" - CPU still at 100% but temp is low, it would take forever and the Gflop value would be way less than normal. Just like the symptoms the article describes. But it only happens for me when less than max free ram is used.
 
Back
Top Bottom