Post Your Overclock!

[tw33k]

Raising the HT Link to 2500 was a last resort. I've been pushing the RAM as far as I can today. Wish this was stable..

 

[Britton30]

If you lower the HT Link and NB Freq, then raise the RAM freq multiplier does that work and maintain your CPU overclock?
As Essenbe told me, overclockong is more art than science.

 

[tw33k]

Honestly..I've tried everything. I'm settling with this...



Not too bad at all

 

[Britton30]

Know how you feel. Oddly those setting are a bit better than your other snip with 7-8-7-21 timings.

 

[tw33k]

Dropping the command rate to 1T helped

 

[smarteyeball]

Just a quick and dirty IBT run @ 4.6 using +offset 0.020 - max vcore spike I saw was 1.320



Temps are a bit meh, but it is IBT and the ambient is ~28c, so not that unsurprising. And since it's only a 5 IBT run, there's noway I'm calling it stable, but at least it's showing promise.

 

[Britton30]

Dropping the command rate to 1T helped

I noticed it was changed. I changed mine to 1T and I would always have core 0 throttle to x4 and around 1GHz.

Just a quick and dirty IBT run @ 4.6 using +offset 0.020 - max vcore spike I saw was 1.320

View attachment 185474

Temps are a bit meh, but it is IBT and the ambient is ~28c, so not that unsurprising. And since it's only a 5 IBT run, there's noway I'm calling it stable, but at least it's showing promise.

On Maximum that'll be more stress than you can put on it in regular use, unlees you run folding@home or do NASA flight calcs.

 

[smarteyeball]

Aside from a 'quick' vcore/temp test, I don't really trust IBT and variants (ie LinX) all that much when it comes to stability. With previous chips I've had them @ 20-50 pass runs that breezed through the tests only to BSOD in certain games/apps etc.

5 runs is equivalent to a single pass of memtest for proper testing, comparatively speaking. (Not definitive )




But it's been 'Skyrim' stable so far, so that's a bonus. It's also a more interesting 'stress test' than P95

 

[Dave76]

Dropping the command rate to 1T helped

Command Rate of 1T is higher/tighter than 2T.

If your having any RAM stability issues change the Command Rate to 2T, it will help.
For high CPU frequencies you may need to underclock the RAM a little, loosen the timings or drop the frequency back to 667MHz.

 

[Wishmaster]

wishmaster, good job. Congratulations on the new build and good luck with the magical 5.0.

Well I managed 4.8Ghz with it.

I also managed x50 5.0 Ghz. It booted and loaded everything in Windows without issue..
But, it wasn't completely stable though.

Under Intel Burn Test, it would freeze and BSOD on the very first test.
It was probably stable enough to get a CPUz Validation but ... thats cheating IMO if not truly stable.


I was sitting at x50 Multi and 1.4volts, (1.39 in bios) so maybe a vCore of 1.44-1.45 would stabilize it enough to get a few IBT runs in at that speed.

But, Im not sure Im comfortable with that kind of Vcore especially for a 24/7 OC.
That and at 4.8, it was already around 77C, and that was at a 1.36 vCore.

So once it started crapping out at 1.4 vCore I just stopped right there, since that was my personal limit for a bench run.




At least I know its possibly capable of it. From reports Ive seen, most need something around 1.5v for any stability at x50.
Im thinking maybe this CPU would run 5.0 at around 1.44-1.46ish.


~~~ Oh, and these benches were run with all 4 cores/8threads enabled.

I am think without Hyperthrteading it would probably be a bit easier.

 

[jeepmann4x4]

www.sevenforums.com/overclocking-case-mods/2309-post-your-overclock-68.html#post1412235

Nice job Wishmaster, The above link is my older post in this Thread. I would think the i-2700's should be able to meet and beat the i-2600's

Link

 

[Wishmaster]

www.sevenforums.com/overclocking-case-mods/2309-post-your-overclock-68.html#post1412235

Nice job Wishmaster, The above link is my older post in this Thread. I would think the i-2700's should be able to meet and beat the i-2600's

Link

WOW!
Looks like you have a very good 2600. From what Ive read, not all them can do 5.0+ at all.



I did go back, and start looking for lowest voltages though, to see what I could by with for a 24/7 OC.
And while I have not run a proper 10hr+ Prime run, I did run a 5 pass Maximum stress IBT run.

4.4 Ghz stable at 1.23v
4.5 Ghz stable at 1.26v
from there...
4.6 Ghz jumps up to 1.3v
vcore needed start increasing rapidly from here ...
4.8 around 1.36-1.37ish was my last stable run

 

[Dave76]

Nice OC Wishmaster.

 

[StalkeR]

Good job Wishmaster!

 

[StalkeR]

I was playing a bit with LinX at 3AM....





So someting is not good here, the MB Temp... In Speccy it shows extremlly high tems(~65C) and in Speedfan/Aida is much lower...
So what temps are true. 60+ C is not good for the mobo.

That was quick tests(1st around 2/3min, and 2nd 5min). What do you think about the temps ? As I can see, CPU temp is also very high, about 60C.

 

[Wishmaster]

Yea, Linpack is vicious!!!

CPU temp under linpack load at 60C is good. In fact 70C and below is fine & to be expected for LinPack with a OC'd CPU.

just keep in mind, nothing in any real world heavy load will ever push the CPU that hard. Prime doesnt really push one that hard either.


-- But that MOBO temp Speccy reports ...

Im conviced its reporting the the CPU temp.

It may also be that particular temp is the Motherboards CPU socket sensor. Not neccesarilly the CPUs core temps.



I noticed on mine, evga eleet was showing each CPU core temp individually, and then the package CPU temp --
Which matches Speccy and my Motherboard CPU TEMP readout LED.


As you can see here, during Linpack the CPU Temp on the left is what speccy says is Motherboard for mine.
But that changes and always matches. As does the Package temps below each Core temp.

So thats what it has to be I think.

 

[smarteyeball]

Linpack is based on IBT, (same libraries) so they're pretty equal in regards to heat and punishment compared to P95


My speccy mobo has not budged off 60c regardless of ambient/load etc Buggered if I know what it's reporting; if anything :huh:

 

[jeepmann4x4]

www.sevenforums.com/overclocking-case-mods/2309-post-your-overclock-68.html#post1412235

Nice job Wishmaster, The above link is my older post in this Thread. I would think the i-2700's should be able to meet and beat the i-2600's

Link

WOW!
Looks like you have a very good 2600. From what Ive read, not all them can do 5.0+ at all.



I did go back, and start looking for lowest voltages though, to see what I could by with for a 24/7 OC.
And while I have not run a proper 10hr+ Prime run, I did run a 5 pass Maximum stress IBT run.

4.4 Ghz stable at 1.23v
4.5 Ghz stable at 1.26v
from there...
4.6 Ghz jumps up to 1.3v
vcore needed start increasing rapidly from here ...
4.8 around 1.36-1.37ish was my last stable run

Yeah It's a good chip! I was just playing with settings when I tried to boot at 5.2-5.5ghz, glad it booted at all of them I never even went back to try for stable because I dont need it or want it running that hard....but I also never tried for anything between 4.6-5.1ghz, I just put it at the 24/7 clock I liked ([email protected]).
Here are the 24/7 clock Prime95 and IBT pics.

 

[Wishmaster]

Thats chips a keeper

 

[smarteyeball]

Same old 4.6ghz now using -.030 offset and culling more AUTO voltage settings. So far it's netted me a nice 10c drop across the cores during testing. A nice 'summer overclock' if you will...




There's no doubt in my mind I could sail past 4.6; but I'm having more 'fun' lowering voltages rather than going for max clock speed.

For now any way