Intel Sandy Bridge 2600K and 2500K Processors @ [H]

Im fairly new to overclocking so if its crashing, does that mean I need to increase the voltage? And thanks for your response
 
Pingis said:
Im fairly new to overclocking so if its crashing, does that mean I need to increase the voltage? And thanks for your response
Click to expand...

Normally, yes. It could be other things, but increasing the voltage is the first thing to try.
 
phatbx133 said:
Mine bios template below.


Running 46 x 104.4 @ 4800, still testing and try change lower vcore etc.

no problem running at most games/desktop no crashes or bsod on me just yet.

ran prime95 for an hour, I stop it and will do again later with better thermal paste( reseat HSF).

120127210136.jpg


120127210156.jpg
Click to expand...

Don't let DRAM voltage auto to 1.6TV! I would really recommend you not run 1600MHz on that ram. Intel service reps have said that only 1333MHz is supported and 1600MHz can void your warranty on 2500Ks. One of my 2500K's integrated memory controllers blew up running ram @ 1600MHz, 1.6TV because the motherboard defaulted some new sticks to that speed and voltage. I force them to 1.5V, 1333MHz now and no problems.

1.4V has gotten me 5.0GHz stable on 2/3 2500Ks so you should be able to get that going with a few other tweaks.
 
Spare-Flair said:
. Intel service reps have said that only 1333MHz is supported and 1600MHz can void your warranty on 2500Ks.
Click to expand...

Huh? Do you have a link for that - I find it hard to believe they would care, or be able to tell, or that it would cause any problems. The voltage, yes, but the speed, no.
 
Forceman said:
Huh? Do you have a link for that - I find it hard to believe they would care, or be able to tell, or that it would cause any problems. The voltage, yes, but the speed, no.
Click to expand...

Sorry, I don't have a link handy, just something I read on a tech forum from another user who had to RMA his 2500K. He said he spoke with an Intel service rep and was told this. The 2500K white paper does specify 1333MHz.

I was told on various forums that 1.6V XMP voltage and 1600MHz combination could be dangerous for the 2500K's IMC and it's the most likely culprit in one of my deaths. I can't tell the difference between 1333MHz and 1600MHz performance anyway so I went to the safe side so I don't kill any more CPUs. The first time wasn't my fault, it was the default profile Kingston set for my ram but it definitely fried my first 2500K's memory controller.
 
This is a 2500K that I got from CL last week, with a NZXT Havik 140 cooler that came today.

I Started at 4800Mhz with stock vcore and Primed for about an hour, so this looks pretty good so far.

4900MHz, 100 x49
1.255 vcore

29C idle (1600Mhz)
43C load (65-70C on the cores, is that right?)



edit: ran Prime for 70mins before core 0 error.
 
Last edited:
i know its a year late but can anybody give me a simple 4.4ghz oc for an i7 2600k, i used the auto tuner and its stable but it gave me a max voltage of 1.392 which i feel is too high even though its offset.

i'd just like an everyday 4.4ghz with a safe voltage and clocking everything down when idle

if anyone can help i'd appreciate it, mobo is p8p67 pro
 
ag55 said:
i know its a year late but can anybody give me a simple 4.4ghz oc for an i7 2600k, i used the auto tuner and its stable but it gave me a max voltage of 1.392 which i feel is too high even though its offset.

i'd just like an everyday 4.4ghz with a safe voltage and clocking everything down when idle

if anyone can help i'd appreciate it, mobo is p8p67 pro
Click to expand...

Just set your vcore manually and keep lowering it and running IBT and prime95 to check stability until it is unstable, then back off. Then when you get there, you can try offset, if you haven't from the outset, so you can get better idle voltages and temps.

I have a 2500k stable at 4.4 and Vcore at 1.216V under load with all of the regular settings (LLC regular, VRM auto, standard phase and T.probe, EIST enabled, C states disabled) on a z68. so you should be able to do similar at that frequency I would think.
 
4.4 should be achievable with a voltage of 1.2-1.28 without having to touch any other voltages. I know my 2500k takes around 1.3 for 4.5 GHZ without touching any other settings.
 
Interesting.. I've been running 16GB of 1.65 rated HyperX ram for nearly a year now, however I've been running it at 1.52v since that's all it takes to achieve its rated speeds.
 
How are the current 2500k steppings OC-wise compared to the "older" ones (where you see people with stable +4.5GHz overclocks)? I remember the old P4 days when the stepping was everything when you wanted to achieve a good overclock. :)
Also, back then there was a site where users could post their cpu overclocks among with the stepping of the processor, I just can't remember what that site was called. Is it still alive?
 
Having a hard time keeping this 2600k cool. Old water block isn't compatible with 1155 so sticking with air cooling for now.

Initial testing with vcore at 1.26v (+0.01v) with LLC off, so bring on the droop. 4.3ghz is stable, but IBT will make the stock cooler hit tjmax and start throttling.

Got a hyper 212+ today and am idling at 25-30*c with ambient 23*c. After 10 passes of IBT, max temp recorded is 89*c. At least it's not hitting tjmax anymore. Next test, 1.3v real. Again, I'm hitting 98*c too easily. What?

Is the hyper 212+ supposed to be able to rotate slightly when properly mounted? Swabbed both IHS and heatsink with alcohol. Applied ASC to heatsink and used paper towel to rub in and wipe off. Applied ASC as dot in center of IHS and mounted. Only using the single fan included with the heatsink currently. Put a piece of tape across the top of the fan and heatsink since there was some air leakage there.

That's to say, I'm more or less fine with 4.3ghz at roughly stock vcore (4.2ghz is fine at stock vcore, or rather +0.005v) since I'm paying three times more for electricity than I used to (really the reason why I upgraded from my trusty q6600. that and the $199 MC deal).
 
The heatsink can rotate slightly when mounted, so that's okay, but you shouldn't be getting those temps with that cooler. Might want to check the mounting, or the fan. That cooler should keep that chip happy at 4.5 and 1.35V easy.
 
Did some reading on the heatsink and found that the traditional dot/line method in the center isn't ideal. Took off the heatsink and found paste distribution to be not nearly as bad as typical results (then again, I tend to apply a bit more paste than average probably). Redid it using a lot more than I normally would, applying three thin lines to the aluminum between the heat pipes. Remounted. Same idle. Load immediately steps coretemps to 55*C or so, with temps seeming to peak around 68-74*C, although the max recorded temps show around 80*C. So about 10-15*C cooler than before, but still not terribly impressive, especially considering I'm seeing 1.23v actual under load. Gave the heatsink a few twists and saw no difference in performance. Finding about the same temperatures with P95 as with IBT. For testing purposes, fan is set to 100%. I'll take it for now, since power savings is the goal. Might even sell the watercooling equipment for now (figure an 80W drop in idle consumption must be more than just the cpu. all those fans and pump for watercooling can't be helping).
 
solkku said:
How are the current 2500k steppings OC-wise compared to the "older" ones (where you see people with stable +4.5GHz overclocks)? I remember the old P4 days when the stepping was everything when you wanted to achieve a good overclock. :)
Also, back then there was a site where users could post their cpu overclocks among with the stepping of the processor, I just can't remember what that site was called. Is it still alive?
Click to expand...

I have a few different batches since they have come out and I noticed that my earlier batches ran cooler and with less volts, but the overclocks have been about the same.
 
is there someone here who upgraded from Sandy to Ivy?
Does it worth the money?

Is there some comparative review that shows the improvements with ivy?
thanks
 
ziddey said:
Having a hard time keeping this 2600k cool. Old water block isn't compatible with 1155 so sticking with air cooling for now.

Initial testing with vcore at 1.26v (+0.01v) with LLC off, so bring on the droop. 4.3ghz is stable, but IBT will make the stock cooler hit tjmax and start throttling.

Got a hyper 212+ today and am idling at 25-30*c with ambient 23*c. After 10 passes of IBT, max temp recorded is 89*c. At least it's not hitting tjmax anymore. Next test, 1.3v real. Again, I'm hitting 98*c too easily. What?

Is the hyper 212+ supposed to be able to rotate slightly when properly mounted? Swabbed both IHS and heatsink with alcohol. Applied ASC to heatsink and used paper towel to rub in and wipe off. Applied ASC as dot in center of IHS and mounted. Only using the single fan included with the heatsink currently. Put a piece of tape across the top of the fan and heatsink since there was some air leakage there.

That's to say, I'm more or less fine with 4.3ghz at roughly stock vcore (4.2ghz is fine at stock vcore, or rather +0.005v) since I'm paying three times more for electricity than I used to (really the reason why I upgraded from my trusty q6600. that and the $199 MC deal).
Click to expand...
I'm getting temps of around 45C under full load (prime95) with the i7 2600k with a hyper 212+ at stock clocks (turbo puts it at 3800MHz under full load). before anyone boos me using a K at stock speeds, know that I'm mostly running this at stock to make sure everything is stable.

Edit, Running IBT to see how hot it gets...
64bit mode
standard dtress level 1024MB
10 times to run
8 threads
 
Update to my last post... The ASUS software is where I was getting those numbers from before. just used coretemp to get the temps close to what you're getting.

Note max temp below:


Uploaded with ImageShack.us
 
edit: I had some idiotic settings and was throttling myself to get "5.0 ghz." With a proper setup (and using LinX w/ AVX instead of IBT), the settings I posted here didn't come close to stable.

I'll probably need ~1.49V idle, ~1.45V load to hit 5.0 on this chip, but I stopped at 4.9 (~1.44/1.40) for now, and am just going to run at 4.8 for a stable 24/7 OC.

I'm not done testing stability, but these settings are looking pretty stable (just finished 25 loops):

48 x 100 (100.2 in bios)
vcore 1.41 (~1.37v load)
LLC level 4
PLL 1.7
VTT ~1.115
C3/C6 disabled
EIST, C1E, Turbo, everything else enabled
Max Power/Current both @ 150
RAM set with the auto XMP profile.

Nothing else much different from stock

xY03C.png
 
Last edited:
I can send you screenshots of my configuration. I have 4.4GHz. I can't seem to get a higher OC without slowing down my RAM. I bought some 2133 Gskill Ram and if i run it at the rated speeds, i can't get past 4.4Ghz. For me 4.4Ghz and still being able to run my ram at the rated speed is good enough. I have an ASROCK Z68 board but the basic settings should be the same.
 
I don't understand this line of thinking at all. The thing is, you'd get better actual system performance in applications with less memory speed and higher CPU clocks.
 
Champ said:
Is 4.0 on air/cooler master 212 easy on a 2500k?
Click to expand...

The 212 Hyper are well known for their performance/price point. Since the SB pretty much comes with a stock 4.0 (Turbo) setting, they're kinda made to for it. With the 212 cooler, I don't think you'll have any issues. The main thing is just to make sure your PC's general airflow is good. Only having a good cooler on the CPU won't help overall performance much. But, tbh I don't see you having any issues with what you're asking. OC away, and be careful!:cool:
 
And just to follow up on that, I used a 212 Hyper (I think it's the newer version) for quite some time on my 2600k and never had any problems. I did however upgrade to a H70 (closed water loop) which brought the temps down almost 10 degrees. This isn't always the case though...
 
kostonn760 said:
The 212 Hyper are well known for their performance/price point. Since the SB pretty much comes with a stock 4.0 (Turbo) setting, they're kinda made to for it. With the 212 cooler, I don't think you'll have any issues. The main thing is just to make sure your PC's general airflow is good. Only having a good cooler on the CPU won't help overall performance much. But, tbh I don't see you having any issues with what you're asking. OC away, and be careful!:cool:
Click to expand...

kostonn760 said:
And just to follow up on that, I used a 212 Hyper (I think it's the newer version) for quite some time on my 2600k and never had any problems. I did however upgrade to a H70 (closed water loop) which brought the temps down almost 10 degrees. This isn't always the case though...
Click to expand...

Perfect. My computer knowledge is far behind now. I use to know everything when it during the P4/Celeron D/C2D era, but haven't been paying attention. Seems like Water cooling is a lot easier and more reliable then before. Thanks
 
I just got the H100. Would that be sufficient to get 4.5-4.6 at very low temps? And does anyone have an OC guide for beginners? Am I just raising the multiplier?
 
Champ said:
I just got the H100. Would that be sufficient to get 4.5-4.6 at very low temps? And does anyone have an OC guide for beginners? Am I just raising the multiplier?
Click to expand...

How low is your low? I have an H100 and it's running my 3960X at 4.8 Ghz. It can handle IBT Max Extreme just fine.

For normal usage, it never goes up to 60 even when I'm using all 6 cores (video processing). I think the H100 can handle 4 cores just fine.
 
Champ said:
Yeah I want to hit 4.8 and not go over 70c load. Is that realistic?
Click to expand...

Depends on what voltage you need. I'd say it is do-able, but on the edge. Are you talking Prime95/Linx load temps, or normal use load temps? Because normal use load temps should be no problem.
 
Normal load temps like in BF3. I'm idling at 29c at stock speeds atm with the h100.
 
That is with EIST on. With EIST off, it peaks at 42c idle (36-42). Is that normal?
 
^yeah. I'm hoping now to just stay at 4.8 and only go 65c max on load with the H100 assuming the H100 is better then the 212.
 
You must log in or register to reply here.
Back
Top