Overclocking 2500k and 560ti

Tar.Aldarion

1.33 was fine for 4.5 for an hour.


1.37 is failing for 4.6 after 2 mins.

Going for 1.4v and 4.6ghz to see what happens. If that fails then I will stick with 4.5 and run the test for longer. If it passes at 4.6ghz I will try to lower vcore bit by bit and stay at that.

Tar.Aldarion

Well I've ran it at 4.6ghz and 1.376v for a few hours and these are the temps I get, average of 72C, high of 76C.

Are these acceptable/safe for 24/7? I know normally my chip wont be this high gaming etc.

sink

Tar.Aldarion wrote: »

Well I've ran it at 4.6ghz and 1.376v for a few hours and these are the temps I get, average of 72C, high of 76C.

Are these acceptable/safe for 24/7? I know normally my chip wont be this high gaming etc.

That's way too hot. 65C would be the limit, maybe let the high get a little over when stressing but you want the average below 65C or your chip will have a very limited lifespan. I'm running mine with a max of 58C at 4.6GHz.

Also your voltage is high for that chip, going by my own experience and other peoples posts you shouldn't need that much voltage to reach 4.6Ghz, more like 1.315V. Excess voltage is also likely what's causing your chip to get that hot.

IMO your RAM voltage is what's causing the problem 1.5V is what you should be running it at stock, you're going 40% above that. Try 4.5GHz @ 1.315V which should be enough voltage for your CPU and your RAM @ 1.645V, and manually set the latency to stock CL 9-9-9-24. If you're stable at that you can then trim back the RAM voltage and latency.

I have my RAM running stable for months at 1.6V, it wouldn't even successfully boot windows at 1.5V

Tar.Aldarion

sink wrote: »

That's way too hot. 65C would be the limit, maybe let the high get a little over when stressing but you want the average below 65C or your chip will have a very limited lifespan. I'm running mine with a max of 58C at 4.6GHz.

Also your voltage is high for that chip, going by my own experience and other peoples posts you shouldn't need that much voltage to reach 4.6Ghz, more like 1.315V. Excess voltage is also likely what's causing your chip to get that hot.

IMO your RAM voltage is what's causing the problem 1.5V is what you should be running it at stock, you're going 40% above that. Try 4.5GHz @ 1.315V which should be enough voltage for your CPU and your RAM @ 1.645V, and manually set the latency to stock CL 9-9-9-24. If you're stable at that you can then trim back the RAM voltage and latency.

I have my RAM running stable for months at 1.6V, it wouldn't even successfully boot windows at 1.5V

Oh is it? crap! most people were telling me under 80 max and everything is fine.

for 4.5 I could fo 1.33 fine, but needed a big jump up for 4.6, 1.375. it was bsoding 2 mins in otherwise, what do you mean that im going 40% above that? I have it set at 1.5? my ram latency should be 9-9-9-24 by default, so should I bother setting it manually? Did you trim back the timings on yours or leave em at that?

So basically I should lower cpu voltage while raising ram voltage to 1.55-1.645?


i will try that now, i know you say my V is too high, i seem to be in the range seen on a lot of forums but i dont think they went near ram voltage.

sink

Tar.Aldarion wrote: »

Oh is it? crap! most people were telling me under 80 max and everything is fine.

for 4.5 I could fo 1.33 fine, but needed a big jump up for 4.6, 1.375. it was bsoding 2 mins in otherwise, what do you mean that im going 40% above that? I have it set at 1.5? my ram latency should be 9-9-9-24 by default, so should I bother setting it manually? Did you trim back the timings on yours or leave em at that?

So basically I should lower cpu voltage while raising ram voltage to 1.55-1.645?

Don't know who was telling you 80C was OK. That would melt your processor in no time.

http://ark.intel.com/products/52210/Intel-Core-i5-2500K-Processor-(6M-Cache-3_30-GHz)

The TCase of the 2500K is 72.6C, that is the maximum that CPU can handle before suffering degradation and going over that will damage your CPU in no time. In reality you want to keep it a good bit below that to give you a bit of head room all it takes is a hot day a bit of dust build up to push your core temp up a few degrees. The temperature your CPU runs at over it's life will directly impact it's lifespan. If you want to keep it a good number of years you want it to be cool

I just have a feeling that your RAM voltage is too low. I'm no expert I've only overclocked a few CPU's but for every one of them I've had to up the RAM voltage for it to be stable. You won't need to keep it at 1.645V permanently, this is just to rule our under voltage as your problem. Your RAM won't be damaged at that voltage and once you know it's not under voltage that's causing the problem you can look at other factors. If it is stable at that, then it is under voltage that was causing the problem and you can try lowering it and seeing what the minimum voltage it will take to be stable.

Tar.Aldarion

ah ok, they said that that tcase didnt matter, and that mid 70s 24/7 use was fine, and that it was the tmax of 98C that was the real figure to worry about. And that since i would never be near prime95 figures anyway that those figures are fine.

Still i am uncomfortable with that hot a chip as you say. I dont want a dead chip too fast!
I put the ram voltage up to 1.645 and even then got bsod in prime at 1.33, 1.34, 1.35, 1.36,and 1.37 so the ram is ruled out, i think its my chip.

still that will yield a high temp id guess so i might have to go to 4.5ghz

Tar.Aldarion

think ill drop to 4.5 as sink said that is very hot! Am testing out 4.5 at 1.335 which should run fine.

So I can choose between:

4.5Ghz @ 1.335v with temps of 66-70C at highest prime 95
4.6Ghz @ 1.375v with temps of 72-76C at highest prime 95


I wish it wouldnt take so much more voltage to jump to 4.6ghz

Tar.Aldarion

i thought it was stable 4.5Ghz @ 1.335v but after a few hours it failed the stress test, same again with all voltages up to 1.35. testing 1.355 atm, so im not too far off the 1.375 needed for 4.6 after all.

Tar.Aldarion

think ill settle for 1.352v @ 4.5Ghz. 1 core is 60, the other cores are generally 65-67 with a high of 70 after a few hours. Temps will generally be like 10-20C under these gaming anyway. Still, it might crash yet!

wish my chip wasnt so power hungry, as it would be easier to go higher then!

Tar.Aldarion

well that was ok for 4 hours so ill just leave it at that until i see a crash during normal use. gonna test temps using counter strike source to see what my real temps will be!

savemejebus

Hmm, long time since i overclocked properly but it used to be good to leave it running prime 95 or occt for 24 hours so that you can be sure it's stable and you won't get any file corruption.

Tar.Aldarion

will do that when i get the chance! As I really should be sure.

for now i will game!

Tar.Aldarion

gaming for a few hours max settings, highest temp of package is 60. this will do i think!

tuxy

sink wrote: »

Don't know who was telling you 80C was OK. That would melt your processor in no time.

http://ark.intel.com/products/52210/Intel-Core-i5-2500K-Processor-(6M-Cache-3_30-GHz)

The TCase of the 2500K is 72.6C, that is the maximum that CPU can handle before suffering degradation and going over that will damage your CPU in no time.

I don't think this is correct. Tcase is the max temperature as measured on the centre of the outside metal top of the processor.

TjMax max is the maximum heat of the core, which is what is measured by the sensor that gives the reading in bios and OS(windows ect).
I believe tjmak for the 2500k is about 98C

*Figured I would reply to this even though it's old and my people use old threads for guides

Tar.Aldarion

my temps never go near the stress tests even when gaming for ages etc. fairly sure i could go much that 4.5 because of that even though the tests were hot but I shan't for safety

/old thread update