OCing my 2500k

litebeer

Distinguished
Jan 20, 2012
22
0
18,520
So, I know I'm late to the game here, but I finally OC'd my 2500k. I can get to 4.8ghz no problem, but the temps go way up.

What voltage do you think is the "lowest" i could go and still hit these temps?

Never done this before, but figured I should take advantage of my Corsair H60.

Thanks for the advice...im currently at 1.415v (suggested on [H]...old asus post)
 
Solution
The difference between the two temps you recorded was that the Ai Suite is the cpu case temp and the other the actual core temp. 4.8 is very respectable and you could with some tweaking possibly/likely get it down which would of course lower your temps. The silicone lottery and other hardware variables make it hard and unreliable to borrow settings but in the most general terms it is quite possible to reduce your 1.415 to 1.385 @4.8 but that's a guesstimate and perhaps even lower but also maybe not ?? lottery.

For instance come down a bit as suggested by Photonboy and test whether you can boot with lower voltages. As an example based on what I am familiar with on the 2700k try for 4.2 at 1.25V. if it does it lower the voltage and boot...
My advice is drop down to no more than 4.6GHz.

The i5-2500K is still a good CPU and at these frequencies is rarely the bottleneck in games.

Thus, use 4.6GHz and bring up the Voltage until you remain stable.

(Even if it appears to be stable at 4.8GHz, you may get errors you aren't aware of causing corruption, and as well CPU's can't necessarily maintain a high overclock state.)
 
The difference between the two temps you recorded was that the Ai Suite is the cpu case temp and the other the actual core temp. 4.8 is very respectable and you could with some tweaking possibly/likely get it down which would of course lower your temps. The silicone lottery and other hardware variables make it hard and unreliable to borrow settings but in the most general terms it is quite possible to reduce your 1.415 to 1.385 @4.8 but that's a guesstimate and perhaps even lower but also maybe not ?? lottery.

For instance come down a bit as suggested by Photonboy and test whether you can boot with lower voltages. As an example based on what I am familiar with on the 2700k try for 4.2 at 1.25V. if it does it lower the voltage and boot again and repeat until it won't boot. Once you've found that lowest threshold for a setting or two you'll likely be able to approximate the next higher setting and voltage required though the increase as you head higher won't necessarily always be the same increment.

Then stability will require uping the voltage but you will be certain that it's no more than your actual gear requires instead of some arbitrary settings picked up by other folks settings that may in fact turn out to be more taxing on your gear than needs be.

That said as a reference only here are some I use that I tested for stability long ago. I was able to boot 4.8 at 1.34v (but boot was all I could do) Prime was stable at 1.37v and OCCT all the highest setting 1.385v with temps reaching 82C. Beyond that stability was precarious I believe more to do with my MB than the chip nonetheless I've benchmarked various 2d benches at 5@1.435v, 5.1@1.45v, 5.2@1.475v and 5.3 required1.495v (2 cores) just enough to get the CPU_Z recorded. Then to complete SuperPI @ 5.3 required yet more V's and that was as I recall 1.512V just enough to submit the score and it was also as high I as will go until the near end of this puters life.

Practice at lower setting then experiment where ever sensible temps and voltages allow you to go. Beyond there, invest enough time and serious study to really understand the risks and weight whether they are worth it. It stops being fun or smart if you blow or fry gear.
 
Solution