i5 2500k OC question
i5 2500k OC question
I have been experimenting with an overclock of 4.5Ghz on my 2500k for several years now. I’m not very experienced with this process, but I followed a guide that matched my Gigabyte motherboard. Initially, around three years ago, I was quite stable at about 1.3V. Over time, I noticed I needed to raise the voltage further to maintain stability, eventually reaching around 1.416V. This also caused my temperatures to rise significantly—around 75 degrees after running Prime95 for 20 minutes.
Could this be a sign the chip is failing? Is 1.4V even safe? And is it acceptable to reach 75 degrees under load at that temperature?
My BIOS settings are mostly similar to the guide I used, but I’m unsure about things like "QPI/Vtt voltage" and "system agent voltage." Would adjusting any of these settings help improve stability and allow me to run at a lower voltage? Thank you in advance for your help.
i wouldn't make those changes. my cpu matches yours, and i have a gigabyte bios; the only adjustments i've made are the offset voltage. everything else runs automatically. if i were you, i'd reset it back to factory settings and attempt a new overclock. cpuz or similar tools are displaying a low vcore, around 0.80v, since the cpu is throttling under load. this is normal and acceptable. for testing initial overclocks, you can disable c1e, c3, c6 and speedstep to avoid throttling, ensuring accurate vcore readings. however, make sure to reactivate power savings once a stable overclock is achieved. it's possible that certain power-saving features are contributing to instability due to vdroop. the 1.416v is...
The safe voltage for the 2500K is around 1.52v. I wouldn't usually run it above 1.4v every day, that would be too risky. The 75c is fine, no major concerns. Other voltages shouldn't really affect stability unless the RAM is the issue. Probably not likely.
Is it cooler? Have you experimented with a different thermal compound?
it's normal for the chip to require higher voltage as time goes on. if you've replaced the thermal compound properly, i think it would have made a noticeable difference in temperatures. i wouldn't alter the vtt or system agent voltage settings. most people agree that 1.38v under load should be the maximum, though a slight increase is unlikely to cause issues as long as temperatures remain acceptable. 75c is acceptable for sandy bridge, and most recommend staying under 80c in the long run. however, a small margin above 80c for sandy bridge should be okay. regarding temperatures and voltage, sandy bridge is considered one of the more resilient intel core processors.
about your overclocking approach... did you simply copy and paste from that review? if so, it could lead to problems. are you using an offset/dynamic vcore or a fixed one? how do you determine your vcore? are you mentioning 1.416v in the bios, or is that what the vcore readings from cpu/cuza/aida/realtemp/ect indicate?
the chip tends to require increased voltage as time progresses. if you have adjusted the thermal paste correctly, it would have improved temperatures somewhat. i wouldn't alter the vtt or system agent voltage settings. most people agree that 1.38v under load should be the upper limit, though a slight excess probably won't damage most processors as long as temperatures remain acceptable. 75c is acceptable for sandy bridge, and most recommend keeping below 80c in the long run. however, a slight rise above 80c for sandy bridge should be acceptable. regarding temperatures and voltage, sandy bridge is considered one of the more resilient intel core processors.
about your overclocking approach... did you simply copy and paste from that review? if so, it could lead to problems. are you employing an offset/dynamic vcore or a fixed one? how do you monitor your vcore? are you suggesting you inserted 1.416v in the bios, or is that what the vcore on cpu, aida64, or realtemp/ect indicates?
i just copied similar values for the system agent voltage and power saving settings, not the vcore. i’m using a dynamic vcore because my bios prevents fixed vcore settings. both my bios and aida64 show 1.416v, while cpu reports a strange low vcore reading—this seems typical for 2500ks and is often seen in cpuz when it’s experiencing issues. under load, my vcore drops to 1.368v in cpu64.
I wouldn't make those changes. My CPU is quite similar to yours, and I only adjusted the offset voltage. All other settings remain unchanged. If I were you, I’d revert everything to the original configuration and attempt a new overclock.
CPU models like this typically display a low vcore, around 0.80v, since the CPU is throttling under load. This behavior is normal and acceptable. For testing overclocking, you can disable certain control registers and speedstep to avoid throttling, allowing more accurate readings. However, once a stable overclock is achieved, you should reactivate power-saving features.
It’s possible that some power-saving mechanisms are introducing instability due to voltage droop. The 1.416v spike is typical when reducing load, while the 1.368v represents the actual vcore we’re concerned about at 4.5ghz—this is within normal limits for your chip. I’d be more concerned if the load was dropping to 1.40v with a droop above 1.45v. You might consider using an LLCC, but it’s probably not necessary for a 4.5ghz setup.
The system agent, QPI/VTT, and PLV should remain unchanged for a modest 4.5ghz overclock. I’d set the PLV overvoltage to auto or even lower it slightly to help dissipate heat, but only if the system remains stable.