• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

OC testing

M

manny1222

Gawd
Joined
Mar 11, 2012
Messages
732
Just installed my 8700K last night and dozed off during windows installation. Can i enable MCE and XMP at the same time? I wont have time for a couple of weeks to dial in a serious overclock (and I'm headed to work) but I'm eager to start gaming again. I figure 4.7 all core and 3000 MHz memory should be good for now
 
Last edited:
yeah you will be fine, should be good, just try to keep an eye monitoring temps, as some boards love to overvolt causing higher than expected temps.
 
I was able to hit 5.0 at 1.20 volts.

I found there was no return vs. Heat and power past 4.6ghz on that chip.

In gaming at 4.7plus it would hit 70 regularly while gaming.
 
So what program is everyone using to test overclock stability these days?
Not sure i want to pay for AIDA64 since I'm not a reviewer and i wont be using this until next time (hopefully not for a few years).
I tried testing with Prime95 but its failing the test after 10 mins. I tested using custom settings, in place fft, 8 min, 4096 max, 3 min runs. My attempted overclock is 4.8GHz at 1.25v, adaptive mode (cpu-z and coretemp report vid 1.283 and vcore 1.255), LLC mode at mode 4 (almost flat i think) and xmp enabled for the 3000MHz memory.
I'm reading online a lot of people saying Prime95 is not appropriate for testing overclock because it puts undue torture on the cpu which I'm inclined to believe because even though the overclock failed Prime95 with a couple of cores hitting 94C, but i decided to try it out anyway. I played iracing (what i play most of the time) for 5 hours without issues and my temps below 40C (only going up to 43C when i had OBS running).
 
manny1222 said:
So what program is everyone using to test overclock stability these days?
Not sure i want to pay for AIDA64 since I'm not a reviewer and i wont be using this until next time (hopefully not for a few years).
I tried testing with Prime95 but its failing the test after 10 mins. I tested using custom settings, in place fft, 8 min, 4096 max, 3 min runs. My attempted overclock is 4.8GHz at 1.25v, adaptive mode (cpu-z and coretemp report vid 1.283 and vcore 1.255), LLC mode at mode 4 (almost flat i think) and xmp enabled for the 3000MHz memory.
I'm reading online a lot of people saying Prime95 is not appropriate for testing overclock because it puts undue torture on the cpu which I'm inclined to believe because even though the overclock failed Prime95 with a couple of cores hitting 94C, but i decided to try it out anyway. I played iracing (what i play most of the time) for 5 hours without issues and my temps below 40C (only going up to 43C when i had OBS running).
Click to expand...
Have you changed your AVX offset? If not, I would set it to -2. I would also use Prime95 v26.6 to test without AVX as well as using the currect version to test with. Also, Realbench for a couple hours is a good idea.

If it is failing in Prime95...it is not really stable. :(
 
I dont know much about avx, but from what i see it the offset reduces the clock by 200MHz during the prime95 testing. Feels kinna like cheating since its not really running at the speed I'm intending. Is avx a special kind of load to the system? If so, how much does it apply in real life scenarios? Is there any games that put that kinna load on the system or is it all benchmarking? I could technically set my overclock to 5.0GHz with an avx offset of 5. Is 5.0GHz a stable overclock with that kinna testing?
 
So I downloaded the v26.6 and its going fine for 15mins now. Will i call it stable if its passes with v26.6 but is failing with the latest version?
 
If you are failing stability tests, then merely ignoring the failures won't magically make your system stable.

Here is the guide to actually achieving stability.

https://hardforum.com/threads/intel...a-1151-6c-12t.1930226/page-84#post-1043318983

Failing any test that would return a non-failure on a stock system means you are by definition not stable.

If you are hitting TJmax with any workload, then you cannot reasonably test for stability.

If you are hitting TJmax with any workload, then you need to increase you cooling capacity, along with, if necessary, delidding your CPU.

This is also one of the best ways of testing how able your cooling system is to stave off hitting TJmax.

https://hardforum.com/threads/testi...ath-kernel-library-optimized-linpack.1947249/

Prime95 is, of course, also a good way of doing so.
 
Last edited:
manny1222 said:
I dont know much about avx, but from what i see it the offset reduces the clock by 200MHz during the prime95 testing. Feels kinna like cheating since its not really running at the speed I'm intending. Is avx a special kind of load to the system? If so, how much does it apply in real life scenarios? Is there any games that put that kinna load on the system or is it all benchmarking? I could technically set my overclock to 5.0GHz with an avx offset of 5. Is 5.0GHz a stable overclock with that kinna testing?
Click to expand...
Yes, AVX is a special kind of instruction set that requires more current to run. No, it does not apply to real life scenarios very often...mostly when encoding. I have not noticed any of the games I play using AVX.


manny1222 said:
So I downloaded the v26.6 and its going fine for 15mins now. Will i call it stable if its passes with v26.6 but is failing with the latest version?
Click to expand...
No. You need to get it so it runs both without failing with acceptable temps.

If you want to get the most out of your system, set your overclock initially using non AVX testing. Then start testing with AVX increasing the AVX offset as needed to control temps.

Example; if you end up with a 50 multi and -3 AVX offset, when you game, benchmark, and do normal tasks it will run 5GHz. If you start encoding or running something with AVX instructions it will run them at 4.7GHz helping to maintain temps.

Just because you can not cool 5GHz while running AVX does not you should not or cannot run 5GHz for everything else.

The AVX offset feature in the bios is there for a reason. It is definitely not "cheating".
 
Thanks for the replies.
So I've had running v26.6 running for 30 minutes now with the CPU @ 4.8GHz with 1.275v. Max temp is 80C so far. So if this stable for another 90 minutes, i can go try with the newest version with the avx offset. If that's stable through the night, am i good then?
I tried Cinebench R15 with these settings and it completed with a score of 1561.
 
I usually start testing with games after I pass 3-4 hours. I will sometimes bump the vcore up just a notch to ensure stability as well.
 
Well I'll be headed to bed soon. I decided to try 5.0 @ 1.3v. I was able to run Cinebench. Now running Prime95 v26.6. I will do that for an hour then switch to the latest with the avx offset and see what happens.
 
You must log in or register to reply here.
Back
Top