Hello everyone,
I've recently overclocked my i7-6700k to 4.6 GHz @ 1.312V and I'm not sure if my temps are what they should be.
After running a 24 hour P95 run, my temps are about 70C, which seems a little high for a watercooler.
Is my cooler defective? Maybe I have it mounted wrong.
Any help is appreciated.
have you oil it up before putting on the cooler?
(joke)
1) check pump speeds
2) check fan speeds
3) check where are you blowing out the hot air...
4) why 1.3V? you don't need that much.
5) what is your room temp.
6) check thermal compound ( you have used it right? )
7) check how your compound spread out itself...
Well I need 1.3V to run @ 4.6 GHz, so nothing I can do there.
Pump is running at 2880 rpm, fans 1440 rpm.
Re-installed the cooler with fresh compound (the previous one was spread out nicely). Still running at 70 C.
My room is at 25-26C.
But the question is - what temps should I expect? I've heard that CLCs provide with almost the same performance as a really good air cooler. Maybe my temps are normal. I honestly have no idea, so that's why I'm asking.
you should have around ~65 from what i've seen from forums on this cpu.
why do you need 4.6GHz?
that's about typical... intel chips run hot under water. my i5 4690k @ 1.38 runs that hot under a custom loop. Intel just specializes in shitty thermal interfaces.
Honestly i do recall intel stating that only the intel and OCCT tests were valid for stress testing and that prime95 is actually too harsh on the parts. Id say 70 C after a 24 hour P95 run is perfectly permissable and Dont do it again because realistically an AIO is never going to be as good as a full loop and realistically your never going to reach that kind of load and if that kind of computation is your intent you should have bought a CPU validated for scientific number crunching.. AKA intel Xeons and AMD Opterons
The better question is - why not? :)
I mean I don't need it but it is still nice to have and compared to other people, I'm running 4.6 GHz at a pretty low voltage.
70°C is totaly fine, and not hot at all.
Also i would recommend to not use P95 on intel cpu´s.
A little on the high side for my taste, but it sounds about right, temp wise.
Curious, though. Do skylakes really need that much voltage?? I haven't looked into it... My 4790k O/C'ed to 4.8ghz @1.29v tops out at 66c. It's cooled by an H100i. At 4.6ghz, it needs 1.18v. And at 4.4ghz, it needs 1.10v.
Skylake has been harder to get pushed beyond the 4.5Ghz mark then previous generations. Although he is using a lot of voltage...
1.312V for Skylake is not abnormal either.
Ive 'been well past 1.4v on an FX 6300 at work and pushed it to 4.7GHz stable... never had a problem on an Evo 212 and it's hit 80 pretty regularly... for 3 years... you're more than fine
well AMD FX cpu´s are ofc a totaly diffrent story for that matter.
But wenn it comes to skylake he should basicly be fine.
I just dont recommend to use P95 for 24hours on an intel cpu.
I was simply illustrating that 70c isn't anything to worry about on CPUs... I get that they are completely different architecture and operate at different temps... but it doesn't downplay the fact that CPUs can run hot and still compute without issues... and TBH I hear all this don't stress test me jargin on here and I don't get why Skylake, which naturally runs much cooler than Haswell... "can't handle" benchmarks...
Has anyone actually done anything to a CPU to render it less capable during a benchmark.... personally? Cause I call BS @ this...of course prime95 works your CPU to the max... but the only time I've heard this "don't run P95" shit is when the 9590 came out and that was because most people were running it with a tiny "it came with the CPU" Seidon AIO and that was the hottest CPU that's hit the market in my time of paying attention...
But to say a central processing unit shouldn't be ran on bench to test an overclock is ludicrous...
i would recommend to read this.
Basicly you could use prime95 on Haswell and newer,
but i would not recommend to let it run for 24 hours tbh.
There are also other stress test tools which are a bit more safe to use.
like OCCT, Aida64, intel burnin test, Asus real bench.
If a rep from Asus says their overclocking series motherboards weren't designed to be able to pass a stress test... then you should not buy Asus motherboards... OF COURSE more current is soaked up by a CPU in a stress test... the motherboard is very specifically designed to deliver a safe amount of current/voltage for the CPU before it trips a blue screen...
to say it even closely mimics a real world blue screen is also absurd... but to insinuate that the motherboard would let a CPU fry itself would also insinuate that the motherboard itself is not designed properly... a Z-series or AM3+ board, by definition, are designed for overclocking...and you check overclocks with stress tests...
It's something to be said, however, that if it doesn't bluescreen in 10 minutes, you're good :P