Temp and Overclock check on i7-4790k please!

I recently overclocked my i7-4790k and just want to post the temperatures and such to see if its a safe overclock or if y'all have any advice for me to change.

Pc specs
i7-4790k
asus maximus hero vii mobo
h100i cpu cooler
gtx 980
corsair 450d two from intake fans, rear exhaust, two fans up top on h100i rad

in the bios, i turned the profile on xmp, adjusted all cores synced to 45, and manual voltage override of 1.2v

after 1 HR 24 MIN on aida64, my temps were( from core #1 to #4 ) max temp of 73 72 73 72 / And average temp of 63 63 63 60
http://gyazo.com/e2fb2002db4a342e6fb7c5fb34dd1a95 ( aida64 temp screenshot )

my Rad fans were only running @ 1300 rpm so not cranked up at all on my h100i

everything went smooth, no crashes or reboots etc.

just wondering if this is solid for a 4.5ghz overclock or if i should adjust anything differently like voltage since it was perfectly stable, FIRST TIME EVER overclocking so im pretty new, be easy on me haha ! Also i dont care about going past 4.5GHZ right now but if you would like to give me advice to hit 4.6 etc, im interested.

gabebrunelle, welcome to the forum. If you passed 1.24 hours of Aida and are not having issues when running the Pc then you're good to go. How did you go about ocing it though? If you Oced it thought the BIOS manually then going for 4.6 is the same as going for 4.5. Raise the multiplier and stress test, if it fails add voltage and keep the temps below 90C. The temps you have now are fine, no worries there. You can also post screen shots of the following tabs in Cpu-Z, Cpu , Spd and Memory so we can have a peek under the hood so to speak. Use the inforum picture hosting tool though, some people can't access outside links. Here is how, if you right click and hit view image you be able to read it better.

Those temps are certainly safe but AIDA64 is not a particularly good heater upper. I'd like to see what kind of temps IBT ("very high" setting) or Prime95 (v. 2.79) would give. With IBT you will get max temps in about 5 minutes. Run it for about 20 rounds to check for stability and Prime95 for about 2 hrs.

Congrats on your first ever overclock!!
The only thing i would suggest (if you have the time of course) is testing it longer under the same conditions. Or maybe try a couple of other stress testing programs ^^ like Trents said =)
Just make sure you keep an eye on the temperatures when running new test for the first time. It will literally only take a few seconds for your CPU to hit 100+ degrees C even from a casual idle temperature of say 30. And if you open up other programs while testing.. this will use more ram and in turn create more overall heat.
I personally like to go for short 5 hour tests at first.
If you chip holds up and your pc doesn't BSOD (Blue screen of death) or shut down and restart.. go for 12 hours!
I like to use diffeent programs too, to ensure my CPU is fully stable. In my eyes.. if it holds out for 24 hours under stress + general use. It's stable!
Good luck!

with IBT i averaged 82 degrees through all 20 tests, is that good?

also PC idles at around 35 overclocked right now, and if i play a steam game (CS:GO in particular, i run about 45 degrees)

gabebrunelle said:

with IBT i averaged 82 degrees through all 20 tests, is that good?

also PC idles at around 35 overclocked right now, and if i play a steam game (CS:GO in particular, i run about 45 degrees)

Click to expand...

Everyone has their own standard for stability gabebrunelle. For me personally it is 2 hours of Prime Blend using the older version that Trents mentioned.If one of my rigs can pass 2 hours of Prime Blend I know I can throw the kitchen sink at it without fail. I do also use my rig for "real work" not just gaming and if it were to crash on me while working it could be costly for me, which it has not ever, If it passes the aformentioned Prime blend. Idle temperatures do not mean all that much, it's really when one is putting a load on the cpu that counts. This is a pretty good guide to read and it is also done on an Asus motherboard so it should be fairly easy to follow.

90c and below is safe. But you say you "averaged" 82c. What was your max temp? If you will monitor temps and voltages with HWMonitor (non pro version) then it will keep track of max, min and current temp.

I agree with Manny. Nothing like Prime95 for two hours. I also would follow that up with the 3DMark battery. I have found that 3DMark is a sleeper when it comes to testing for stability. It may do a better job than P95 or IBT in ferreting out memory settings that aren't quite stable, like timings that are too tight that might need a touch more voltage to the DRAM.

- - - Updated - - -

90c and below is safe. But you say you "averaged" 82c. What was your max temp? If you will monitor temps and voltages with HWMonitor (non pro version) then it will keep track of max, min and current temp.

I agree with Manny. Nothing like Prime95 for two hours. I also would follow that up with the 3DMark battery. I have found that 3DMark is a sleeper when it comes to testing for stability. It may do a better job than P95 or IBT in ferreting out memory settings that aren't quite stable, like timings that are too tight that might need a touch more voltage to the DRAM.

I've never been in a situation where p95 small fft and blend pass for say an hour, and then have any 3dmark fail. Are you using the non avx p95?

I'm curious as to why yoy see those results!

Non avx version of P95?

The older versions that don't get your cpu as hot. I believe 28.5 and newer do.

I understand wanting to make sure that you have complete stability. Your temps are exactly where they should be. Prime and aida are nice stress tests to run but absolutely useless with 4770k and 4790k's. The only stress test you should have to use with these 2 cpu's are XTU and Realbench. Everything else is just worthless. Your temps look like they are right on par though and should never see them get that high in any other application. You do want to work on your cache/uncore though. Go for that 1:1 ratio but if you cant get it there you will be fine with 200-300MHz behind core speed without losing any performance. Just make sure you keep uncore voltage below core voltage. Also if you really do not ever want to worry about temps Delid that 4790k. You will be very happy with the temperature results after a delid, I promise you.

I would start by setting your XMP profile, set all cores to 46 and set voltage on core to 1.25. Boot in, run a few loops of realbench and if stable set all cores to 47 ...etc... If not stable bump core voltage up 0.025 at a time until u find stability. With 4770k and 4790k you really do not have to touch anything else but cores and voltage and set XMP. Im sure someone will chime in here with the easiest ways for them to overclock these chips but I have built systems and delidded at least 20 of these chips and have a lot of experience with them. Once you have a nice stable overclock on the cores move to the cache and do the same thing. Now if you can find a stable overclock at say 4.8 under 1.3v I would say you have a golden chip and then you should absolutely delid it so you have more room for overclocking even further.

Starting to ramble here. Let me know if you have any questions and I would be more than happy to walk you through it.

ed_biii said:

. . . Prime and aida are nice stress tests to run but absolutely useless with 4770k and 4790k's. The only stress test you should have to use with these 2 cpu's are XTU and Realbench. Everything else is just worthless . . .

Click to expand...

I'd like to know your reasoning here. Prime . . . worthless? Seems to me it's the other way around. I would agree with you about AIDA but not Prime. Prime and IBT are the only two stress tests that even make an overclocked system breathe hard. Those others are wimpy. I'm running the XTU CPU stress test as I type on a G3269 and there's not even any lag at all when I type. I couldn't say that if I was stressing with Prime95.

Interesting how there are so many opinions on stress testing software. I fond this:

"ASUS warned us against using synthetic stress tests like are shown above.

Unvalidated stress tests are not advised ( such as Prime 95 or LinX or OCCT, Intel Burn Test or other comparable applications ). For high grade CPU/IMC and System Bus testing Aida64 is recommended along with general applications usage like PC Mark 7. Aida has an advantage as it is stability test has been designed for the Haswell architecture and test specific functions like AES, AVX and other instruction sets that prime and like synthetics do not touch. As such not only does it load the CPU 100% but will also test other parts of CPU not used under applications like Prime95. Other applications to consider are SiSoft 2013 or Passmark BurnIn. Additionally this generation has a more specialized point of consideration for synthetic stress tests. When using an adaptive vid voltage control will be automatically controlled by the iVR when a complex concurrent AVX load is initialized from Applications like Prime95 or Aida Or LinX more voltage will be supplied than has been defined/requested." http://www.hardocp.com/article/2013...74770k_ipc_overclocking_review/7#.VWo8EUYYE9o

Okay, just for comparison's sake, I ran the Intel Xtreme Tuner stress test for 8 hr. yesterday and passed without a hiccup. With the same overclock settings running Prime95 v27.9 I blue screened after little more than an hour. I'm going to leave the settings as they are for awhile to see if it's stable in daily use.

But then again, I just ran Prime95 v2.85 with the AVX and stopped it after 27.5 hr. Funny that v28.5 didn't blue screen like v27.9 did. Temp was only 1c higher with the later version. My cooling is overkill, however and I have seen 8-10c difference between the non AVX and AVX versions of Prime with air. So, I think the apprehensions about using the newer versions of Prime would apply to those with more marginal cooling solutions.