Are these bad temps for stressing a 4790K with IBT?

[ScottyN91]

Are these abnormal temps for a 4790K at 4.7GHz 1.275V (fixed)? I'm currently using a newly installed Kraken X62 with performance pump and fan profiles.

 

I'm currently running the same test again but this time with the fan and pump at 100%, since the performance profile last time meant the the fans were only at 50% and the pump at 70% throughout, although I doubt much will change.

 

In the past I've tried 2 successive Corsair H110i GTX's with the same chip at 4.7GHz and was getting similar if not slightly higher temperatures (hitting 100C on some of the cores), even after re-applying paste and re-seating. This lead me to believe that either that cooler was inadequate for overclocking this chip, or I somehow had a faulty chip that was putting out way too much heat.

 

Edit: I forgot to mention that the liquid temps were 26C before and 32C after, and the motherboard/case temps were 24C before and 20C after. Ambient room temps were 20-21C throughout.

 

Edit 2: So I've finished the 2nd test and the max temps this time were 84C 89C 91C 83C, so only a 7-8C difference and they're still too hot IMO.

[i_build_nanosuits]

double check your installation your CPU is trottling hard and it's not normal...this is a fairly low voltage and temps bellow 75c should be expected.

something is wrong.

[Jdnoss1234]

I noticed that the burn in test puts causes very unrealistic temps. I tried it on my on my 4790K @ 4.6Ghz 1.30V and It will run my system straight to thermal throttle with a Thermaltake Water 3.0 360. On Prime95 v26.6 Small FFT test I will run at 78C all day long. I would recommend running a couple of other test software, but be aware don't run a prime95 newer that v26.6 as it will cause the same unrealistic temps. Aida64 has a 30 day free trial and you can try that as well.  

[ScottyN91]

On 4/15/2018 at 10:38 PM, i_build_nanosuits said:

double check your installation your CPU is trottling hard and it's not normal...this is a fairly low voltage and temps bellow 75c should be expected.

something is wrong.

Ok, I'll try re-applying thermal paste and re-seating, but as I mentioned, I've had similar results despite re-applying in the past so I'm not sure if that's the problem.

 

On 4/15/2018 at 10:50 PM, Jdnoss1234 said:

I noticed that the burn in test puts causes very unrealistic temps. I tried it on my on my 4790K @ 4.6Ghz 1.30V and It will run my system straight to thermal throttle with a Thermaltake Water 3.0 360. On Prime95 v26.6 Small FFT test I will run at 78C all day long. I would recommend running a couple of other test software, but be aware don't run a prime95 newer that v26.6 as it will cause the same unrealistic temps. Aida64 has a 30 day free trial and you can try that as well. 

Yeah, I always wondered if it was IBT causing the ridiculous temps, I'll take a look at those after I've re-installed the hardware.

 

Ok, so I've just finished the second run this time with Thermal Grizzly Kryonaut applied, and the difference was only 2-3C. Although I know that IBT is particularly stressful compared to P95 and AIDA64, is there anything else it could be, like my CPU is damaged somehow?

[FlatBrokeRacing]

my delidded 4790k @ 5ghz 1.35V runs low 90's on IBT, and will run low/mid 80's with Aida64, and will virtually never go above 75*C while doing heavy every day workloads with room temps in the 22*C mark with an H100iV2, i run fans/pump at max anything over 65C. delidding is what made all the difference 

[Froody129]

3 hours ago, FlatBrokeRacing said:

my delidded 4790k @ 5ghz 1.35V runs low 90's on IBT, and will run low/mid 80's with Aida64, and will virtually never go above 75*C while doing heavy every day workloads with room temps in the 22*C mark with an H100iV2, i run fans/pump at max anything over 65C. delidding is what made all the difference 

This.

 

Although better than the 4770k the 4790k still has terrible TIM and delidding will make the biggest difference here. I would think you could hit 5GHz since your voltage is pretty low for the clock 

[Ebony Falcon]

On 16/04/2018 at 7:44 AM, FlatBrokeRacing said:

my delidded 4790k @ 5ghz 1.35V runs low 90's on IBT, and will run low/mid 80's with Aida64, and will virtually never go above 75*C while doing heavy every day workloads with room temps in the 22*C mark with an H100iV2, i run fans/pump at max anything over 65C. delidding is what made all the difference 

And ur ok with it at 90c 

[Ace McPlane]

On 4/16/2018 at 12:30 AM, ScottyN91 said:

Are these abnormal temps for a 4790K at 4.7GHz 1.275V (fixed)? I'm currently using a newly installed Kraken X62 with performance pump and fan profiles.

 

I'm currently running the same test again but this time with the fan and pump at 100%, since the performance profile last time meant the the fans were only at 50% and the pump at 70% throughout, although I doubt much will change.

 

In the past I've tried 2 successive Corsair H110i GTX's with the same chip at 4.7GHz and was getting similar if not slightly higher temperatures (hitting 100C on some of the cores), even after re-applying paste and re-seating. This lead me to believe that either that cooler was inadequate for overclocking this chip, or I somehow had a faulty chip that was putting out way too much heat.

 

Edit: I forgot to mention that the liquid temps were 26C before and 32C after, and the motherboard/case temps were 24C before and 20C after. Ambient room temps were 20-21C throughout.

 

Edit 2: So I've finished the 2nd test and the max temps this time were 84C 89C 91C 83C, so only a 7-8C difference and they're still too hot IMO.

Actually, the 4790k is just a really, really hot chip.

Early hyper-threading you know.. that shit was hot.

 

1.275 is actually pretty high for this chip. Most people see that and are thinking that's relatively low.

Not for this chip it isn't.

A 4690k would run fine at 1.29 without getting that hot, but the 4790k is a different beast.

This is one of the #1 chips I recommend delidding as a result.

 

It's just a really hot chip and you're shooting for too high of an overclock.

Pull it back to 4.5 or 4.6 and lower that voltage until you stay under 80 degrees celcius..

Oh and don't use intel burn test.

Use Aida64.

 

[FlatBrokeRacing]

19 hours ago, Ebony Falcon said:

And ur ok with it at 90c 

90c under a complete and udder bullshit stress test, hell yes i'm fine with that.

 

daily usage even going on some very heavy workload is easily 20c less than that. 

 

You need to delid these chips in order to really get results temp wise, the TIM is junk and the lid height on the haswell/devilscanyon chips are really shit. I even push mine past what Intel says is max (1.3V is max daily voltage on haswell). Even if you are too scared to use Liquid metal, using something like MX4 for TIM will cut temps easily 10*C or better. 

[ScottyN91]

Ok, I'll consider delidding for the future, since I was trying for 4.8GHz originally, but even after pushing to 1.320V and getting 99C with 100% fan and pump, it still wasn't stable, so voltage (and heat) requirements seem to skyrocket after 4.6GHz at 1.200V (same as at stock), up to 1.280V (more stable than 1.275V) for 4.7GHz and over 1.320V for 4.8GHz+.

 

For now I think I'll stick to 4.7GHz as the voltage isn't too high and the temps should hopefully be ok when I switch to P95 for less intense stability testing.

 

So far in general usage and gaming with AC: Origins (which uses a fair amount of CPU), I've only seen a max of 67C (with 32C max liquid temp and 22C max ambient) so it seems to be ok when not running IBT.

[Ebony Falcon]

On 18/04/2018 at 5:28 AM, FlatBrokeRacing said:

90c under a complete and udder bullshit stress test, hell yes i'm fine with that.

 

daily usage even going on some very heavy workload is easily 20c less than that. 

 

You need to delid these chips in order to really get results temp wise, the TIM is junk and the lid height on the haswell/devilscanyon chips are really shit. I even push mine past what Intel says is max (1.3V is max daily voltage on haswell). Even if you are too scared to use Liquid metal, using something like MX4 for TIM will cut temps easily 10*C or better. 

 

[Mick Naughty]

Seems like accurate temps for subpar aio’s. 

[ScottyN91]

On 20/04/2018 at 1:48 AM, Mick Naughty said:

Seems like accurate temps for subpar aio’s. 

Have you got any suggestions for better AIOs that are 280mm or less?

[Mick Naughty]

1 hour ago, ScottyN91 said:

Have you got any suggestions for better AIOs that are 280mm or less?

Nope. I quit using them. Nexus has a good review if them l though. 

 

Delid or drop voltage. 

[airdeano]

why not try a more standardly used test like aida64 or p95 (1344 fft) or even intel XTU.

 

using the very high stress is like using AVX on p95 versions. this creates high thermal loads.