High Temps during stress testing and gaming.

[alitomr1979]

This is 80C peak while stress testing with Intel Extreme Tuning Utility. I run the test for 25 minutes.

When gaming with CPU demanding games like Battlefield 1 it goes into the low 70s.

My cpu is the i7 7700k with the Cryorig h7.

Idle temp is about +7C over room temp. The temp under the stress test is +52C above room temp, which is consistent to the performance to be expected from the H7 with the i7 6700k, according to most reviews online.

Even though this seems ok, given the max temp of the 7700k is 100C, my unit is being benchmarked only 733 in the intel extreme tuning utility, and in the cpu z app it is getting subpar results too. Could this be due to cooling problems or I just got an underperforming chip?

I also have only one ?DIMM rated 2133Mhz, and this causes some apps to show its speed at 1066Mhz. Could the RAM speed be affecting the CPU performance in the tests and in real life scenarios like gaming??

 

[LucaFire]

For the RAM problem, what is happening is that since it's DDR (Double data rate), they divide the speed in two (because of the double part) - this means that when it says 1066, it really means that the total is at 2133.

As for the CPU, have you tried reapplying the thermal paste? Indeed, some chips are just hotter than other chips of the same type. There could be some thermal throttling occurring if the chip is getting too hot.

 

[alitomr1979]

For the RAM problem, what is happening is that since it's DDR (Double data rate), they divide the speed in two (because of the double part) - this means that when it says 1066, it really means that the total is at 2133.

As for the CPU, have you tried reapplying the thermal paste? Indeed, some chips are just hotter than other chips of the same type. There could be some thermal throttling occurring if the chip is getting too hot.

Thanks, Lucafire. I applied the thermal paste that came with the Cryorig H7. Intel XTU is saying that no thermal throttle has occurred in all of my tests.

 

[LucaFire]

Considering you have a Cyrorig H7, what size case do you have? If it's in a very small space that could be contributing to it.

 

[LucaFire]

There is also the fact that with the new generation of Intel CPUs, they've been using relatively poor TIM (thermal paste) under the lid of the CPU.

Many have used a strategy called "delidding" (I personally would NOT endorse this for the common PC user) where you remove the lid of the CPU and remove the previous TIM and apply your own. Many of the people that have tried this found that there were reductions in temps by as much as 30C.

Again, I would not heavily discourage you from performing this procedure unless you are extremely experienced. Doing this will also void any warranties you have on the CPU.

 

[Crumpet 1]

There is also the fact that with the new generation of Intel CPUs, they've been using relatively poor TIM (thermal paste) under the lid of the CPU.

Many have used a strategy called "delidding" (I personally would NOT endorse this for the common PC user) where you remove the lid of the CPU and remove the previous TIM and apply your own. Many of the people that have tried this found that there were reductions in temps by as much as 30C.

Again, I would not heavily discourage you from performing this procedure unless you are extremely experienced. Doing this will also void any warranties you have on the CPU.

Intel saves what.. 2p per cpu, and the user loses out on temperatures or their warranty.

Good business practice!

But yes, I agree with LucaFire, make sure your cpu cooler has enough room to breathe, and if you're not satisfied by the results consider sending it back saying you are dissatisfied with the quality of your purchase if you are within a reasonable amount of time from your purchase. (be sure to read your receipts terms and conditions)

 

[alitomr1979]

Thank you for taking time to help! I ordered a noctua D14, with mx4 thermal paste. They are both supposedly great.

I have been getting the expected performance from my Cryorig H7 cooler and the 7700k. That pairing should get a 52-53C delta above room temperature under max load in XTU stress testing.

 

[The_Staplergun]

Was this for the stock clock settings?

 

[alitomr1979]

Was this for the stock clock settings?

Yes. The 770k is reaching 4.48Ghz, stock.

 

[The_Staplergun]

That IS pretty high for stock settings. You should see a pretty fair difference with the d14. Hopefully it works for you!

It's the unfortunate facts of the 7th gen architecture and poor TIM under the lid, they run excessively hot and require premium cooling.

 

[alitomr1979]

That is what I´ve read. As I said, a few reviewers have posted that the 7700k with the Cryorig H7 should peak with around 53C above room temperature, and that is what I am getting.

In all honesty, I dont see much of a point is overclocking. You get maybe 6-10% speed improvement and need a motherboard and cooling solution that is at least 30% above the total investment in CPU-motherboard-CPU Cooler.

For example, I am currently using MSI B150M mortar + 7700k + cryorig H7, thats 89+345+35 = 469USD. In order for me to get 10% (tops) speed improvement, I would need at least the Z270 Extreme4 + 7700K + Noctua NH-D14, and that is 150+ 345 + 65 = 560USD, which is a 20% increase on 469...for the 10% improvement which is very optimistic. Realistically you can expect a 6-7% speed improvement. Bad investment, if you ask me. I did it anyway, because I wanted a motherboard that gives me the option to do SLI when the time to upgrade comes...but I also think SLI is not a very good or good value upgrade.

 

[The_Staplergun]

It's 13-14% at 4.8 16% at 4.9 and 19% at 5.0ghz, but yeah. With the d14 you could get probably up to 4.8 if your voltages run low. Possibly 4.9 if you're super lucky.

Yeah d14 should do you well all around.

 

[alitomr1979]

It's 13-14% at 4.8 16% at 4.9 and 19% at 5.0ghz, but yeah. With the d14 you could get probably up to 4.8 if your voltages run low. Possibly 4.9 if you're super lucky.

Yeah d14 should do you well all around.

The actual performance of the CPU is not linear with the clock speed. You wont get 10% more performance out of a CPU that is OCd from 4Ghz to 4.4Ghz. When you run benchmarks and do actual measurements your results are going to be closer to that I said in my previous message.

 

[The_Staplergun]

Yeah there's plenty of factors, but in raw numbers that's what it shows. I know my mobo post screen shows how much % I've overclocked.

It's based on raw computing calculations, not much other factors here. There's cache speeds and voltages and all kinds of fun mess otherwise.

All good.