High CPU temp GTA V with new cpu cooler?

[biffzinker]

it's a 6700K, isn't it notorious for overheating because of the IHS-DIE TIM?)

It's not the TIM so much as the sealant glue used to hold the IHS on that causes an increased gap the TIM has to fill in between the die, and heatspreader. Although when I de-lidded my 4790K the TIM layer on the die looked thinned out enough to not of been the cause for the temperature issue. So the quality of the TIM that Intel claims they improved on was still sub-par? Although the PCB from the 4790K to the 6700K is thinner, don't know if anyone remembers the issue of warped, and cracked substrate.

http://www.modders-inc.com/psa-cooler-mounts-can-damage-skylake-cpus/

 

[EsaT]

What TIM? Note some take several days and multiple heatup/cool down cycles to reach optimal cooling efficiency. While this is typically 2 - 3°C improvement, I have seen up to 5°C. Some TIMs take up to 200 hours to reach maximum efficiency. If the Grizzly Kryonaut shown in your specs, then curing time is not an issue.

Also Noctua's own TIM should be good from the start.

I don't think that there is a one to one relationship between ambient and CPU temp.

Unless airflow over cooler (=fan speed) is ramped up to lower cooling system's thermal resistance, every increase in ambient temperature goes to temperature of cooled component.
Even for constant heat load producing component...
What CPU actually isn't, with higher temperature increasing both leakage and power used by transistors to do their job.
End of the page has explanation straight from Intel:
http://www.madshrimps.be/articles/a...n-45nm-Performance-Overclocking-Power-Usage/6

Thats what I was thinking as well, especially considering that I'm using a 7-heatpipe high-end cooler.

Well, who knows how much of toothpaste there's under the heatspreader.
Isn't it rather well established that Intel doesn't want to endanger their excessive profit margin for making consistently flat heatspreaders and attaching them accurately?

You might not have to delid. Grab a straight edge and check your IHS and make sure its not convex. You might be able to get away with a quick lap job. Delidding isn't going to help you at all if you can spin your CPU on its IHS like a top.

At least slightly towards convex surface would make sure of good contact with heatsink in center closest to heat producing area.
Unlike in case of concave surface.
Though who knows if it's double concave IHS...

 

[P4-630]

Did some tweaking, the average CPU temp was around 62C with 21C ambient:

 

[advanced3]

There is really no reason to cool with air anymore, unless you're one of those hipsters who still think a huge, ugly heat sink looks good. Ill take the much better performance and looks of a AIO.

 

[TheMadDutchDude]

By that logic, quit cheaping out and go full custom...

Jeeeeeeeez.

 

[thebluebumblebee]

There is really no reason to cool with air anymore, unless you're one of those hipsters who still think a huge, ugly heat sink looks good. Ill take the much better performance and looks of a AIO.

Okay, so now that you know what a wack-a-mole feels like, back it up. Present your argument to support what you've said. Make sure to compare noise, performance and cost. Go.
P.S.: Bonus points for pointing out the one inherent advantage the air cooler has over the AIO.

 

[Caring1]

There is really no reason to cool with air anymore, unless you're one of those hipsters who still think a huge, ugly heat sink looks good. Ill take the much better performance and looks of a AIO.

I smell a troll with a fishing rod.

 

[freeagent]

I went from an H100, with two 120x38s to a Thermalright. I don't miss the H100 at all. Not once did I feel the H100 was better. Although, when I was sporting the H100 I did think it was better than my old D14.

Edit:

To be fair, I really hated that cooler. So big and obnoxious, and not any better than the ultra 120 extreme I was using before that. I literally sold it for fifty bucks a week after I paid a hundred for it. I felt bad for charging him that much. Even with stronger fans it was junk. Maybe I got a dud..

 

[Mac2580]

Seems normal, its summer here but still very early. Ambient must be around 20C.
Only played for 4 mins.
My PC is stock with boost off.

 

[P4-630]

Seems normal, its summer here but still very early. Ambient must be around 20C.
Only played for 4 mins.
My PC is stock with boost off.

Yeah lol , you only played for 4 minutes haha, when I startup GTA V the first 4 minutes in the game I see a CPU temp from high 40s - low 50s.

I have a Noctua NH-U12A.

Anyways, I have been tweaking with fans and now I see an average CPU temp of 57 degrees (19C ambient) with playing for an hour or so.

Also I play a modded GTA V with an average of 100 fps @ 1440p.

 

[Vario]

Looks like temperatures are on target, I typically sit around 55C CPU after an hour of GTAV/FiveM gameplay with 8600k @ 4.6 1.18V and Phanteks PHTC14PE and two big Thermalright TY143 Fans.
I have a 24*C room ambient.

 

[Mac2580]

Yeah lol , you only played for 4 minutes haha, when I startup GTA V the first 4 minutes in the game I see a CPU temp from high 40s - low 50s.

I have a Noctua NH-U12A.

Anyways, I have been tweaking with fans and now I see an average CPU temp of 57 degrees (19C ambient) with playing for an hour or so.

Also I play a modded GTA V with an average of 100 fps @ 1440p.

View attachment 121097

Yeah thats good to know, i am voluntarily GPU bottlenecked on both my pcs till GTA VI launches. Its expected but either way my temps dont actually rise over time as fan curve doesnt allow more really. As you can see they were very calm, all at 45pwm except cpu, on the standard bios fan curve. Ive noticed on my Ryzen it does rise slowly but surely over time but mot here. Noctuas are expensive right? They dont really sell them here. (I wouldnt consider shipping) And i know my 212 was dirt cheap, so you should actually expect better. Correct me if im wrong but even though my card is worse my vsync was off so our cpus were giving their all, regardless of (gpu bottlenecked) output.

 

[OneMoar]

>80c on air at 1.35v
then you can be concerned
untill then send it

 

[Vayra86]

There is really no reason to cool with air anymore, unless you're one of those hipsters who still think a huge, ugly heat sink looks good. Ill take the much better performance and looks of a AIO.

Yep, I love the sound of shitty pumps, tiny rads and meh fans along with a limited lifetime and high price. Its só 2019.

-signed Dark Rock Pro3

 

[P4-630]

Got some new fans installed and did some more tweaking, these are the results sofar @21°C ambient:

CPU temp around 55°C average during GTA V.

 

[delshay]

Yep, I love the sound of shitty pumps, tiny rads and meh fans along with a limited lifetime and high price. Its só 2019.

-signed Dark Rock Pro3

A user here on TPU said something along the lines of

"I prefer a slow death of a air cooler, than a sudden death of a leaking AIO liquid cooler"

Whoever said that, has my vote.

EDIT: I now have two of those Noctua NF-A12X25 PWM fans. They are awesome especially when it comes to noise lever. Mounted on Phantek PH-TC14CS

 

[Vario]

Air cooler will run forever, you can use it on an infinite number of builds as long as Intel sticks with the current hole spacing, its been a decade now so probably. Dual ball bearing fans can last a decade. Its just a solid purchase when you can run it for so long. I don't need to think about it ever. It sits in my windowless case and just simply works. No pumps to fail, nothing. Even if the fans fail the unit will still work fine.

Yeah thats good to know, i am voluntarily GPU bottlenecked on both my pcs till GTA VI launches. Its expected but either way my temps dont actually rise over time as fan curve doesnt allow more really. As you can see they were very calm, all at 45pwm except cpu, on the standard bios fan curve. Ive noticed on my Ryzen it does rise slowly but surely over time but mot here. Noctuas are expensive right? They dont really sell them here. (I wouldnt consider shipping) And i know my 212 was dirt cheap, so you should actually expect better. Correct me if im wrong but even though my card is worse my vsync was off so our cpus were giving their all, regardless of (gpu bottlenecked) output.

212 gets you about 80% of the performance of the top end air for about 50% of the price, there are diminishing returns on air. Especially if your case has a poor airflow design or inadequate fans.

 

[John Naylor]

1. I have not found a reason to delid any Intel CPU since the 3500k / 3700k ... all CPUs since have hit voltage wall long before temperature wall.

2. There are very good $99 coolers (Cryorig R1, Noctia NH-D15, Phantes PH-TS14PE) but the $45 Scythe not only tops those but every CLC type AIO on the market (at time of testing) but one and that once is, like almost all others, extremely loud. The numbers are the numbers ... If you want to spend 2 -3 times as much for an AIO that is going to be much louder and cool almost as good as the Scythe, that's a user's perogative. At 55% load, I would want to leave the room:
https://martinsliquidlab.wordpress.com/2013/03/12/swiftech-h220-vs-corsair-h100i-noise-testing/

Look at the temps and noise differences below ... the $120 Cryorig A40 gets the same temp result as the $45 Scythe of 69C but it is 1.75 times as loud:
https://tpucdn.com/reviews/Scythe/Fuma/images/temp_oc_aida64.png
https://tpucdn.com/reviews/Scythe/Fuma/images/fan_noise_100.png

Yes, liquid cooling is more effective cooling when we are talking about custom loops... AIO coolers with aluminum rads need extreme rpm fans to get close to cooling the aforementioned air coolers. Can they do better than these coolers... yes... with 2200 rpm fans ... but so can your air cooler with 2200 rpm fans. But do you want to spend 3 times the price of a Scythe for say a Cryorig A40 to get something that is 1.8 times as loud ? Again, the test results speak for themselves. If ya ever decide to go w/ an AIO, would recommend an all copper unit like the EK or Swiftech offerings.

3. I agree, I would not put in a CC type AIO if they were free ... way too loud, (galvanic corrosion science experiment) ,, I have no issue with OLC types will all copper componentry and 1.0 + gpm pumps like EK and Swiftech
h ttps://www.youtube.com/watch?reload=9&v=a_r4tWpMHHI&feature=youtu.be&t=8
https://martinsliquidlab.wordpress.com/2012/01/24/corrosion-explored/

4. If you take Noctua fans off a Noctua cooler and replace with Phanteks, temps drop by 6C at same rpm. However, this is a new cooler with new fans.
https://www.overclockersclub.com/reviews/phenteks_f140/3.htm

5. It's all about delta T ... that is, it is consistent with changes in ambient temps.

6. You can accelerate TIM curing by thermal cycling. While I find P95 useless for stability testing, I do use it to accelerate TIM curing ... cycling it up to 80+ and back to room temps 4 or 5 times. TIMs have different curing rates ...AS5 for example can take 7-8 weeks (w/ Shin Etsu and P95 cycling, I find only marginal improvement after about 2 hours). All TIMs have cure time ...
https://www.techpowerup.com/forums/...-vs-conductonaut-liquid-metal-73-w-mk.229167/

7. You want 50% (dusty air intake filter restriction) more intakes than exhaust to insure that you are not sucking in hot exhaust from PSU and GFX card thru the rear grilles and vented slot covers. Rule of thumb for "quiet system" ... one 140mm case fan for every 75 - 100 watts of component wattage (50-75 for 120mm).

8. I'm pretty sure that the Asus Z170 Pro Gaming comes equipped with AI Suite which includes FanXpert utility ... you can install AI Suite to get FanXpert and then disable other components of the suite if you do not wish to use them. Use FanXpert to set up fan curves for each MoBo header.

9. I have 2600ks thru 8700ks running in low - mid 70s at 4.6 - 4.8 GHz ... no delidding.

10. Connect Cooler to CPU or CPU_OPT ... usually 2nd fan has no rpm monitor... alternately cooler may have splitter in which case plug into CPU header. Suggest putting three intake fans on CHA_1 and 2 exhaust fans on CHA_2 ... make sire combined amperage on any channel is < 1 amp... most case fans < 1500 rpm are around 0.14 amps.

11. Recommend testing on a standard load .. we use RoG Real Bench (8 minute benchmark) and HWiNFO for monitoring. Note the 'default OC settings for voltage that some MoBos assign are usua;lly too high ... they basically say "what voltage will be needed for the worst OCing specimen we ever made".

http://dlcdnmkt.asus.com/rog/RealBench_v2.56.zip

12. CPUs and coolers are make curved on purpose, IHS matching cooler with the concave / convex thing. Some manufacturers void their warranty if lapped as they believe it affects performance.... Most testing however shows no difference from lapping unless you have a low budget cooler with bad surface.

13. Your HWiNFO numbers look fine, again suggest RoG real Bench which is an application based rather than an unrelaistic sythetic load.

 

[Komshija]

What risk? some have 5 year warranty, they very rarely leak

But they do leak. Maybe one in 15 or so, but still air cooling is a lot safer. My former cooworker had a leak on his Corsair H100 which destroyed his system. From time to time water leaks on AIO coolers and even custom cooling systems happen, some people even share their misfortunes on the web. Even more, high-performance air coolers will outperform any 120 mm AIO and majority of 240 AIO coolers, even a few 280 mm coolers.

It's very strange to have such high temps with good air cooler like OP's Noctua. I would say that either CPU lid is bent, damaged or cooler base/support has some manufacturing flaw. I have the same CPU but paired with Be Quiet Shadow Rock Slim. Highest ever temperature I had was 82°C on a hot summer day - CPU was overclocked and running Prime 95 28.9, but even when overclocked I never saw temperatures above 73°C during normal operation and heavy gaming (no synthetic benchmarks).

 

[John Naylor]

There is really no reason to cool with air anymore, unless you're one of those hipsters who still think a huge, ugly heat sink looks good. Ill take the much better performance and looks of a AIO.

Air cooled systems are low noise but still, there's noise. I build custom loops not because I need to for overclocking, but because if I can perceive ANY noise, this is unacceptable. To effectively cool with water and obtain low noise, you need several things:

a) Thermally efficient materials like copper ... CLCs use aluminum rads
b) Low speed fans .... not workable with thermally inefficient aluminum rads
c) Flow rates of 1.0+ gpm ... for example, H100i is 0.11 gpm

Much better performance ... ? let's use the ever popular Hyper 212 as a universal benchmark along with the venerable Cryorig R1

https://tpucdn.com/reviews/Corsair/Hydro_Series_H105/images/CPU_OC_typical_b.gif
https://tpucdn.com/reviews/Scythe/Fuma/images/temp_oc_aida64.png

The $120 Corsair H105 you have chosen manages 10C better than the Hyper 212
The $45 Scythe Fuma manages 11C better than the Hyper 212

The $120 Corsair H105 you have chosen manages 3C better than the Cryorig R1 Universal
The $45 Scythe Fuma manages 3C better than the Cryorig R1 Universal

So, after paying almost 3 times the price, you "almost did as well" as the air cooler

https://tpucdn.com/reviews/Scythe/Fuma/images/fan_noise_100.png
https://tpucdn.com/reviews/Corsair/Hydro_Series_H105/images/fan_noise100.gif

The $120 Corsair H105 you have chosen is 8C louder than the Hyper 212
The $45 Scythe Fuma is 4C lower than the Hyper 212
The combined 12 dba means that your H105 is 2.3 times louder than the Fuma

The $120 Corsair H105 you have chosen is 9C louder than the Cryorig R1 Universal
The $45 Scythe Fuma is 3C lower than the Cryorig R1 Universal
The combined 12 dba means that your H105 is 2.3 times louder than the Fuma

So tell us again why spending almost 2.6 times as much to get "almost equal" thermal performance while being 2.3 times louder equated to "better performance"

 

[Deleted member 24505]

Each to their own, basically. I prefer the cleaner look with a AIO, and it seems so do a lot of commercially available Systems now.

Air cooled systems are low noise but still, there's noise. I build custom loops not because I need to for overclocking, but because if I can perceive ANY noise, this is unacceptable. To effectively cool with water and obtain low noise, you need several things:

a) Thermally efficient materials like copper ... CLCs use aluminum rads
b) Low speed fans .... not workable with thermally inefficient aluminum rads
c) Flow rates of 1.0+ gpm ... for example, H100i is 0.11 gpm

Much better performance ... ? let's use the ever popular Hyper 212 as a universal benchmark along with the venerable Cryorig R1

https://tpucdn.com/reviews/Corsair/Hydro_Series_H105/images/CPU_OC_typical_b.gif
https://tpucdn.com/reviews/Scythe/Fuma/images/temp_oc_aida64.png

The $120 Corsair H105 you have chosen manages 10C better than the Hyper 212
The $45 Scythe Fuma manages 11C better than the Hyper 212

The $120 Corsair H105 you have chosen manages 3C better than the Cryorig R1 Universal
The $45 Scythe Fuma manages 3C better than the Cryorig R1 Universal

So, after paying almost 3 times the price, you "almost did as well" as the air cooler

https://tpucdn.com/reviews/Scythe/Fuma/images/fan_noise_100.png
https://tpucdn.com/reviews/Corsair/Hydro_Series_H105/images/fan_noise100.gif

The $120 Corsair H105 you have chosen is 8C louder than the Hyper 212
The $45 Scythe Fuma is 4C lower than the Hyper 212
The combined 12 dba means that your H105 is 2.3 times louder than the Fuma

The $120 Corsair H105 you have chosen is 9C louder than the Cryorig R1 Universal
The $45 Scythe Fuma is 3C lower than the Cryorig R1 Universal
The combined 12 dba means that your H105 is 2.3 times louder than the Fuma

So tell us again why spending almost 2.6 times as much to get "almost equal" thermal performance while being 2.3 times louder equated to "better performance"

Quite a few high performance air coolers are nearer £80-99 now, which is nowhere the 3 times cheaper Than a aio you are stating. As I said each to their own.

 

[delshay]

Another way to look at it, Air vs Liquid cooled. If you had to leave your house for 6 hours with your computer left on doing some task, which one would you trust the most.

A. Air
or
B. Liquid Cool.

I'm no expert, but I do know electricity & liquid are not the best of friends.

 

[Deleted member 24505]

Another way to look it, Air vs Liquid cooled. If you had to leave your house for 6 hours with your computer left on doing some task, which one would you trust the most.

A. Air
or
B. Liquid Cool.

I'm no expert, but I do know electricity & liquid are not the best of friends.

Mine is on all day, even if i leave the house. if you are so worried about it breaking, don't turn it on.

 

[John Naylor]

But they do leak. Maybe one in 15 or so, but still air cooling is a lot safer. My former cooworker had a leak on his Corsair H100 which destroyed his system. From time to time water leaks on AIO coolers and even custom cooling systems happen, some people even share their misfortunes on the web. Even more, high-performance air coolers will outperform any 120 mm AIO and majority of 240 AIO coolers, even a few 280 mm coolers.

It happens .

https://forums.tomshardware.com/threads/my-corsair-h60-exploded.326466/

https://www.reddit.com/r/buildapc/comments/4pxjp2

https://forums.tomshardware.com/threads/aio-liquid-cooler-bursts-3000-in-drain.2611486/
https://forums.anandtech.com/threads/aio-cooler-leak.2487981/

Not sure why this happens... doesn't seem as if the weak pumps could even build up enough pressure but something is happening. Wise to put in a pressure relief valve in custom loops o to of reservoir. Can do this with a OLC type AIO (EK and Swiftech). On the Swiftech for example there is a side rad port. Take out the plug and install a M x M threaded connector to an elbow. Put the pressure relief valve, on the top of elbow or if needed use a extention to get it right up under the top grille and then the PRV. You can even take off the PRV and use as a fill or bleed port.

 

[delshay]

It happens .

https://forums.tomshardware.com/threads/my-corsair-h60-exploded.326466/

https://www.reddit.com/r/buildapc/comments/4pxjp2

https://forums.tomshardware.com/threads/aio-liquid-cooler-bursts-3000-in-drain.2611486/
https://forums.anandtech.com/threads/aio-cooler-leak.2487981/

Not sure why this happens... doesn't seem as if the weak pumps could even build up enough pressure but something is happening. Wise to put in a pressure relief valve in custom loops o to of reservoir. Can do this with a OLC type AIO (EK and Swiftech). On the Swiftech for example there is a side rad port. Take out the plug and install a M x M threaded connector to an elbow. Put the pressure relief valve, on the top of elbow or if needed use a extention to get it right up under the top grille and then the PRV. You can even take off the PRV and use as a fill or bleed port.

Question:

Does the pressure relief valve need to be set-up, ie calibrated. You can't just add one in & hope it will work. Or am I wrong. Please correct me on this.