Run away train! 100C temps... on 7600k

[SPL Tech]

7600K @ 4.8 GHz 1.31 vcore running a 7-year-old Evo 212 HSF. I already delided the CPU. I was getting as high as 100C in game and around 90C on average. I replaced the paste with Thermal Grizzly since tests seem to show it's the best on the market. That dropped the temps a bit but I'm still seeing 75 - 85C while gaming with highs in the low 90s every few min and a maximum peak temp of about 95-100C on the hottest core across a gaming session. I tried maxing out all my case and CPU fans. That helps a tad and brings the temps in the 70 - 80C range, but I still see peaks in the 90s. So I dont know what the deal is. Maybe I need to replace the liquid metal on the CPU die itself? I have had this CPU for about 2.5 years now. I did clean the HSF so it's free of all dust. That dident do anything at all.

Also, for whatever reason the #0 core gets a LOT hotter than the #1 core while gaming. It may peak out at 100C while the #1 core will peak at 80C although Task Manager shows the CPU utilization is pretty close to max on all cores..

Regardless, I did order a Corsair H115i liquid cooler so that should help... a lot... hopefully. It's a 280mm radiator so that should be much better.

 

[Zerileous]

I was going to say, the 212 is limited with OC capability. Is this a new problem that developed or have the temps been like this for a while? New setup or did something change, or were you running with OK temps that now degraded. The high core-core delta indeed indicates a problem under the IHS or with mounting.

 

[EarthDog]

Wondering if the delid did it? Are you running bare die or did you put the IHS back on? If the latter....do it again. I'm guessing the internal paste ismt doing the job.

 

[trents]

I would reapply the LM. Make sure you apply it both to the underside of the lid and the die. There is some anecdotal evidence out there that LM tends to pump out over time and lose good contact.

That said, the fact that when you ramp up your case fans you see a significant temp drop suggests to me that you have a case ventilation problem.

 

[EarthDog]

Case fans may be a part of it, but something clearly isn't right there. I would also reapply under the hood as some cores are 80C, others hitting 100C. If you've repeated the block and reapplied paste, that leaves only what is under the IHS that needs tweaked.

 

[SPL Tech]

Changing the speed on the case fans only helps slightly. I am running the HSF on the IHS--no bare install. I did reapply the TIM and that helped a bit. I don't know what my temps used to be as I haven't checked in over 18 months, but I recall them being lower. I'll try to reapply the LM TIM and see if that helps.

 

[EarthDog]

I am running the HSF on the IHS--no bare install.

THen it seems like you don't have enough pressure on it...which, why the heck are you running naked? You're going to destroy that chip without a delid die guard of some sort... not worth it unless you are benchmarking competitively..

Throw some of the recommended paste between the IHS and die and slap that bad boy back together.

 

[trents]

"no bare install." I think he's saying he's not running naked.

 

[SPL Tech]

THen it seems like you don't have enough pressure on it...which, why the heck are you running naked? You're going to destroy that chip without a delid die guard of some sort... not worth it unless you are benchmarking competitively..

Throw some of the recommended paste between the IHS and die and slap that bad boy back together.

As in it's not a direct mount IHS.

CPU Die - LM TIM - IHS - Standard TIM - HSF

Like that.

 

[EarthDog]

Holy... I need my glasses cleaned. Lol

Reapply the middle tim as was said initially.

 

[SPL Tech]

maybe it's time for a 9700k or something. I have been holding out for the 10x series since it doesent look like the 9700k is really that much more powerful than the 7600k.

 

[Zerileous]

Well not in terms of single thread performance. You would get a soldered IHS, but you've already de-lidded your CPU, so unless you're just tired of dealing with this and don't want to re-apply the LM on your 7600K, that shouldn't be a factor. You would also be required to purchase a new motherboard.

 

[trents]

Yeah, that's the thing. A 9xxx CPU would require you to purchase a new motherboard and then you'd be stuck with dead end technology as the 10th gen Intel's will use a different socket. https://wccftech.com/intel-lga-1200-socket-confirmed-for-upcoming-intel-10th-generation-cpus/

 

[SPL Tech]

Yeah, that's the thing. A 9xxx CPU would require you to purchase a new motherboard and then you'd be stuck with dead end technology as the 10th gen Intel's will use a different socket. https://wccftech.com/intel-lga-1200-socket-confirmed-for-upcoming-intel-10th-generation-cpus/

Are you kidding me, another version of the 14nm? This is Intel's 4th version of the same CPU. They should be on gen 2 of 10nm by now. What a load of crap. Are they going to call this 14nm++++ now?

 

[trents]

They have laptop 10th generation out already but some thermal issues with it I understand. Yeah, it its truly amazing how the CPU war has changed and so fast. AMD has risen from the ashes. I read the other day that that AMD sold 40% more CPUs than Intel in ?the last year? Intel is in trouble. They are struggling to get out a viable 10nm chip and AMD already has had a 7nm CPU out for about a year. Intel is currently surviving on their leftover partner commitment contract fumes.

 

[EarthDog]

Intel will be back in the game. If AMD could be second for as long as they were and come back...not worried about intel.

Besides, its not the size of the wave (process) but the motion of the ocean (architecture). Look at nvidia and amd gpus... 12nm vs 7...

Intel has rode this arch and process long enough. 2021 and truly new chips should bring back more equal core/thread counts and pricing...better power too.

 

[Zerileous]

Indeed. While I'm not a huge fan of waiting to upgrade, since there is always something bigger and better around the corner, I doubt either team has a chipset that will be useful beyond the end of this year. Intel is going to require a new chipset to refresh the same old stuff, with 2 more cores crammed in, and AMD has already stated that they will be supporting AM4 through the end of 2020. That means possibly one more generation of CPUs on AM4, but we don't know for sure. Also we're probably going to see DDR5 coming in 2021.

There's a lot of time left in this year so don't despair if something ends up being really wrong with your overclock or you find yourself needing 6+ threads. But I wouldn't think of this as a convenient time to upgrade.

 

[SPL Tech]

Well I installed my Corsair 115i cooler and that thing is amazing. It's a night and day difference. Temps dropped in the high 30s idle and 65 - 70 while gaming using a game that maxes out all of my cores. It was basically a 20C drop. This means I should be able to get 4.9 or maybe even 5.0 GHz out of the processor. For sure liquid cooling is the way to go. Air doesent even compare and with these all in one coolers, liquid cooling is just as easy as air. No custom crap that's going to leak everywhere. Corsair's stuff is great. I want to get one of their PSUs now so I can get a PSU that doesent run the fan when the computer is idle.

 

[Zerileous]

To be fair to air, thing's have come a long way compared to the 212. The best air coolers will hang with your 115i, but they will be similar in cost and pose their own fitment and compatibility challenges.

 

[trents]

Ditto. As we pointed out earlier, the hyper 212 is not actually "the best" air cooler for that CPU. Not by a long shot. If you hung a Noctual NH-D15 on that CPU you would see similar temps to what you are seeing with you AIO water cooler.