new i9-9900k build - over 100c

ococ

Limp Gawd
Joined
Oct 20, 2006
Messages
326
Bear with me as it's been many years since I've messed with overclocking.

I just put together a new build with the following parts:
i9-9900k
Aorus Master Z390 (firmware F8b)
Corsair Vengeance 32GB 3200mhz ram
Seasonic 850W PSU
Corsair H80vi2 (from old system) (noctua thermal paste)
Silverstone FT02 case (from old system)
old SSDs and HDDs (from old system)
Radeon RX560 (from old system)



with stock settings (vcore auto shows 1.20v) I'm seeing idle temps around 25-27c which appears to be normal. When I run Intel XTU benchmark, it's peaking at 81c. But when I run Prime95 small ffts, coretemp quickly jumps into the 90s and even hit 100 before I shut it off.

Is the H80vi2 just not good enough? and say you're not overclocking, how do people run this with stock fans?

Thanks!


**Update...new problem**

I don't know what just happened. I tried to do some overclocking following Gigabyte's guide - https://www.gigabyte.com/FileUpload/Global/multimedia/2/file/525/946.pdf
I set the vcore to 1.3...then 1.31 all the way up to 1.35, a 5ghz and everytime it rebooted during a benchmark test. Resigning to not be able to get 5ghz at 1.35 (and temps were letting me get any higher voltage), I decided to back down to 4.8ghz at 1.25. This time the crash seemed harder and the computer wouldn't boot to bios. After a few power cycles, it reverted itself back to stock firmware (F4) and reset to default. I then updated it back to F8b, but now my lowest temps are 31c. 5c higher than before. Also now when I run the XTU benchmark, it's jumping up to 100c vs 75c before. I was ok with not being able to overclock, but now I can't even run stock.

Any ideas?
 
Last edited:
Remount the cooler before anything else. Check the contact and the thermal paste spread. It should be enough at minimum for stock settings at least.
 
Will try that, but wouldn't that issue manifest itself at stock temps? Maybe I should install the corsair software to make sure the cooler is actually working.
 
took my like 5 times to get my heatsink correctly applied with thermal paste.....it happens......just redo it...were talking 20+degrees easy..paste has to be installed just perfect:)
 
I just remounted the cpu cooler with fresh paste. Also i tightened down the cooler more. The stock idle temps are within 1-2 deg (25-26c) now. When I run Intel XTU, it went from topping at 81 down to topping at 75, so some improvement. But P95 is still hitting 100 after only running for ~10 secs and the XTU is showing Thermal Throttling.

Couple of question. The H80vi2 is mounted to the mobo on CPU_Fan. There's a splitter that comes out from the cooler as well, which is what I plugged the fans into. So all three (corsair pump, 2 fans) are all going into CPU_fan. When I'm running the tests, it doesn't seem like the fans are going any faster, which is ~1000rpm. What's the right configuration for these fans.

Secondly...it seems like I'm still pretty far from having <90 temps with stock cpu settings running P95. Is that unrealistic?
 
I just looked up some other tests and I don't think I'm too far off. One guy is running stock speeds with a Noctua D15 - idle temps 33, I'm at 26. AIDA 64 temps at 72c, I'm at 65c. wPrime 1024 temps at 72, I'm at 54c.
 
I just looked up some other tests and I don't think I'm too far off. One guy is running stock speeds with a Noctua D15 - idle temps 33, I'm at 26. AIDA 64 temps at 72c, I'm at 65c. wPrime 1024 temps at 72, I'm at 54c.
probably just need to tweak the pump and fan speed at this point....i just disable qfan all together on mine since its dead quite even at full speed lol..but everyone hears different
 
I would definitely spin the fans up and see how it goes. I get the itch between the shoulder blades at 100*
 
Hook your PUMP to a secondary fan header and set any control (in bios) to disabled to ensure it's running at 100%. Then set control to auto/standard for the main fans. they should definitely ramp up normally if setup 'standard' in bios.
 
I don't know what just happened. I tried to do some overclocking following Gigabyte's guide - https://www.gigabyte.com/FileUpload/Global/multimedia/2/file/525/946.pdf
I set the vcore to 1.3...then 1.31 all the way up to 1.35, a 5ghz and everytime it rebooted during a benchmark test. Resigning to not be able to get 5ghz at 1.35 (and temps were letting me get any higher voltage), I decided to back down to 4.8ghz at 1.25. This time the crash seemed harder and the computer wouldn't boot to bios. After a few power cycles, it reverted itself back to stock firmware (F4) and reset to default. I then updated it back to F8b, but now my lowest temps are 31c. 5c higher than before. Also now when I run the XTU benchmark, it's jumping up to 100c vs 75c before. I was ok with not being able to overclock, but now I can't even run stock.

Any ideas?
 
is there a pump only header on your mb? or a pump/fan header? if so make sure to connect the pump to that. sometimes if the pump is connected to a fan header it doesnt allow it to run at full speed. also if there is a pump header go into your bios and make sure its set to run at 100%. i would also make sure all of the case fans and your rad fans are running full blast while your ocing just to be sure you are getting the most out of your rad/case cooling. those rad fans have very good static pressure ratings but you have to run them at 100% to get the best performance out of them. if none of that helps im thinking your pump is probably dying. time to call corsair for a replacement.
 
Ok, I messed around with the fans and while the idle temps are still a bit higher than before ~28-30, the temps under load are ~80 for XTU. All of this is running stock speeds.

With the Aorus Master I have the following: CPU _FAN, CPU_OPT, SYS_FAN1, SYS_FAN2, SYS_FAN3, SYS_FAN4, SYS_FAN5 PUMP, SYS_FAN6 PUMP. The H80vi2 has two connectors coming out of it, one being male 3 pin which is what powers it, and the other splitting into two 3 (or 4)pin females. This is where the cooler fans plug into H80vi2. I have all this plugged into CPU_FAN. I also have 3 larger case fans that I've plugged into SYS Fan 1, 3 and 4.

Not sure if this is the optimal fan set up, but it's working for now, but it still won't do P95 at stock speeds. While I've given up hope of overclocking to 5ghz, i'm still pretty happy with the jump from an old 2600k to the 9900k for Lightroom and Premiere Pro export times.

I may do another cooler reseating this weekend, but this is good for now.
 
I tried this. I took the main power cable from the H80vi2 which has the 2 fans also pigtailed into it, and plugged it into SYS_FAN6 PUMP. When I did that it throttled the pump as well as the fan speeds down to around 750 rpms? It also had some weird behavior. Coretemps showed the idle temps around 27, so not much change there. But when I ran XTU, the temps maxed out in the 50s, and the benchmark wouldn't finish. I didn't see it give any Thermal throttling errors but it just wouldn't finish. So I decided to change it back. Now I have the pump in CPU1, took the pump fans off the pigtail and plugged them into CPU_OPT and SYSFan1. Pump is running at MAX while the other two are set to CPU temps. This seems like the best setup I've found so far and the XTU benchmark is completing with a max temp of around 75-80.

btw, I also ran some conversions in Premiere Pro and seeing max temps at 65 so I should be fine with this config. Question is do I want to bother eeking out some more gains.
 
You did set your pump to 100%/Full in Bios ?? in smart fan 5 make sure the fan is set to Full or 100%..NOT AUTO..Auto will not give enough power to run the AIO Pump under stress..
 
Yes, I've made several changes at this point. Full speed on both the pump and the 2 fans, but it's apparent that the h80i isn't strong enough for the 9900k heat.
 
Yes, I've made several changes at this point. Full speed on both the pump and the 2 fans, but it's apparent that the h80i isn't strong enough for the 9900k heat.

That doesn't seem right to me... I know the 9900k puts out plenty of heat, but a properly running H80i with a good fan (or fans, if in push-pull) should be at least enough.
 
That doesn't seem right to me... I know the 9900k puts out plenty of heat, but a properly running H80i with a good fan (or fans, if in push-pull) should be at least enough.

I would’ve agreed with you 3 weeks ago. But after messing with this thing for too many hours and getting inconsistent results, I’m going to RMA the CPU, then swap out the cooler next.

I’ve tried different bios. All fans and pumps set to full speeds. One one reboot I’m getting 28/81 (idle/peak) on other reboots I’m getting 38/100. Zero bios changes, same tests. This is using loaded optimized defaults with full fans.

Maybe it’s not the cooler, but at stock clocks it’s peaking around 81, with any overclocked voltage I’d easily be in the mid 90s so I’m thinking a Noctua 15s would do me better.
 
I would’ve agreed with you 3 weeks ago. But after messing with this thing for too many hours and getting inconsistent results, I’m going to RMA the CPU, then swap out the cooler next.

I’ve tried different bios. All fans and pumps set to full speeds. One one reboot I’m getting 28/81 (idle/peak) on other reboots I’m getting 38/100. Zero bios changes, same tests. This is using loaded optimized defaults with full fans.

Maybe it’s not the cooler, but at stock clocks it’s peaking around 81, with any overclocked voltage I’d easily be in the mid 90s so I’m thinking a Noctua 15s would do me better.

Why would you RMA the CPU?? The mobo seems to be more of a suspect than any other component.

[QUOTE**Update...new problem**

I don't know what just happened. I tried to do some overclocking following Gigabyte's guide - https://www.gigabyte.com/FileUpload/Global/multimedia/2/file/525/946.pdf
I set the vcore to 1.3...then 1.31 all the way up to 1.35, a 5ghz and everytime it rebooted during a benchmark test. Resigning to not be able to get 5ghz at 1.35 (and temps were letting me get any higher voltage), I decided to back down to 4.8ghz at 1.25. This time the crash seemed harder and the computer wouldn't boot to bios. After a few power cycles, it reverted itself back to stock firmware (F4) and reset to default. I then updated it back to F8b, but now my lowest temps are 31c. 5c higher than before. Also now when I run the XTU benchmark, it's jumping up to 100c vs 75c before. I was ok with not being able to overclock, but now I can't even run stock.][/QUOTE]

I would reset the BIOS again and maybe even re-flash it.
 
A few drive by suggestions

Ensure the pump is actually running, Corsair says you can connect it to the CPU fan header, but I don't know if that can cause problems.
Load optimized default in BIOS after updating to remove old crud and snafus from the update.
Try feeling around on the cooler to figure out if you even get the heat away from the CPU.
Check if Enhanced Multi-Core Performance is enabled/AUTO (M.I.T. -> Advanced frequency settings), this is technically an overclock.

Not something I can pinpoint, but I remember my Gigabyte board was a bit odd with the CPU voltage when it was set to AUTO, try keeping an eye on the vcore in XTU.

(Edited suggestion on the pump)
 
Last edited:
Yes I’ve run with the fans connected to the pump which then goes into a cpu fan header, as well as the fans plugged in separately, all running at 100%.

Yes, load optimized defaults (with maxes fans) was setup. LOD has MCE on by default.
 
MCE screws with a lot of settings and can result in very high temperatures, despite everything looks like it is running "stock"
 
I keep seeing p95, what version/settings are you using? Could have sworn most people/sites are recommending a non current build for stress testing.
 
Back
Top