Temp too high with 5930K. Hardware and Clock issues?

Toggle sidebar Toggle sidebar
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630
I've built my new PC, and will be watercooling it once everything is tested and I know it's working properly. I'm currently using the reference Intel cooler on my CPU because I wanted to save money, and I wont be using it once I've tested parts.

I was running Prime95 last night as part of the Real Temp sensor test. I was hitting temps of 95 degrees, and 2 of my cores received a "LOG" status, indicating that they had hit thermal limit at one point, and were now being throttled. Once I saw that, I shut my system off and went to bed.

How concerned should I be? Is it just because I have a crappy cooler, or could something else be amiss?


Edit: I took off my cooler to redo the thermal paste, just in case. Hopefully it goes better. Didnt realise just how small the contact patch is for the cooler onto the CPU.

EDIT 2: Found out that clock speeds havent been running at default. Now the struggles is getting them to go to stock and stay there....
 
Sort by date Sort by votes
K

kunthakenthe

Admirable
May 20, 2014
2,682
0
7,160


the cooler is crappy and these haswell's run hotter than ever. just watercool it.
 
Upvote 0 Downvote
U

unksol

Splendid
Sep 12, 2011
7,060
0
27,960
The stock cooler is designed to cool it at stock. Obviously if you hit 95C at stock settings on a stock cooler something is wrong.

Did your motherboard over volt or over clock it? I suppose with prime 95 in a warm environment it could throttle. but its D@mn cold here lol
 
Upvote 0 Downvote
K

kunthakenthe

Admirable
May 20, 2014
2,682
0
7,160


lol hold on prime 95 on STOCK COOLER with one of the HOTTEST CHIPS FROM INTEL lol no. Thats your problem the stock cooler can barely handle it stock load much more a extreme benchmark, water cool then bench/OC.
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630
I was running everything at stock clock speeds, ambient temperature around 20C, and idle temp is around 35C. I know the stock cooler is never going to be amazing, but I wanted to make sure my components had no faults before I switched from air to water.

I had 1 BSOD a week ago while nothing was under load, so I'm trying to work out what the issue is. Memtest86+ showed no issues overnight. I havent run Prime 95 for longer than 20 mins because temps scared me off. Means I cant rule out the CPU :'(



EDIT: To elaborate with temps, I'm getting a 10 degree difference across cores. some sit at 85C, others at 95C. While reapplying my thermal paste, I decided to check how flat the CPU was. Seems like it's convexed, sitting about .5mm high in the center. I'm guessing that's one of my culprits?
 
Upvote 0 Downvote
frag06

frag06

Honorable
Mar 17, 2013
1,353
0
11,960


Are the fans running at full speed? Try manually setting them to max RPM.

Yes, if the heatsink isn't making full and even contact, that could possibly cause your different temp readings.

Prime95 is very hard on Haswell CPU's. Try a different program.
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630
all fans are running at max.

Current fan setup is 2 fans drawing air in from the front, 1 fan drawing air in from the back, towards the cpu. I have 1 fan on top venting air out. PSU is partitioned off.

What program would you suggest I use to make sure the CPU has no faults? Everywhere I look recommends Prime95.
 
Upvote 0 Downvote
frag06

frag06

Honorable
Mar 17, 2013
1,353
0
11,960


Sorry, I worded that poorly. I meant is the stock cooler running at max RPM?

I say give AIDA64 a try.

What water cooler are you using, a loop or an AIO? I would just put it under water and see how it performs. Plus, I would test with games as well, since you aren't likely to ever see these high temps when gaming.
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630


CPU fan is running at max.

I'm going to do a custom loop. I'm still waiting for some fittings to arrive, so I cant watercool it just yet.

Given the cpu is convexed, I'm considering lapping it once I know the CPU isn't faulty.


I'll give AIDA64 a go and see what happens there...
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630
Derp.

So after reinstalling the CPU and looking through the BIOS, I noticed that the mobo was running a XMP profile that was OCing the CPU to 3.8GHz. For reference, I'm using an ASUS x99 Deluxe mobo.

I'm just going to set the OC tuner to auto instead, with a target CPU frequency of 3.5GHz.
 
Upvote 0 Downvote
frag06

frag06

Honorable
Mar 17, 2013
1,353
0
11,960


That will do it. Auto OC usually jacks up the voltage to undesirable levels, too.
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630
I neglected to mention in OP that I was using Prime95 as part of the sensor test for Real Temp GT 3.70.

Tested again with Realtemp and Prime95, and results were better for a bit. The cores were all sitting at 87C till the last 5% of the heat up test, where they all spiked to 98C and were throttled. Needless to say I panicked and stopped the test. Not sure what caused it this time. May have to doublecheck my thermal paste application. I'm using Gelid Extreme.

When I did the test last night, hottest 2 cores were sitting at 90C for most of the test and then reached 95C.




Would the Auto OC explain my spike at the very end of temp testing?

Also, if I shouldnt use auto, should I use the other XMP profile? sets CPU to 3.75GHz. No idea how to manually set everything to "stock" in the BIOS because I'm a noob at most things BIOS.
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630
Got everything set to default, and disabled turbo mode. Had 2 cores running 10C hotter than the rest again, and all cores hit 100 degrees in the final part of the heat test. I'm going to put it down to it being daytime, and ambient temp is much higher than last night.

At least I now have my clock speeds at something sensible. Now to use AIDA64 for a bit and check stability.

Last musing for now, is that the temp difference is still high between cores. Really thinking I'm going to lap the CPU now. That's at least something I know can do properly.
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630
I'm starting to think that I will just wait till its cooled before I do stability tests, even if it means more work.

I tried a stability test with AIDA64 but I only see CPU, Cores 1-4 and Mobo for temps. max temps have hit 79 for CPU, cores are hitting 93 and Mobo is 32. Ambient temp is 22C.

The thing that really throws me is that despite bios being set to default and turbo mode turned off, clock speed is being recorded as 4.1GHz, and no throttling detected at any point. WTF??
 
Upvote 0 Downvote
frag06

frag06

Honorable
Mar 17, 2013
1,353
0
11,960


Try resetting the BIOS to default and see if that changes anything.
 
Upvote 0 Downvote
K

kunthakenthe

Admirable
May 20, 2014
2,682
0
7,160


i think somehow it still is overclocked, try resetting the bios as suggested or checked if turbo or any other overclocking features are on, also check if you accidentaly overclocked it.
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630
OK, finally some luck.

Whenever I set to defaults, the PC would lock up soon in to testing, even after disabling turbo mode [which wants to make it run at 4.1GHz instead of 3.5GHz]. For some reason my RAM was trying to run at 3333MHz even though it wasnt showing up as such until I tried to change settings away from default in BIOS :/

Anyway,after some searching, I also disabled "CPU C-States" and "Enhanced Intel Speedstep Technology" in the Power Management Configs, and was able to test on both AIDA64 and Prime95. Clock speed stayed at 3.5GHz, and it seems like RAM was behaving too.

Hottest temperature reached was 82C on Prime95, which is much better than before [24C ambient temp]. The spread of temps was much better too, with there only being a maximum of 4C difference across cores. Just when I though all was going to be fine though, the system locked up during the cool down phase of Real Temp's sensor test.

So, things are not perfect, but it's been progress.

I'm currently getting the last bits to flush my rads, and hopefully my fittings arrive this week so I can get this bitch underwater.


EDIT: I'm starting to want to give up. every time my PC locks up and needs to be reset, I lose whatever the bios was set to, and it goes back to wanting to run everything at 4.1GHz. Feeling in over my head now because I just cant make sense of it.
 
Upvote 0 Downvote
frag06

frag06

Honorable
Mar 17, 2013
1,353
0
11,960



It's sound like your board is the cause, so you may want to return it for a different one if possible. An Asus Rampage V if you want to go higher priced or something like the ASRock Fatal1ty X99X Killer if you want to go cheaper.

The Deluxe has some problems it seems. It actually even fried $1500 worth of equipment, (a 5960X and the board) in one test setup. No one knows the exact cause, but last I saw they were saying Asus was at fault. I would just return it and buy a different board if it is still giving you problems.
 
Upvote 0 Downvote
K

kunthakenthe

Admirable
May 20, 2014
2,682
0
7,160


you can try a new board, but whats happening now is that the ram(ddr4 new technology) is trying to clock itself a 3333mhz when the software isnt ready, highest stable right now is 2888 or 3000mhz. Check for updates or calll asus to see if they have a solution. Have you been saving at all. Maybe that test is causing the lock up, virus maybe check.
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630


My parts are as follows:

CPU: i7 5930K
MOBO: ASUS X99 Deluxe
GPU: Radeon R9 295X2
RAM: Corsair Vengeance LPX CMK16GX4M4A2666C16 16GB (4x4GB) 2666MHz DDR4
PSU: Seasonic 1200W Platinum

I've played a bit of Starcraft 2 and had no issues with maxed graphics, but that was before I started messing with BIOS settings. That means I was running one of the ASUS XMP profiles, where the ram was clocked to 2800MHz, and the CPU was running at 4.1GHz. It was very hot [hence starting the thread] but it was mostly stable. Only issue was 1 BSOD, but I cannot remember what I was doing at the time. I know I wasnt trying to bench or game though.

I got the last of my fittings today and was going to try watercooling it and seeing how I go.

Is there a way to check the mobo is faulty without getting a second board? I'd rather not have to buy a new board just to test.
 
Upvote 0 Downvote
frag06

frag06

Honorable
Mar 17, 2013
1,353
0
11,960



Not really. If you reset it to default and it is still auto overclocking everything, something isn't right.

Maybe try setting all of the settings that are giving you problems to manual. Set the CPU multiplier to x35 (sync the cores) and see if that fixes the problem. If not, try setting voltage to manual as well. Also, set your RAM to it's default timings and voltage.

Update the BIOS too, if you haven't already.
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630



I'll have another go at that. I just dont know what to set things to and am scared to mess around with them. I am using the latest BIOS.

Maybe I'm misinterpreting the wording, but the only "reset to default" option is "load optimised defaults" so maybe ASUS is guessing what settings I should be able to run, opposed to actual defaults? I'll have another look into how to reset BIOS just in case.
 
Upvote 0 Downvote
K

kunthakenthe

Admirable
May 20, 2014
2,682
0
7,160

good luck on your watercooling adventures, even when it does overclock you dont have to worry anymore temps should be around 20c on idle 40c max @ 4.4-4.5ghz.
 
Upvote 0 Downvote
A

Apoca1ypse

Distinguished
Dec 9, 2008
56
0
18,630




Here's hoping. I'm just finishing my radiator flush and am off to the store to buy more distilled water.


I just flashed the bios, and it seems like it wants to run on a premade profile. 4.1GHz CPU clock, and ram at 2800MHz.

I'm going to just leave it as is, and strip the PC down so I can start putting on waterblocks.

This is ending up being much more hassle than expected. May as well test with what my final cooling solution will be. If a part ends up being faulty, it wont seem like much more work now....
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom