5930K Temps don't make sense...

[Solaris]

Bought a 5930K back at the beginning of the year... It's on a custom water loop with plenty of radiator capacity.

But even without being overclocked, the temperatures are spiking up very quickly.

This is from 10 standard iterations of Intel burn test, and three very high iterations. I was monitoring radiator temperature at the same time and they only rose by about 3 degrees over the whole test.

Spoiler

The line at the start of the graph that is highest is the package temperature, which starts at 41 degrees idle and then gets lost with the rest of the lines.

The rest of the lines are individual core temperatures, which start out between 30-35 and then jump to around 65-69 at peak load.
So what's happening?? I'd almost be led to believe that the loop wasn't pulling any heat away from the CPU, but the maximum temperatures never seem to rise...

I'm using an EK-Supremacy EVO - Full Nickel waterblock and have re-seated and reapplied thermal paste a number of times, have tried rotating it 90 degrees, and have the correct jet plate and insert for this CPU inserted.

The temperature of waterblock, and the area around the CPU never get hot to the touch.

So what do you think? Did I strike out in the silicon lottery and get a CPU with either bad inbuilt temperature probes, or a poor adhesion to the IHS?

 

[tyler_jrb]

They do seem quite high especially for a stock 5930k, mine was in the 30s-40s maxed out on stock. I wouldnt reccommend intel burn, use Aida64 or ROGbench. The 6 cores dont like it intel burn or Prime95. Stresses them way way over anything your cpu will experience.

41c idle is very hot, mine idles at 26-32c on 4.7ghz with a custom loop but does rise a few degrees after a few hours gaming because of my Gpu's, and ambient temperature does have impact aswell. The Cpu maxes out at around 60c (4.5ghz) and under 70c (4.7ghz).

Other than that depends on other specifications aswell.

The temperatures spiking is quite odd though, try different monitoring software like Coretemp as it could be as easy as a software issue. Aswell as Rogbench or Aida64.

 

[r7slayer]

I'm not sure what it could be as my first thought would be that you didn't seat your EK block correctly on the CPU but you've redone it a number of times? So im sure you've got it seated correctly. On a okayish custom water loop you shouldn't be hitting between 65 and 70 degrees more like top end 50s then maybe after some duration it will top out near 65. Intel burn test is quite overkill for cpu stress testing but they made it like that so if it passes its a good indication of temps and stability though you can pass one test and fail another. Aida is better tbh, and you'll never see those temps from intel burn test in real workloads or gaming. EVER lol.

 

[Silent_Scone]

Does sound like a contact issue.

 

[5960xNvidiaGamer]

Those temps definately don't seem right, at 4.6ghz my 5960x at idle is under 40 at 1.28v

 

[Scougar]

Are either of the mating surfaces concave or convex? Also, is there enough pressure?

I have found wide variations in temps between software, I don't really trust any of it tbh.

 

[traveyb]

Are either of the mating surfaces concave or convex? Also, is there enough pressure?

I have found wide variations in temps between software, I don't really trust any of it tbh.

This, try taking the block off and look at the surface, either it's not sitting on tight enough of its curved not creating pressure in the centre

 

[Solaris]

Sorry to necro the thread, but after finally disassembling my loop to do some upgrades... I figured out what was wrong...

Where I'd disassembled the block to change the internals, I'd mounted the base 90 degrees out of whack. So instead of the water going in and flowing through the channels, it was going in and I dunno what...

In short, I'm a total idiot...