Overclock.net › Forums › Cooling › Water Cooling › H80i - Very high temps
New Posts  All Forums:Forum Nav:

H80i - Very high temps

post #1 of 2
Thread Starter 

I got myself an H80i the other day as I had just upgraded to Haswell (4670k) from my old Lynnfield i5 (760) that had been running at stock all this time. Figured I'd do some semi-decent OC in the 4.3 - 4.5 range, but I've been having major heatproblems. So just now I decided to set everything back to default in BIOS and see temps while running the new AVX2 Linpack. I set manual vcore to 1.2 however as I didn't want the auto-settings to rush vcore up too much during the test, everything else was left at default however. In the following screenshots you can see my results, and that I didn't even finish the test as the temps were running so high I didn't feel safe letting it go on:

I'm using 2 Gentle Typhoon AP-15's on the radiator. The "CPU fan" value you see in HWMonitor on the screenshot is actually the top exhaust-fan as it's running on the CPU_OPT slot. Corsair Link shows both radiator fans running at 1680-1689rpm. I have the two default radiator fans as front intakes (no idea if that was a good plan or not, but my old original Fractal Design silent fans that came with the case were just about dead). Besides those 2 I have 2 Nexus PWM fans as well, one as bottom intake and the earlier mentioned top exhaust. The h80i is set up as rear exhaust. In my amateur mind, this should be a decent setup for PAP and should be able to do better than this.

The two things I can think of:
1) Pump poorly mounted / TIM not properly applied. When I reseated the pump after switching the fans from stock to GT's I applied AS5 TIM using the preferred method under 'Instructions' on the AS website.
2) 3(!) of the radiator threads on the side of the radiator where I connect it to the case broke pretty much instantly upon first install, even as I was being very careful and screwing them in by hand. I 'solved' this by using tape on the screws so that they would get in there, but it's obviously not a perfect solution. I can "wiggle" the radiator a little bit in one direction, as in I can push on the side of it to get it into a tilted position. It does not appear to move in any shape while running however.

As you can see the idle temps are decent, and in tests like Prime95 it appears to be only the short FFT's that push it up really high. That is probably completely normal though, but I haven't been OCing in forever so figured I'd mention it as well.

Hopefully this is clear enough as I'm not quite sure how to best explain this in english, and I'd be thankful for any help / tips. Thanks!

Edit: As the image appears to be rather small even when clicked on, here is a link to it in full size: http://i.imgur.com/p3VSBQf.jpg
post #2 of 2
Thread Starter 
Small update incase this is helpful in any way (sorry if this counts as a bump, in the future I'll just edit the OP instead if so!)

Changed the lininput_xeon64 file like so after reading a post on the Asus ROG forums:
Sample Intel(R) Optimized LINPACK Benchmark data file (lininput_xeon64)
Intel(R) Optimized LINPACK Benchmark data
1       # number of tests
20000   # problem sizes
20000   # leading dimensions
5       # times to run a test
4       # alignment values (in KBytes)

Ran it like this with 1.1V at 40 multi (not sure why I bothered changing the multi at all really as the point was to try with lower vcore, but oh well). Even at this low volt, I'd max out at 84C on the cores during the test. Now I do realize that linpack with AVX2 pushes things to the extreme of the extreme, but 84C on 1.1V just seems way off considering the temps I've seen from people with similar/worse coolers at higher voltages.

Results from the test:
Warning: Spoiler! (Click to show)
Intel(R) Optimized LINPACK Benchmark data

Current date/time: Thu Jul 25 02:56:04 2013

CPU frequency: 4.005 GHz
Number of CPUs: 1
Number of cores: 4
Number of threads: 4

Parameters are set to:

Number of tests: 1
Number of equations to solve (problem size) : 20000
Leading dimension of array : 20000
Number of trials to run : 5
Data alignment value (in Kbytes) : 4

Maximum memory requested that can be used=3200404096, at the size=20000

=================== Timing linear equation system solver ===================

Size LDA Align. Time(s) GFlops Residual Residual(norm) Check
20000 20000 4 28.829 185.0258 3.504283e-010 3.102058e-002 pass
20000 20000 4 28.615 186.4110 3.504283e-010 3.102058e-002 pass
20000 20000 4 28.521 187.0273 3.504283e-010 3.102058e-002 pass
20000 20000 4 28.246 188.8487 3.504283e-010 3.102058e-002 pass
20000 20000 4 28.597 186.5298 3.504283e-010 3.102058e-002 pass

Performance Summary (GFlops)

Size LDA Align. Average Maximal
20000 20000 4 186.7685 188.8487

Residual checks PASSED

End of tests


Edit: Some more testing shows that 1.25V (43 mult, 34 uncore, SVID Control and CPU Spread Spectrum disabled, LLC 8, Initial CPU Input Volt 1.7V, Eventual CPU Input Volt 1.65V) goes to 100C+ instantly in new Linpack, with temps of well over 90C in IBT and Prime95 Small FFT's.
Edited by katatoni - 7/25/13 at 8:28am
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Water Cooling
Overclock.net › Forums › Cooling › Water Cooling › H80i - Very high temps