Core i7-2600K overclocked temperatures

Soldato
Joined
16 Jul 2004
Posts
14,075
Hi,

I recently built a Core i7-2600K / Asus P8Z68-V system that I've overclocked, and I am wondering about temperatures.

The CPU is interfaced by Noctua NT-H1 to a Noctua NH-D14 with 2x Akasa Viper fans. The air from the heatsink blows directly towards the rear exhaust vent, which has an Akasa Apache assisting the airflow. The top of the case has an Akasa Apache venting air. The front of the case has an Akasa Viper drawing in cool air. The Vipers are all on PWM from the CPU header (via a splitter), and the Apache fans are being duty cycled with voltage regulation based on CPU temp. I've set them all to come on full at a CPU case temperature of 50C.

The CPU is running at 4.5 GHz by a x45 multiplier and 100 MHz BCLK. Load Line Calibration is set to "High" and voltage regulation is on offset mode with +0.01v dialed in. Before I figured out offset voltages, I first set a fixed voltage of 1.32v and encountered a BSOD during stability testing. An offset of +0.01v seems to keep it at a minimum of 1.32v under heavy Prime95 load. Everything else is default.

I have been running Prime95 on large FFTs for an hour and my CPU case temperature (or whatever is reported by the motherboard) has risen to between 56 and 59C. Using Core Temp, I can see that typically Core #0 is ~61C, Core #1 ~68C, Core #2 66C and Core #3 64C. Peak temperatures are Core #0 64C, Core #1 71C, Core #2 69C and Core #3 67C. Power draw is around 97 to 102 watts and voltage hovers around 1.32v to 1.328v to 1.336v.

Do those temperatures seem normal? They 'feel' a little high, but I'm far from an expert!
 
Last edited:
They seem normal to me. I personally wouldnt want to go above 85c on a SB cpu so 71c is fine.

If you really think you should be seeing lower temps try reapplying the paste and reseating the HS.
 
I did that earlier today, but I also switched the TIM from AS5 to NT-H1, so I'm unsure how much of the temperature drop I saw (2-3C) is the uncured AS5 vs. the no-cure-needed NT-H1 and how much is better 'installation'. I tinted the CPU case and heatsink per the AS instructions, but I think I overdid it and had too much paste left. The most comprehensive review of TIMs puts the Noctua just 0.9C behind a fully cured AS5, so I have gone with the Noctua for ease of setup.

http://benchmarkreviews.com/index.php?option=com_content&task=view&id=150&Itemid=62

Prime has been running for 2 hours now and temperatures are:

CPU 56C
Core #0 62C
Core #1 67C
Core #2 66C
Core #3 63C

Tj. Max is 98C so I seem quite far away from that; over 30C away in all cases. I might be able to find stability at a +0.000v offset which should help temperatures.

I've just seen a lot of voltage/temperature quotes around the net that are lower, but I suppose thinking about I should treat these with suspicion if the exact benchmark and measurement type isn't listed. I'd be curious to know what other people are getting running the same test with the same CPU using the same per-core temperature measurements.
 
Last edited:
Back
Top Bottom