A good OC?

[AD-15]

hello,

I've just done my first ever overclock of my C2D E6600.

I'm at 355MHz FSB, x9 multiplier, 1.36V Vcore, and have tried Orthos for about 15 min, with a max temp of about 44C.

Should I try getting to a lower Vcore voltage?


It feels awesome getting an extra 0.8GHz out of your CPU without any cost except a 10C increase in temperature.

[staffsMike]

Lower Vcore's are always nice yeah

It will show instability far sooner than high vcore anyway.

So thats nigh on 3.2GHz, on what cooler? Certainly good temps so far.

[AD-15]

hello,

I'm using one of these:

http://scan.co.uk/Product.aspx?WebProductId=505971

A Zalman 9700NT, with some Arctic Silver 5 compound.

By the way, for how long should I run orthos? Are there any other tests I should do?

Thanks!

[staffsMike]

OCCT is a good one. It's quite quick and is a good general indication of stability.

Orthos is up to you really. 24 hours is a bit overkill I find but if you want to do that then go for it. I've ever gone as far as 6 hours though.

[cpemma]

44C (and the 34C at stock) looks surprisingly low for a heavily-stressed quad temperature.

What monitor software are you using?

[staffsMike]

E6600 It's a dual core.

[Thorsson]

The question about which software, and also which measure (is it Core or Die?) are still relevant.

[Webby]

24 hours of orthos should give you a high level of confidence that the system is stable personally I have never had it fail after an hour has passed but to be totally sure I have left it for 24 especially if I plan to run at that speed for extended periods.

[Blitzen]

44C (and the 34C at stock) looks surprisingly low for a heavily-stressed quad temperature.

What monitor software are you using?

E6600 isnt a Quad

[AD-15]

hello,

I was using the Abit uGuru utility. I will try using the Intel TAT ASAP, to check if it gives the same readings.

I left Orthos on for four hours yesterday, but didn't come across any problems.

I won't be running this speed all the time though, only for when I use MS flight simulator, as it is incredibly CPU intensive. I have saved overclocked and non overclocked profiles in the BIOS, so it's easy to switch back.

[cpemma]

hello,

I was using the Abit uGuru utility.

I would check using Coretemp or Speedfan.

44C (and the 34C at stock) also looks surprisingly low for a heavily-stressed dual temperature.

And keep on overclocking, it's what these processors are made for. Otherwise why buy the decent cooler?

[AD-15]

hello,

I've just tried using coretemp, and it reports 5C higher than abit uGuru. Whilst uGuru maxes out at about 46C, coretemp maxes out at about 51C.

Which should I trust?

[Webby]

You shouldn't trust either! The sensors have not been calibrated so they are not to be trusted equally the sensor was developed to protect the CPU from overheating so is more accurate in the higher temperature ranges.

This is not supposed to be a rant but the very fact that two or more programs which are reporting information from the same source disagree shows you how subjective it all is.

If you want to you can attempt to calibrate your sensors Real Temp will allow you to do this (to an extent, not real calibration but an attempt to remove some of the error) http://www.techpowerup.com/realtemp/ the is a link there to the calibration page which tells you what you would need to do, try it if you want to

Personally I just use Everest and take the readings with a pinch of salt using them more as estimates rather than exact figures since my CPU readings never exceed 50 I don't expect I will be having any problems any time soon.

Oh and if you are measuring temps to make sure your not damaging your CPU then take the most pessimistic view i.e. the highest reading

[Thorsson]

Totally agree with Webby. It's a point I keep making. Although these programs use the same on chip sensor, they make certain assumptions (like the value of TJMax) in reporting the temps that they do.

If they report close to ambient case temp after 10 minutes of zero CPU use then you can probably trust them at low CPU use. Under full load you should see the die temp a few degrees cooler than the cores. If that's the case then that measure is probably showing reasonably accurate temps.

[Blitzen]

I would check using Coretemp or Speedfan.

44C (and the 34C at stock) also looks surprisingly low for a heavily-stressed dual temperature.

And keep on overclocking, it's what these processors are made for. Otherwise why buy the decent cooler?

eh?
UGuru is certainly more accurate than Speedfan. In fact, touching the core and having a rough guess is more accurate than Speedfan.

[Thorsson]

Not used it but I believe the latest Speedfan gives the same temps as Core Temp & Everest (most of the time anyway). Doesn't mean they're right of course...