BF2S Forums abnormal cpu temps... reapply thermalpaste?

TheEternalPessimist wrote:

Every Q66 rig ive built (of which there have been about 5 now) has had almost identical temps to mine (+5* or so max) so there's something wrong here.

E6600 has TDP of 65w like most of the C2D's do. Q6600 G0 has TDP of 95w. And when you increase vcore on the quad and OC it to 3.33ghz, it hawt.Increasing vcore on dual core had much less impact on temps than with quad.

EDIT: to make some comparison between overclocked E6750 and Q6600 G0 for example, the load temps are 20c higher and idle temps are 15c higher with the quad than with overclocked E6750. TDP is at least on par with the B3, maybe even a bit more thanks to higher vcore of 1.35v that I have.

kylef wrote:

TheEternalPessimist wrote:

Every Q66 rig ive built (of which there have been about 5 now) has had almost identical temps to mine (+5* or so max) so there's something wrong here.

It could simply be the fact that Panzer has significantly overclocked his chip.

That is correct. you hardly notice difference at stock settings but when you OC both, the quad's heatload is a lot higher. My quad had 50-55c load temps at stock at zalman in silent mode. but when OC'ed it is 70c.

Last edited by GC_PaNzerFIN (2008-03-31 06:00:09)

TheEternalPessimist wrote:

OK to make it simpler, the 5 rigs with Q66s in have all been clocked to at 3.3-3.7Ghz with v cores between 1.4 and 1.55 and they all stay a hell of a lot cooler than 70*, 1 had a CNPS9700, 2 had Thermalirght 120 Extremes passive and one had the Tr 120 extreme with a fan on, last had an Ultima-90 coz it was a small case, hence it only made 3.3 with what i'd call acceptable temps.

read this.

The advantages of dual-core Core 2 Duo E6850 over the quad-core Core 2 Quad Q6600 look as follows:

    * 25% higher clock speed ensuring higher performance in applications that aren’t optimized for multi-threading;
    * Higher bus frequency that increases the memory subsystem performance. Peak bandwidth of the bus between the processor and memory for CPUs with 1333MHz bus is 10.7GB/s vs. 8.5GB/s by CPUs with 1067MHz bus.
    * 46% lower heat dissipation that allows using relatively simple and inexpensive cooling solutions, even during overclocking.

The advantages of the quad-core Core 2 Quad Q6600 over dual-core Core 2 Duo E6850:

    * Twice as many computational cores that ensure higher performance in multi-threaded applications.
    * Twice as large L2 cache memory. It is especially crucial for Windows Vista users as this operating system can distribute computational threads intellectually over the cores with shared or exclusive L2 cache.

http://www.xbitlabs.com/articles/cpu/di … 600_2.html

TheEternalPessimist wrote:

Yeh ok fine, we're ignoring my e6600 if you hadn't noticed, that doesnt change the fact your CPU is running massively hotter than it should be does it?

Q66 on a CNPS9700 at 3.3Ghz should hit 60*c tops.

I've bult a rig with that cooler and that CPU and ran it at 3.4 on 1.5v core and it only hit 60*.

70* it's going to start to burn out.

it is less than 10w dude! Q6600 B3 is 105w and Q6600 G0 is 95w!

from vr-zone...

Last edited by GC_PaNzerFIN (2008-03-31 06:17:04)

kylef wrote:

TheEternalPessimist wrote:

Every Q66 rig ive built (of which there have been about 5 now) has had almost identical temps to mine (+5* or so max) so there's something wrong here.

It could simply be the fact that Panzer has significantly overclocked his chip.

It wouldn't get that hot from that OC. It isn't *that* big.

TheEternalPessimist wrote:

ghettoperson wrote:

kylef wrote:

It could simply be the fact that Panzer has significantly overclocked his chip.

It wouldn't get that hot from that OC. It isn't *that* big.

Exactly, 3.3 is a very modest OC, that chip with that cooler shouldn't get anywhere near 70*, a point that seems to be very hard to get across lol.

you might have misunderstood me completely. I'm talking about CORE TEMPS for individual cores not the CPU overall temp.

TheEternalPessimist wrote:

What are you on about now? they're all G0 and all cooler than yours.

Stop arguing and go fix your cooler lol.

I'm beginning to think theres some sort on confusion here, you do know I'm talking about rigs that I myself have built don't you?


And now you're posting benchs of QX6850s?

Plus it's 300Mhz faster and uses 0.1v more...

EDIT: I find it hard to believe that my chip using the same voltage and is (marginally) faster is supposed to run at 30+ degrees less than yours. They aren't that different.

Last edited by ghettoperson (2008-03-31 06:35:42)

Your cores are also only running at 60* which is a long way from the 70* u posted back on page 1, 60 is reasonable, 70 would be fucking aweful.


EDIT: I'm guessing that, core temp is bugged or your CPU hasn't kicked out of SpeedStep.

Last edited by TheEternalPessimist (2008-03-31 06:48:52)

That's just about acceptable then I guess, I'd still be fiddling with it but I know what a pain in the ass those coolers can be.

Check the edit above BTW, is that a bug with Core Temp or is it actually still running with a 6x multi coz of speedstep?

Last edited by TheEternalPessimist (2008-03-31 06:52:09)

TheEternalPessimist wrote:

Every Q66 rig ive built (of which there have been about 5 now) has had almost identical temps to mine (+5* or so max) so there's something wrong here.

Use real temp to measure temps. Some processors have bad dts.

TheEternalPessimist wrote:

I use a variety of measuring software, Everest, Speedfan, Core Temp, they're all the same. You need to remember building custom PCs is basically my second income, I know what I'm doing and make quite a lot of cash doing it, all through word of mouth.

And the programs above all measure the same temps right? Dunno bout the 65nm processors but for 45nm you need to use Real temp in order to view temps correctly.

.Sup wrote:

TheEternalPessimist wrote:

I use a variety of measuring software, Everest, Speedfan, Core Temp, they're all the same. You need to remember building custom PCs is basically my second income, I know what I'm doing and make quite a lot of cash doing it, all through word of mouth.

And the programs above all measure the same temps right? Dunno bout the 65nm processors but for 45nm you need to use Real temp in order to view temps correctly.

Yeah they all did, real temp is reporting the same temps too (I'm home now and running Orthos).