Overclocked my Retail E8400...WOW!

[Xvys]

Thanks for the info, unclewebb. I will use that calculation and get an accurate reading of the cores.

But I decided to exchange the cpu, which my local retailer did without question. So I now have a new E8400 installed, from the exact same batch as my old one, an early week 740. The good news is both cores now show reasonable idle readings. At the same 4005 Mhz overclock as before, one core idles rock solid at 38C, the other bounces around from 33C-36C. Under Prime load, one core went up to 44C, the other to 50C?...a little higher than the reading of my old "bad" processor, but at least then the cores were equal under load. No such luck now.

No big deal I guess. If the CoreTemp readings are correct, it is not overheating and this imbalance could be due to sloppy heatsink installation...or perhaps more likely the sensors in the cpu do not read low temperatures well.

vanclas wrote:

My problem is temp

Idle: 43-44 with my Thermalright Ultra 120

It's normal ?

These cpu's are so new it is hard to say what is normal. At the same 4.0GHz overclock my cores read 38/33C at idle, but I'm not sure if you have the Extreme heatsink like mine, which may account for 2C or so. But I would say that if under full load your core temps don't exceed 54C, then all is well!

[yenclas]

Thanks for the info, unclewebb. I will use that calculation and get an accurate reading of the cores.

But I decided to exchange the cpu, which my local retailer did without question. So I now have a new E8400 installed, from the exact same batch as my old one, an early week 740. The good news is both cores now show reasonable idle readings. At the same 4005 Mhz overclock as before, one core idles rock solid at 38C, the other bounces around from 33C-36C. Under Prime load, one core went up to 44C, the other to 50C?...a little higher than the reading of my old "bad" processor, but at least then the cores were equal under load. No such luck now.

No big deal I guess. If the CoreTemp readings are correct, it is not overheating and this imbalance could be due to sloppy IHS installation...or perhaps more likely the sensors in the cpu do not read low temperatures well.

vanclas wrote:


These cpu's are so new it is hard to say what is normal. At the same 4.0GHz overclock my cores read 38/33C at idle, but I'm not sure if you have the Extreme heatsink like mine, which may account for 2C or so. But I would say that if under full load your core temps don't exceed 54C, then all is well!

In full load with orthos, with 4,0Ghz and 1,3575 vcore mi temps are 58-59ºC

I have a normal TR Ultra 120.

Is it dangerous ?

Why my processor show high temps ?

[Xvys]

Those temps do seem a little on the high side, vanclas, but not abnormally high. I believe I read that the maximum recommended core temps of these Wolfdale's is 64C. There was also a comment in one of the threads that 54C is the maximum. Not sure on that one. The previous generation Conroe cpu's had a maximum core temp of 65C, but many people routinely overclocked them to much higher temps. These processors have a built-in throttle down feature (at perhaps 120C as was pointed out) to protect the cpu from overheating. It was almost unheard of to burn out a Conroe, they would just shutdown if overheated.

These Wolfdale processors have the same throttle-down technology and I doubt you will be hurting your processors with core temperatures even in the 60C range.

I had no problem with heating my old E6300 to 70C cores under load, and it ran 24/7 for 18 months and still works perfectly...overclocking to 533FSB x 7 on both old my vanilla P5B and current P5K Deluxe. I would say that 65C is the maximum I would heat up these E8400 cores.

[stasio]

According to INTEL max.temperature for E8400 is 72.4C.
Refer HERE

[unclewebb]

According to INTEL max.temperature for E8400 is 72.4C.

Just keep in mind that those Intel published maximum numbers have nothing to do with the temperature number that CoreTemp is returning. The Intel number is measured on top of the heat spreader at the geometric center and is intended for heatsink developers. The reason Intel went to on die digital thermal sensors is so you can operate your processor at a higher temperature without hurting it. In the previous P4 era, processors would start to throttle too soon.

With a core processor as long as it isn't throttling then it is running within spec. Unless your heatsink falls off you should never have a problem with heat.

[Xvys]

Just got my replacement cpu all setup, it seems a bit more stable than my previous one. I didn't try to optimize the voltage or go for the maxiumum run yet, but so far so good!
------------------------------------------------
Asus P5K Deluxe / E8400 / 2G Crucial 10th Anniv. / ATI 1950pro / Ultra120 Extreme/ FSP FX600-E

[Ledge]

Man I am seeing some nice OC's! I hope the 680i likes the E8400 later this week???

[unclewebb]

Xvys: Nice OC. Not bad for a day of work!

Just wondering what type of thermal paste are you using and what's your method of applying it? I've always had really good success with spreading it out in a thin rectangular layer on the IHS directly over top of where the two cores are located.

http://regmedia.co.uk/2006/07/25/intel_conroe_1.jpg

[Xvys]

I got the Tuniq TX-2, which I believe is the newest mx-2 product and it was the cheapest at $3.99 for 3.5g at NCIX, which has many stores in my local area.

On a previous installation I put it on fairly small amount, then spread it out as even as possible with an exato blade. But when I removed the cpu it seemed there were too many bare spots. On my next install, I laid on the goop thicker in 3 main plops, but it was too much and a fair amount squeezed out the edges when I secured the heatsink. Not so good.

On my current install I went for a medium amount, with 3 vertical lines of tx-2 and 2 horizontal ones, but not spreading it out. But i'm sure it got spread out fairly evenly when I screwed down the heatsink.

I am at 4275Mhz now (475 x 9) with idle core temps of 38C. I Primed for about 45 minutes and core temps got up to about 60C...which is about as toasty has I want for those tiny circuits! Unfortunately I forgot to Prime overnight, but will try again tonight.

But I would say that less Mx-2 is better, perhaps 3 grains of rice and allow the heatsink pressure to spread it out evenly, rather than trying to scrape it level.

[unclewebb]

I'm a regular at NCIX as well. Can't beat deliveries to my door, 1000 km away, in under 24 hours.

I slightly modify what they recommend for applying AS5 to a dual core cpu:
http://www.arcticsilver.com/pdf/appi..._dual_wcap.pdf

I draw the line and then move it around into a rectangular area so it covers both cores. I think most of the heat transfer is happening close to the center so might as well concentrate on there. I'm sure paste manufacturers like your multi line approach!

People need to worry less about the temperature their core processors run at. With the amount of crazy overclocking that goes on if these things were that feeble, more people would be wearing them around their neck.

I found that when overclocking you will lose stability and have random reboots at full Prime load long before you ever reach a temperature high enough to hurt them. As long as you are Prime stable for a few hours you shouldn't have anything to worry about in normal use.

Here's how my E6400 ran with the Intel heatsink and fan:



Even after a year of abuse it still runs rock solid stable. These E8400 stories have got me tempted to try an upgrade though.

Now that Intel has got a year of experience manufacturing core processors they've decided to raise the TjMax rating for these new 45nm chips. TjMax is the maximum temperature that these processors are guaranteed to safely run at without any long term problems. Definitely no worries at 60C.

[Power5]

Wow, cant believe you got 3.5 with only 1.2v.

You don't see any problems coming from two different core temps? I would think there is a performance difference between the cores if one is 10c hotter.

This is IF the cores ARE actually different temps, not just a sensor error.

[Ace-a-Rue]

Wow, cant believe you got 3.5 with only 1.2v.

You don't see any problems coming from two different core temps? I would think there is a performance difference between the cores if one is 10c hotter.

This is IF the cores ARE actually different temps, not just a sensor error.

look at his VID...very high!...it is the reason he is overclocking so high with literally low vcore...i can't hold 4200-4200 mhz without sending my board PWM temps into the stratosphere with very high temps (100% load)...my VID is 0.9625v...i can hold 3ghz (stock speed) with 1.03 vcore...if i push the vcore to about 1.45v, the PWM temps zoom right into the red after about 2 minutes of 100% load on all four cores.

[unclewebb]

Timeout here. Power5 that was a picture of an older version of CPUz that used to misreport core voltage. I need approximately 1.45 volts to be Prime stable at 3.5 GHz. The temp on both cores was equal at 75C. Tjunction in the pic is what CoreTemp now correctly refers to as TjMax or the maximum temperature before throttling begins which for the original Conroe processors was 85C.

Just trying to show that there is no need to worry when you see a big temperature number as long as things are running stable. The new E8400 has even more headroom than the previous generation, Conroe, desktop processors.

If you are crazy enough you can run them right on past their rated 85C maximum. The digital thermal sensor is a 7 bit number so after it gets to zero it wraps around and becomes 1111111 binary which is 127. SpeedFan and TAT do the calculation of TjMax - 127 or 85 - 127 and comes up with -42. When these programs start reporting negative numbers your processor is definitely getting HOT!

You can see from the graph that it continued to get hotter and kept right on running. When I put the heatsink back on things cooled down again and everything was fine. I think it hit an honest 97C without a nuclear meltdown.

[Xvys]

yep, these are tough processors. Conroes were as tough as they came. Hopefully the Penryn's will be equally as robust.

[DarkjeThe2nd]

Mine shows both cores always the same within a degree.
I even get the same temps on both sensors even if I load one core and idle the other one, dunno what to think of that?

checked the msr on both cores and that also shows the same value in both...

must be borked :| I fear?

[Damien]

when i went outside with my setup last night, one core showed 44c the other 26c on idle, then at load both were the same at 4.5ghz. weird chips.

[Ace-a-Rue]

not weird chips...poor quality in the DTS sensors....i had similar problem with my 1st 9650.

[sbinh]

I finally have my hand on E8400 on P5K Dlx board.
Unfortunately, system could NOT boot even if the board has latest BIOS (0809).

It booted and ran flawlessly with E6600 and E6400.

What should I do now? Please give me some advice.

[sofarfrome]

I believe there is a more recent bios update available on one of Asus's ftp sites (do a google search for it). Hopefully you still have either an E6600 or a E6400 as you will need to boot up withone of those and flash the bios. I know it works on that board as I have seen people running that combo. I run one on the Premium version but that is a different bios.

[sbinh]

On Asus site, 0809 is the latest BIOS version for P5K Dlx.
If you happen to know any other BIOS verison later than that floating around, please let me know. Thanks.

[jcniest5]

I finally have my hand on E8400 on P5K Dlx board.
Unfortunately, system could NOT boot even if the board has latest BIOS (0809).

It booted and ran flawlessly with E6600 and E6400.

What should I do now? Please give me some advice.

Your CPU may be DOA. I have the exact combo and mine boots up right away without a single issue.

[Leeghoofd]

Man I am seeing some nice OC's! I hope the 680i likes the E8400 later this week???

You got yours going already on 680i ? I got my friends Striker to mess with, till now Q6600 runs fab on it with 1503 bios, but I will use this board with E8400

[TheGoat Eater]

my decent clocker is in my sig - Got 4.6GHz with my MARS board and pretty basic watercooling (2x120rad) and I am looking forward to DICE runs

[Grnfinger]