E8400 Stuck Temp = RMA?

[Penley]

Hey folks,

First off, thanks for the help figuring out what parts to get! Now that I finally got my system up and running I am concerned about my E8400. I've tried Coretemp, speedfan, and BIOS readings and one of my core temp diodes shows 46 C idle and ~50-52 C load but the other core temp stays locked at 46 C no matter what I do. I hear that temp sensors are a big issue with this CPU and I was wondering whether I should RMA this CPU and hope for a working one? I updated to the latest BIOS (F11) but it had no effect on the temp readings.

So far I haven't had much time to overclock. I tried out 9x450 at a cpuz vcore of 1.35 and mem at 2.2, but it orthos failed after 4 hours. I am currently trying to find the FSB wall by running 460x6 at vcore 1.38 (I'll find out when I get home if that worked).

 

[n7]

Yes, now that you've voided your warranty you should RMA...

 

[Penley]

Well at the very least, is this a problem or should I just assume that both cores are roughly the same temp as core 1?

 

[boomhower]

Originally posted by: n7
Yes, now that you've voided your warranty you should RMA...

I disagree. If he didn't damage the processor by overclocking it or using and aftermarket heatsink then he did not void the warranty. Same concept as putting a different than OEM air filter on your car. Technically your not supposed to but they have to show that in and of itself caused whatever the problem is before they can just void. I want to say it is the magnuson act or something of the like.

 

[Penley]

The processor seems to be running fine it is faulty temp sensor that concerns me. It had that same problem at stock conditions when I first put it in. Also the temp seems to be 46 idle/52 load under stock or OC conditions, which strikes me as strange.

 

[imported_Bee]

the processor is working without any problems at stock frequency, isn't it?

There is no reason to RMA it.

 

[Tempered81]

its the software, not the temp sensors... you may or may not have the same problem with a new chip if you rma. 90% of e8xxx owners have faulty temp readings in Coretemp/speedfan

 

[boomhower]

I don't think anyone knows just yet if it is hardware or software. I do know on the major vender's sent back a large quantity of CPU's due to the faulty temperature readings. If it is due to the chip I say you have every right to RMA it. Just because it performs at stock speed does not mean it is not defective. Correct temperature readings are part of a fully functional chip.

 

[DSF]

Originally posted by: boomhower

Originally posted by: n7
Yes, now that you've voided your warranty you should RMA...

I disagree. If he didn't damage the processor by overclocking it or using and aftermarket heatsink then he did not void the warranty. Same concept as putting a different than OEM air filter on your car. Technically your not supposed to but they have to show that in and of itself caused whatever the problem is before they can just void. I want to say it is the magnuson act or something of the like.

For what it's worth, Intel's warranty does not apply to processors which have been operated outside the published specs. It doesn't, at least following the letter of the warranty, make any difference whether the chip was damaged that way or not.

 

[theprodigalrebel]

I thought there was a BETA version of Speedfan that worked well with the Wolfdales? Anyone have any idea about that?

 

[taltamir]

Keep the CPU, a few months after a new stepping is released, RMA it... While the chip is defective, its not a defect that prevents you from using it.
It is a defect over which it is a good idea to RMA, and if you wait until a revision you get a free upgrade...
(and intel gets to keep the money it cost to make the replacement CPU longer, and possibly improve production efficiency and cut costs so that it costs them less to replace)...

Everybody ends up happy

 

[imported_Bee]

Originally posted by: jaredpace
its the software, not the temp sensors... you may or may not have the same problem with a new chip if you rma. 90% of e8xxx owners have faulty temp readings in Coretemp/speedfan

umm what?!? This doesn't make any sense to me.... why would the software work on some CPUs (10%) but not on others (90%)

Originally posted by: taltamir
Keep the CPU, a few months after a new stepping is released, RMA it... While the chip is defective, its not a defect that prevents you from using it.

wrong, the chip is NOT defective. Maybe, and only MAYBE, the temperature sensor(s) is (are) defective. No RMA folks, I tell you.

 

[Zenoth]

Right now 99% of the motherboards out there do not fully support the 45nm models, the temperature readings are off by a certain extent for pretty much everyone. In my case the best temperature "reader" I've seen is Lavalys EVEREST 4.20. There's also SiSoftware Sandra Lite, but that one shows temps a little higher than EVEREST's. And I see that you (topic creator) are using a Gigabyte motherboard, and right now Gigabyte are actively updating the BIOS of all their P35 and X38 motherboards since the past two months to improve compatibility issues with these new Penryns.

My suggestion before you RMA is to wait for newer BIOS revisions to come. Because keep in mind that if the motherboard itself and the BIOS can't support those chips fully then how could software be more precise than the BIOS itself, I don't think that'd be possible. In my case I am lucky that Gigabyte released the F7 BIOS only a few weeks ago in prevision for the 45nm models, but even BIOS F7 was released before the actual retail release of the E8x00 models. And right now there's an F8b revision floating around which might well again improve compatibilities.

 

[Anubis]

yea temps are all over the place
case in point

http://pics.bbzzdd.com/users/TheEvil1/temps.jpg

abit tool and speed fan are about the same, sandra is close to those and everest is about 12C higher :/

gonna push the chip when my Ultima 90 gets here, just chilling it with the stock cooler and paste ATM

 

[Killrose]

Make sure you are using an app that stresses both cores. Some apps only use one core and the other will stay cooler. Run two instances of prime at the same time and see what happens.

 

[imported_Bee]

a BIOS problem? I doubt it. If the CPU reports bogus temps, the BIOS will report bogus temps and so will CoreTemp, Speedfan etc. etc. Why would ALL the mainboard vendors out there have the same problem? Has any vendor managed to fix this "little" issue yet?

 

[Penley]

Originally posted by: Bee
a BIOS problem? I doubt it. If the CPU reports bogus temps, the BIOS will report bogus temps and so will CoreTemp, Speedfan etc. etc. Why would ALL the mainboard vendors out there have the same problem? Has any vendor managed to fix this "little" issue yet?

Which brings me back to the original conclusion that the chip is defective. Bad temp sensors=defective. And it is a risk to the system because thermal shutoff protection may be compromised if the temps are wrong.

Though it seems to me there is no point in sending back right now as I would probably get another of the same batch with the same problem. Hopefully waiting 15 days is enough time to get a better batch of chips in.