Huge temp difference between cores

[terrytpot]

I can't think of anything other than a faulty processor which would give me a difference of 30 deg C between the two cores of the new E8200 cpu I've bought...unless you include the fact that this is my first build (I know,noob error?)
My system specs are:
Antec 300 case
Intel Core 2 Duo E8200 2.6GHz (Retail 775)
Gigabyte GA-EP45-DS5
4GB Corsair Dominator 1066MHz DDR2
Sapphire ATI Radeon HD 4770
Zalman CNPS9500-AT
600W OCZ ModXStream Pro Modular PSU
1TB Western Digital Caviar Green 3.5" Hard Disk Drive
1 x LG GH22N40 DVDRW-RAM SATA

My first thought was that I hadn't properly secured the Zalman cooler to the cpu (I recalled that of the 2 screws securing it to its bracket,one was quite awkward to get to as the fans in the way and that screwhead was getting "chewed" up and I wasn't sure if I'd done it up tight enough..so, I removed the cooler and the as5 I'd used,cleaned it all up and reinstalled everything making sure this time that both screws were screwed in all the way)..no change to the temps though and I've used real temp,core temp and speedfan all of which report a similar difference between the 2 cores temps. I'm not absolutely panicking though as I cant believe that 2 cores which are so close together can have such a huge difference in temp (core 0 runs at 38ish while core1 runs at 68ish whilst idling...and no,its not ockd,thats at stock speeds!) Any suggestions as to the cause of this would be greatly appreciated.

 

[jalpaugh1978]

I, personally, dont care for the Zalman heatsinks. IMO, it's much easier to get good steady temps with and tower cooler w/a backplate like a Xigmatek 1283. Ran into the same problem on a friends pc with a Zalman 9500. Finally we just put my old Arctic freezer 7 Pro on it and all was good

 

[kbits]

Stress the system and watch both cores temperature.

If the core0 goes up, and core1 stay at 68, then the core1 temperature sensor is stuck, which is very common on those cpu.

In that case, the cpu itself is working properly.

 

[randomizer]

It's a high temperature to have a sensor stick at but it's not unheard of. As long as you don't really exceed 68C it will never budge.

 

[terrytpot]

hmmm,not too sure its stuck as I've just tried running the small test in prime95 and I've managed to nudge it upwards
results: at the 14k point
speedfan...52/76deg C @ 2670.00mhz (multiplier not shown)
Realtemp...48/71deg C @ 2666.41mhz (claimed speeds of 333.3 x 8)
Core temp.47/70deg C @ 2660.00mhz (claimed speeds of 332.5 x 8)

I am a bit concerned as the plan was to oc this rig to 4ghz and the temp on core 1 is allready maxed out at stock speeds..unless of course its telling "pork pies" intel give a max temp of 72.4 I think...

who reckons I should be thinking of chucking it back at Intel for a replacement?

 

[randomizer]

That's odd indeed. The IHS may not be flat, or it may not properly bond with the cores, but since they are the same piece of silicon it seems unlikely that they would be so far apart in temperature. If one of them trips the throttling mechanism well before the other one then something is not right, and right now that looks like what will happen.

What does Speedfan say your "CPU" temp is out of curiosity?

 

[terrytpot]

Not quite sure how to find an overall cpu temp with speedfan but here's what I can get from it:
Temp1 34C
Temp2 37C
Temp3 -2C
Temp 0C
Core 0 51C
Core 1 75C

gulp...

 

[terrytpot]

I've now removed the zalman cooler and fitted the oem cooler which has obviously bumped the temps up a bit..funny thing is all these scary temps are being shown whilst its happily plodding through the blend stress test which still hasn't listed any errors?