Intel QX9650 Drastic Core Temperature differences!

[xgman]

IHS looked fairly flat....it's just the cores running in pairs have a large displacement in temps ...the replacement 9650, under idle conditions with the Freezone Elite, cortemp is showing core #0 & 1 at 43C and core #2 & 3 at 36C

So what is the Apples to Apples compare on the elite at similar cpu speed and vcore to the Xtreme air cooler you were on before in temps or overclockability?

[JohnZS]

No, 90% of them are like this and I don't think it really hurts anything. You would be wasting your time, unless you have a bad clocker anyway.

Whereas that is probably true, I had stability problems on stock clocks when I used anything which stressed the cores 3 and 4.
So you might want to see if you are stable and if so, as xgman said it might not be worth the gamble as so many people are having these problems
John

[Ace-a-Rue]

So what is the Apples to Apples compare on the elite at similar cpu speed and vcore to the Xtreme air cooler you were on before in temps or overclockability?

the elite is NO better than the Ultra Xtreme air cooler (TRUE) when loading all four cores at 100%...NOW, i can get cooler idle temps out of the elite....when i load all four cores up, it matches fairly even with the TRUE...that is based on lapping the TRUE heatsink base....the base on the elite looks good...i am not planning to lap it since it is more cumbersome since the liquid tubing is attached to the heatsink base...i guess that i could take off the permanent clamps but that is something i really do not want to do in case i need warranty on the elite later on.

[xgman]

Is it any quiter?

[Ace-a-Rue]

i'll need to put the side cover on it to determine that....i still have a lot of fans whirling inside the case.

[JohnZS]

Fans certainly do make a noice without a controller and stuff
What is a TRUE?
I'm still waiting on my RMA, this is the latest from the e-tailer

Hi,

Thank you for your email.

We can confirm your return has been received back and we are currently awaiting stock to replace.

Please advise if we can be of any further assistance.

Regards

The fustrating thing is that they have no official ETA from Intel when they will be getting their stock alocation, it could be anything from days to weeks. (to months even). The Good News is that I will automatically get a refund if I have to wait more than 14 days from when they said my chip was faulty (earlier this week).

I wouldn't be surprised if they get stock next week as rumour has it the 21st is the UK launch of the 45nm Dual Cores.

John

[Ace-a-Rue]

Fans certainly do make a noice without a controller and stuff
What is a TRUE?
I'm still waiting on my RMA, this is the latest from the e-tailer

The fustrating thing is that they have no official ETA from Intel when they will be getting their stock alocation, it could be anything from days to weeks. (to months even). The Good News is that I will automatically get a refund if I have to wait more than 14 days from when they said my chip was faulty (earlier this week).

I wouldn't be surprised if they get stock next week as rumour has it the 21st is the UK launch of the 45nm Dual Cores.

John

sorry, john, a TRUE is a "Thermalright Ultra Extreme" heatsink

as far as noise from three 120mm fans are minimal, at least for me, since i am a former US Navy and Airline pilot (retired) who listened to fan noise in the co*ckpit non-stop!...listening to it right now sort of reminds me of being there right now.

the Freezone Elite is a nice unit but not competitive with the TRUE, especially in price!

[Falkentyne]

So that's why an airplane is in your avatar....

How did you like flying anyway? Why did you retire?

[Ace-a-Rue]

So that's why an airplane is in your avatar....

How did you like flying anyway? Why did you retire?

LOVED IT!

navy was due to reaching a desk job after you promote up the chain...it is hard to remain in a flying seat once you become a senior officer.

in commercial airlines, you are forced to retire at age 60 during my era...the FAA just recently up the retirement age to 65...can't go back and i wouldn't either.

if you click on my name to the left, i have my airline and navy career highlighted on my home page.

[Falkentyne]

That page is really awesome, Ace. A very storied history in pictures. +1

[Ace-a-Rue]

thank you

[JohnZS]

Yeah there are some ace planes

Hi as Ace already knows I got my replacement CPU back last week. It has a package date of 01/18/08 and is a L740B120T batch
My idle temps are a bit....interesting, however this time the CPU can sustain a large amount of load (100%) and can survive Intel's Cluster Stress Testing which stresses a CPU more than TAT and more than Prime and even more than Orthos!
There is a thread here on how to get this utility from Intel and how to set it up.

My new idle and load temps



As you can see it gets quite hot, but at least it does not crash out! For those interested it can run up to 3.6Ghz on stock voltage, but requires a slight bump upto 1.38v (1.36CPU-Z) to get 4.0Ghz
Sadly my system cannot run 1600Mhz at any multiplier (even @ 6) due to ASUS not supporting my 1600Mhz RAM. I am stuck @ DDR3 1333 or DDR 1200 if I try to set 1600FSB.

I have made ASUS aware of this issue and am hoping a future BIOS update will rectify the issue. FYI I cannot use their latest 0808 BIOS as that breaks support for my RAM @ 1333Mhz (leaving me stuck @ 1066!)

Anyway I am happy....ish with my replacement, it would have been nice to see slightly lower load temps as I believe 65C is the maximum these CPU's are supposed to get. However it is hard to say what temperature it is really at as I do not have a lot of faith in the DTS/DTS reading software on the new 45nm cores.

John

[Hornet331]

be advised, intel specifices a Tcase of ~65°C, not a core temp of 65°C, as recommended max.

Tcase is measured in the middel of the IHS not in the core itself.

[JohnZS]

Ahh thanks Hornet
But that is hard to measure? as most BIOS' have crap temperature readings (well I think my one does anyway if you see the readings it gives above).
John

[Hornet331]

i dont know if its possible to measure the exact tcase temp without modifing the IHS.

http://download.intel.com/design/pro...x/31873401.pdf

As you can see on page 91 in this document (its for the E8000 series but it doesn't matter since Tcase measurment point is the same since the P4) you can see how the measurment is performend and that you need a carved IHS.

[KTE]

JohnZS; what were your ambient/case temps there?

The difference between core1 and core2 is very abnormal idling.

[Ace-a-Rue]

it seems that is the norm for these 9650's...

[JohnZS]

JohnZS; what were your ambient/case temps there?

The difference between core1 and core2 is very abnormal idling.

The ambient temps are around 20C/70F. I also noticed that one of the cores in particular reads very off the mark @ idle, however it is not quite so bad @ load (at least I can get 100% load this time!)

Sadly I think Ace is right, it seems that abnormal temperature readings are to be expected with this generation of chip.

I might be on to something though which may or may not help.....check out this thread Heat Problems on 45nm CPU's

John

[KTE]

Thanks. In the UK, those temps are basically wrong and impossible even with the low ambients we're having. This again shows bugged DTS. For a >$1000 product, you expect much better.

Have you tested at what temp your chip shuts down and throttles? It may help...

[Ace-a-Rue]

Thanks. In the UK, those temps are basically wrong and impossible even with the low ambients we're having. This again shows bugged DTS. For a >$1000 product, you expect much better.

Have you tested at what temp your chip shuts down and throttles? It may help...

my chip shuts down at 4285 mhz over 80C in coretemp

EDIT:...that's running wPrime 1024M stability bench.