C2Q 9400: 1 core idles a good 8 - 10C hotter than the other 3?

[NTB]

Just upgraded to a new MB + Intel Core2 Quad 9400...seems to be working fine, but after trying both the intel stock cooler and an AC Freezer 7 Pro, there's one core that consistently stays a good 8 to 10 degrees C hotter than the other 3, even at idle. I'm guessing that could be a problem?

Right now I've got 3 cores hovering around 30C; while the last one is sitting at 38C.

Nathan

 

[Rubycon]

Load it up with Linpack and see if the delta persists. Idle deltas can be higher because that one core may be doing something that's too little to notice with the task manager, etc.

 

[NTB]

Originally posted by: Rubycon
Load it up with Linpack and see if the delta persists. Idle deltas can be higher because that one core may be doing something that's too little to notice with the task manager, etc.

will do. Right now I'm working on a bunch of photos that I shot on a recent trip. While the temps for the other 3 cores bounce around as the pics are converted from RAW to TIFF or JPEG, the 'hot' core isn't budging - but they're all being used, to one extent or another.

Nathan

 

[soccerballtux]

If it goes away during loading you've likely got a faulty (but still working) temperature diode. Intel only guarantees them to work accurately at temperatures close to the thermal limit of the chip. After all, that's all that really matters.

If it's stuck then mebbe RMA.

 

[Flipped Gazelle]

Hmm... I have the inverse of this issue. On a Q9550, 1 core idles about 5C lower than the others, and 8C under OCCT.

 

[CJM]

i have the same thing with my 9400. Core 4 idles at a higher temperature, and then under load core 1 is always the highest. I don't really think it's anything to worry about.

 

[BonzaiDuck]

Originally posted by: soccerballtux
If it goes away during loading you've likely got a faulty (but still working) temperature diode. Intel only guarantees them to work accurately at temperatures close to the thermal limit of the chip. After all, that's all that really matters.

If it's stuck then mebbe RMA.

Wolfies and Yorkies -- Yup. Posi-lutely . .. Abso-tively. I'm skeptical about the RMA possibilities given that Intel disclaimer -- so I live with it. But he can try . . .

I still don't understand, though, why many of us are stuck on this core-sensor thing -- while our core-sensor-things are stuck-ha-ha. They still use the legacy TCase sensor; its readings feed BIOS monitor; some of the Windows monitoring software reports TCase; and it's known that TCase is supposed to lag behind the core-average by 5 to 10C. Even if they're neck-and-neck, TCase is THE basis for the Intel thermal spec.

 

[F1shF4t]

Originally posted by: BonzaiDuck

Originally posted by: soccerballtux
If it goes away during loading you've likely got a faulty (but still working) temperature diode. Intel only guarantees them to work accurately at temperatures close to the thermal limit of the chip. After all, that's all that really matters.

If it's stuck then mebbe RMA.

Wolfies and Yorkies -- Yup. Posi-lutely . .. Abso-tively. I'm skeptical about the RMA possibilities given that Intel disclaimer -- so I live with it. But he can try . . .

I still don't understand, though, why many of us are stuck on this core-sensor thing -- while our core-sensor-things are stuck-ha-ha. They still use the legacy TCase sensor; its readings feed BIOS monitor; some of the Windows monitoring software reports TCase; and it's known that TCase is supposed to lag behind the core-average by 5 to 10C. Even if they're neck-and-neck, TCase is THE basis for the Intel thermal spec.

Thats probably because TCase temps are innacurate and vary by proc and mainboard. I have two q6600 one G0 the other B3. For the B3 Tcase is reported as 25V under core temps, for G0 its usually almost same as core temps. Thats on one board on others its different again.

I think TCase is given as thermal spec cause its easier to use that than guess the thermal resistance of the heatspreader etc, when designing a heatsink.

 

[NTB]

Originally posted by: CJM
i have the same thing with my 9400. Core 4 idles at a higher temperature, and then under load core 1 is always the highest. I don't really think it's anything to worry about.

Running Linpack right now and my CPU is doing exactly the same thing: cores 2, 3, & 4 are all right around 45C; core 1 is at 48C.

Nathan

 

[alevasseur14]

Yeah, from what I've read, it's normal. Right now, Core2 in my Q9550 is idling a solid 4-5C higher than the other three cores. It took a while but I got over it.

I still think I'm going to yank the Freezer 7 Pro and put my Tuniq Tower back in. After finals...

 

[NTB]

Originally posted by: alevasseur14
Yeah, from what I've read, it's normal. Right now, Core2 in my Q9550 is idling a solid 4-5C higher than the other three cores. It took a while but I got over it.

I still think I'm going to yank the Freezer 7 Pro and put my Tuniq Tower back in. After finals...

Hah..I have an original (and considerably more expensive) ThermalRight version of that cooler (120mm fan, right?), but I'm not using it because ( A ) I couldn't get the brace off the back of the other board and didn't order another one originally, and ( B ) I didn't feel like ripping the new board *back* out to put a replacement brace in. My luck with PCs lately has been...less than stellar, so I didn't want to push my luck. Maybe later. I got the little Freezer 7 instead and replace the Intel stock cooler with that.

Nathan

 

[alevasseur14]

The only reason I don't feel like staying with the Freezer 7 is those stupid mounting pins. I got it off to swap processors and it took me forever to get them back together. My temps are pretty high (around 38-40C per core, except for 2...) so I'm pretty sure I messed up the paste putting the cooler on.

I'd rather stay with the Freezer 7 because it's quieter and cools a bit better, but I just don't feel good about re-mounting it.

 

[BonzaiDuck]

Well, we're back to the "push-pin" issue, I see.

I hate 'em.

All it took was some thermal issues with an OCZ Vendetta 2, and I took it off the system. "My thumbs! My thumbs! Ow!" And I'm not "all thumbs."

Here's something interesting, though, which I may have mentioned before.

The backplate for a Noctua NH-U12P works interchangeably with a backplate for a ThermalRight Ultima-90 -- which (I think) works interchangeably with one or more other TR coolers.

I'm getting lazy -- somebody mentioned not wanting to pull their motherboard to change heatsinks. So I had decided to try changing the heatsinks without pulling the board.