E6400 overheating throttling

Well, if a system is being temporarily throttled when idle (but the throttling goes away immediately when utilization picks up), that's just Intel's SpeedStep, which is harmless. Are we talking about something different here? Something where throttling kicks in and persists even at low utilization and low temperatures? I never saw that on my system (E6500/NVIDIA).

 

dcp12345678,

It really does make a difference for me. Like a 10-15 degree Celsius difference. That is, throttling occurs at temperatures 10-15 degrees Celsius cooler if I'm docked. It's been a common complaint of many other folks that the throttling problems are worse when docked.

But you seem to have escaped any trace of this problem, which is both confounding and encouraging to me. Perhaps something has been changed on more recently manufactured systems - but what, I wonder? Are you able to reach GPU temps around 90C like Goldeneye? That just blows my mind. There's no way I would ever get there with my E6500/NVIDIA system.

 

wsx,

Can you suggest a good tool to read chipset temp (under XP)?

Thanks

 

I use Everest to monitor temperatures.

 

Readings taken from Speedfan:

idle:
GPU: 61C
HD0: 39C
Temp1: 42C
Core 0: 36C
Core 1: 36C
Core: 57C

load (2 instances of StressPrime running, google earth flying through New York City, and playing a DVD all at the same time):
GPU: 92C
HD0: 41C
Temp1: 42C
Core 0: 61C
Core 1: 63C
Core: 86C

The system ran fine the whole time, and I got no throttling and no event 7's at all.

Hope that helps .

 

Yes, I reach those temperatures with no negative effects! See my last post.

Anyway, thanks for your extensive work in this area, like I said, I used your exact tests as soon as I got the laptop because my plan was to return it immediately if it had the problems and get the Lenovo T400. But I'm happy to say the laptop runs fine.

I don't know if Dell did something different with the newer machines, but if they are replacing heat sinks on existing folks' machines and that's fixing it, maybe whatever "fix" there is was incorporated into the newer machines by default (like maybe improved heat sink design or something?)

 

I can reach 102C my GPU using Everest to monitor my temperature, using Furmark to stress and have my GPU overclocked. Of course this is only stressing the GPU, so I don't know if my CPU/Northbridge is being affected by any throttling. But in games, such as GTA4, Far Cry2, The Witcher and all games in Orange Box, at their best playable settings, I face no throttling., UNLESS I block the fan with my hand which heats up a lot more the system.. but ten why would I do that?!

Yes you are correct, it is normal operation, and this is what Intel GPU user don't complain about.

Yes, the topic of the Intel GPU users, where where their system throttles is that they say when they do basic stuff, like using google earth, the system throttles based on their description. It would seam that when the Northbridge heats a bit, the system acts like it's overheating and slows down it's clock to prevent any damage. But, Nvidia users, like myself, don't have this problem, and infact the Northbridge can heat up A LOT more before it starts doing so.
This problem affect users when they don't dock their laptop (I don't know what happens when it's dock).

 

Well, improved cooling hardware couldn't be the explanation for the difference.

For example, you say your idle temps are:

Well, mine, right now, with an ambient room temp of 24C are (also via SpeedFan):

idle:
GPU: 56C
HD0: 41C
Temp1: 43C
Core 0: 33C
Core 1: 39C
Core: 50C

which are actually a little cooler overall

And, more importantly, your load readings are;

and readings for me just at the starting point of throttling are (testing right now as I write this):

load:
GPU: 67C
HD0: 42C
Temp1: 43C
Core 0: 54C
Core 1: 63C
Core: 62C

(All I did was make some scrolling text with a batch file and run an instance of Calculator working on a big number factorial)

So the explanation couldn't really be a better heat sink setup. If that was the case, then your load temps shouldn't even reach my throttle temps. There must be some other explanation.

 

@tinkerdude!.. yea and no.
The only way to calculate idle, is that your turn off the system for 20min until everything is cool, you turn it, log-in in Windows account, close all programs watch TV for 30min, come back and see the temperature. That is how you calculate idle. Looking at dcp1234567 idle temps, it looks like he just watch a movie/video or was doing something before.

As for load, tinkerdude you are wrong. What you did is not called load. The only thing you did is mildly pushed 1 of the 2 CPU. It might have work back in the old days when we had 300Mhz singel core CPU's, but not today. Load, means that:
CPU1, CPU2, HDD, RAM, and GPU are at 100%
Of course, this is very hard to achieve, as if you run something for the GPU, the CPU is used but not so much, therefor if you run Everest stress test + 3D FurMark either the CPU or GPU won't reach max 100% load, as tehy would battle for the CPU, but close to it.

 

Yeah, I didn't give it a full 20 minutes to cool down, only 3 or 4 minutes, so that could explain why the idle temps were a bit higher. I'm still pretty new to all this monitoring stuff, so I'm not that familiar with all the protocols to use for proper testing.

tinkerdude!
I don't really have an explanation then if it's not the heatsink design. I am running A14 of the BIOS, but from what I've read that doesn't seem to make a difference.