Real Temp - New temp program for Intel Core processors

[jason4207]

Wow! That is great info man. At least I know my HS is mounted well!

I have no idea about vdroop b/n the cores. That's the first time I ever thought about that, but I guess it's possible.

I have a thought! I've read about guys disabling some cores...I'm about to look into that!

BRB

---------------

There has to be a way to do it, but I can't find it. Anybody know how?

edit: I found out how to disable 2 cores using the boot.ini file, but since it's windows based I'm guessing the other 2 cores are still getting power. Plus, I'm not sure which 2 cores I'm looking at now...but temp wise they look like cores 0/1. No help.

[ChrisZ]

chrisZ: Thanks for that screen shot. That makes TjMax=95C look very believable for the 45nm Quad processors. The next RealTemp should make your 4 idle temps a lot more comparable. How about some Prime load temps now?

No problem.
However I was unsure how to define the "ambient" this time, the water in the CPU cooler was quite warm and it is actually the liquid providing the ambient, no? So I did the following: before taking the second screenshot I stopped the load and waited a few seconds to see the immediate temp drop.

No corrections in Real Temp. Background temps are from Everest Ultimate.

Hope this helps.

[RealTelstar]

Nice done!
Because I cant read all 25 pages, somebody can tell me if this works under vista x64?

Edit: Yes, it works. Just tried. I like executables.

[TankGuys]

Awesome job - I'm going to refer customers to this site. We've been getting a *lot* of questions on this. Well done!

[unclewebb]

At full load the 2 dies are different by 5*C, so they are tracking together pretty well. Maybe the 2 dies have different TjMax's?

When you first came up with the "different TjMax" theory within a Quad it sounded pretty crazy but after more testing it's about the only theory that makes any sense.

I use a Kill-a-Watt meter which shows you how much power your computer is using at the wall outlet. When running the single threaded version of Prime, whatever core I run it on, power consumption is the same. When running Orthos which is the dual core version of Prime, I get the same power consumption whether I run it on core0 and core1 or on core2 and core3. With Prime that is designed for Quad cores, I can choose any 3 cores to run it on and power consumption is always the same. If power consumption is equal no matter what side of your processor you're using then you would think that heat output of each individual core should be pretty much equal.

I decided to run 4 cores of Prime and to get the core temp a little higher I turned off the heatsink fan.

I still need to crunch some numbers but this 5C difference between core0, core1 and core2, core3 remained pretty consistent even as the first pair of cores went over 80C.



To me, a 5C difference is significant and there is no logical reason why two cores would be running 5C cooler when they are running the same core voltage and MHz and running the exact same load as the other two cores. Is TjMax different between the two pairs of cores? Maybe. I'll give users a way to individually bump TjMax for each pair of cores in the next release so other users can have a look at this issue.

It's just about time to pull the heatsink and see if that changes anything.

Edit: Here's my log file with TjMax=95C for all cores and no idle correction.

http://www.fileden.com/files/2008/3/...ncorrected.txt

Here's how the log file looks on my Quad using different TjMax for each pair of cores combined with a (+)(+)(+)(0) idle calibration. Much more realistic and likely a lot closer to the real truth.

http://www.fileden.com/files/2008/3/...empLogQuad.txt

and here's a second corrected run from idle to 80C and back to idle again.

http://www.fileden.com/files/2008/3/...mpLogQuad2.txt

[phill]

I wish it worked for me

[unclewebb]

I wish it worked for me

What doesn't work for you? I should have a fresh release of RealTemp ready within a day so the Quad guys can do some testing. If you are using Vista or a limited account you need to right click on the icon and choose Run as Administrator.

[jason4207]

When you first came up with the "different TjMax" theory within a Quad it sounded pretty crazy but after more testing it's about the only theory that makes any sense.

I use a Kill-a-Watt meter which shows you how much power your computer is using at the wall outlet. When running the single threaded version of Prime, whatever core I run it on, power consumption is the same. When running Orthos which is the dual core version of Prime, I get the same power consumption whether I run it on core0 and core1 or on core2 and core3. With Prime that is designed for Quad cores, I can choose any 3 cores to run it on and power consumption is always the same. If power consumption is equal no matter what side of your processor you're using then you would think that heat output of each individual core should be pretty much equal.

I decided to run 4 cores of Prime and to get the core temp a little higher I turned off the heatsink fan.

I still need to crunch some numbers but this 5C difference between core0, core1 and core2, core3 remained pretty consistent even as the first pair of cores went over 80C.



To me, a 5C difference is significant and there is no logical reason why two cores would be running 5C cooler when they are running the same core voltage and MHz and running the exact same load as the other two cores. Is TjMax different between the two pairs of cores? Maybe. I'll give users a way to individually bump TjMax for each pair of cores in the next release so other users can have a look at this issue.

It's just about time to pull the heatsink and see if that changes anything.

Edit: Here's my log file with TjMax=95C for all cores and no idle correction.

http://www.fileden.com/files/2008/3/...ncorrected.txt

Here's how the log file looks on my Quad using different TjMax for each pair of cores combined with a (+)(+)(+)(0) idle calibration. Much more realistic and likely a lot closer to the real truth.

http://www.fileden.com/files/2008/3/...empLogQuad.txt

and here's a second corrected run from idle to 80C and back to idle again.

http://www.fileden.com/files/2008/3/...mpLogQuad2.txt

When I typed that I spent a while re-reading it before I clicked the 'Submit Reply' button. Like you, I couldn't think of another explanation.

I think we need some more data points. If we can get a lot of folks w/ the same 5*C difference then maybe we can say something more concrete.

[unclewebb]

Core0 and core1 on my Quad track each other exactly from idle to full load just like any other dual core I've had. Even the cores in my E8400 track exactly until down low when they both get stuck at slightly different temperatures.

Core2 and core3 don't start tracking each other exactly until the high 70C range and their idle behavior is different as well. If you ever graph two cores the graphs are pretty much identical from idle to full load and back again but not these two cores and they don't match up with core0 and core1 at all.

We need more Quad users to step forward with data including the new 45nm Quads. After seeing this issue it's starting to make sense why Intel doesn't want users using DTS data to figure out core temperatures. Too many anomalies and variables to consider. Writing the software is easy. Trying to write documentation to explain what seems to be going on is going to be impossible.

I don't think this 5C issue effects all users but I know that something is wrong with my Quad as well as yours. Testing is going well.

[axis]

Is that at idle, full load or both?

Right now, i'm running 3.4ghz @ 1.32 Vcore at idle.

Everest: 50-43-42-49
RealTemp: 35-28-29-34

I'm not sure if the difference has to do with the processor being a 135w QX or not. The same temp variations appear on load as well. I do find it awfully strange that this 5C difference keeps showing up. Notice that my 3-4 core stays exactly 5C apart and core 1-2 also does that at load.

Axis

[Uberbob102000]

So what's the consensus on the B3 Q6600 TjMax? I think it may be 85 because in 28º ambient I get 52-50-47-48 idle temps @ 3.2GHz and 1.288 Vcore but drop those by 15 gives me some realistic temps. I have a Tuniq 120 w/ a San Ace in it.
Though on the new 45nm my temps are consistent and seem correct.

[NaeKuh]

The only thing that bugs me about this program is that i supposidely have 2 cores which are breaking the rules of thermodynamics.

There colder then ambient. :T

Other then that its a great program. Its fairly accurate, coretemp is only 10C higher in marking for everything including the TjMax.

[axis]

So what's the consensus on the B3 Q6600 TjMax? I think it may be 85 because in 28º ambient I get 52-50-47-48 idle temps @ 2.4GHz (speedstep) and 1.288 Vcore but drop those by 15 gives me some realistic temps. I have a Tuniq 120 w/ a San Ace in it.
Though on the new 45nm my temps are consistent and seem correct.

You're also running quite a bit of Vcore for a Q6600 at stock speed. Aren't most here running 1.20-1.25 at stock speed? That extra 0.038-0.088 will add quite a bit of heat to a CPU.

Axis

[Ace-a-Rue]

We need more Quad users to step forward with data including the new 45nm Quads. After seeing this issue it's starting to make sense why Intel doesn't want users using DTS data to figure out core temperatures. Too many anomalies and variables to consider. Writing the software is easy. Trying to write documentation to explain what seems to be going on is going to be impossible.

I don't think this 5C issue effects all users but I know that something is wrong with my Quad as well as yours. Testing is going well.

here is an idle shot, case door closed, Coolit Freezone cooler with ambient in room is about 75F....when i run wPrime 1024 bench the temp comes within 5C difference.

my QX9650 is currently at 4.1 Ghz.

i hope this helps.

[jason4207]

The only thing that bugs me about this program is that i supposidely have 2 cores which are breaking the rules of thermodynamics.

There colder then ambient. :T

Other then that its a great program. Its fairly accurate, coretemp is only 10C higher in marking for everything including the TjMax.

Did you try calibrating idle temps as per the 1st post?

[Uberbob102000]

Whoops.. I'm tired I put down Speedstep Freq @ 6x

[unclewebb]

Did you try calibrating idle temps as per the 1st post?

Reading the directions is too much work. Much easier to complain!

Ace-a-Rue: Can you post a screen running Prime on your 4 cores along with RealTemp. Just run it long enough so the temps are fairly stable on all 4 cores and hopefully so I can see if there is a 5C difference between the two pairs of cores. The new version will be ready tomorrow to deal with this situation.

Anyone else with a Quad should do me a favor and post the same. If people want accurate temp monitoring software then the more data I can gather the better.

[phill]

What doesn't work for you? I should have a fresh release of RealTemp ready within a day so the Quad guys can do some testing. If you are using Vista or a limited account you need to right click on the icon and choose Run as Administrator.

I'm using XP Pro SP2 and when I first had my QX I tried Core Temp and Real temp and the values where static, they never moved one degree. Didnt matter if I put it under a load with Prime or if I left it in the desktop with it idling I think something isnt right with the CPU rather than anything else... Although saying that, Speedfan and the bios are only two things I can monitor my temps with I just dont know how accurate they are either

[unclewebb]

I think something isn't right with the CPU rather than anything else.

It sounds like both of your on-chip sensors are stuck. The only other time I've seen that is with an Engineering Sample (ES) core processor. RealTemp or any other software that depends on reading these sensors is not going to work for you. I know SpeedFan lets you read from a motherboard diode and calls it CPU. That's all you'll be able to go by with your chip.

[phill]

Would I be entiled to an RMA with the product from the company I bought the CPU from in the first place??

[unclewebb]

Would I be entiled to an RMA with the product from the company I bought the CPU from in the first place??

Technically, no. Intel does not agree with or recognize any software that reads these sensors and tries to convert the data to an absolute temperature. Even if these sensors are stuck at your typical idle or full load temperature, they might become unstuck and start to work as they should as your cpu gets closer to TjMax. This is what they are designed and tested to do. Does the RealTemp TestSensors feature show zero movement on all cores?

If your processor ever over heats, catches on fire and melts into your motherboard because of these sensors, then you'll have something to complain to Intel about but you probably have a better chance of winning the lottery than that ever happening.

[jas420221]

Works with the Q9450.

I just need that individual core adjustment (and tray temps!!) now and Im all set!!!

EDIT: It says 55/52/52/55 all at full load 30mins.

[Ace-a-Rue]

Ace-a-Rue: Can you post a screen running Prime on your 4 cores along with RealTemp. Just run it long enough so the temps are fairly stable on all 4 cores and hopefully so I can see if there is a 5C difference between the two pairs of cores. The new version will be ready tomorrow to deal with this situation.

Anyone else with a Quad should do me a favor and post the same. If people want accurate temp monitoring software then the more data I can gather the better.

here you go; about 6-7C spread between the last core...pic taken four and half minutes into the wPrime:

[unclewebb]

Ace-a-Rue: Thanks for the info. If you have time could you try running this version of Prime and use Small FFTs:

ftp://mersenne.org/gimps/p95v256.zip

I find the load with this test is very steady and keeps the max temps steady as well.

[Nevrsadie]

Heres my Q6600 GO under load. Im getting the same difference in temps but with different cores. Cores 0 and 2 are the same, and cores 1 and 3 are the same, 6 degrees apart.