Dell Latitude E6400 - Overheating issues with minute tasks such as playing flash videos for extended periods?

This is a problem that Dell is well aware of and has done nothing about, except censor forum users who make to big a fuss about it. The problem appears to be due to an oversensitive BIOS prone to dialing down clockspeeds at the first sign of heat. There used to be a fDell forum user by the name of Tinkerdude that reported ind etail the issue hear, but Dell has since banned him, and he create a 59 page PDF document detailing the problem. There is something wrong when a laptop taht is not even a year old has 17 BIOS revisions already. You can find the document and more info about the issue here - http://www.maximumpc.com/article/news/overheating_problems_plague_e6400_and_e6500based_dell_laptops

After reading Randall Cotton's (aka Tinkerdude's) 59 page pdf clearly outlining this issue, I was motivated to create a tool to try to give a user running Windows a little more control of Clock Modulation and the ability to keep the multiplier at its maximum.

The program is called MaxCore2 and more information about it is available here:

http://www.xtremesystems.org/forums/showthread.php?p=4147775#post4147775

I'm very interested to see if this simple utility gives you any control over this problem.

http://www.fileden.com/files/2008/3/3/1794507/MaxCore2.zip

Edit: I added the ability to reduce the core voltage VID at full load.  Most Core 2 CPUs can run just fine with a little less voltage and should create less heat with this option.  Do some Prime95 testing to make sure that your computer is still 100% reliable with this option.

What else do I have to add to this tool to get users to check it out? 

I think you'll be very surprised at the results.  :)

I'm curious if you feel this has any correlation to the issue (albeit intermittent) of docking an E6500 on the e-dock, and the machine freezing entirely?

I seriously doubt there's any connection, sorry. In my testing, even when throttling was taking place, docking and undocking the system didn't cause it to freeze.

Hi tinkerdude and attwifi,

I agree that these two issues are likely unrelated.  I happen to suffer from both.  It seems that the 'complete and total system freeze' issue disappeared with BIOS A14 or so, but is back with A18 (it's happened twice in a few days).  Basically, the freeze occurs intermittently the moment when an external power source is applied -- either an AC adapter or the docking station.  The pointer and keyboard completely stop responding and all OS activity halts (though the screen retains whatever frame was there when it froze).  I now need to remember to save my work and shut down sensitive applications before plugging it in... is this all really necessary? 

DELL - please fix these fundamental issues.  All I want with this machine, which is the most important tool for my job, is that it "just work."

Dear, Tinkerdude How do you think, buying Dell E6400 I am going to experience same throttle problem, or Dell probably have changes something in their latest releases of products (may be on hardware level). Or I will experience same problem... batter to stay off?! Thank you!!!

Well, I can't say whether you'll experience the same problem. Many people have bought E6400s and have had no problem, I'm sure. And I have only done testing with the E6500, so I don't know how much of what I've found applies to the E6400 (though clearly they are similarly affected).

What I can say is that for the E6500, the problem only happens when the internal temperature of the system gets hot enough (the temperature in general, not any specific component). And it doesn't take as much heat to cause throttling when the system is docked (that is, it is much more likely to throttle and the throttling will be much worse when the system is docked). So if you buy an E6400 and never use it with a dock AND never use it in an environment above, say, low 70's Fahrenheit AND make sure that the system doesn't accumulate any dust in the fan exhaust, then you may not ever have this problem.

But how many people can forecast how they'll need to use their system the whole 3 or 4 or more years they have it? And why would you want to put yourself in a position where you know your system may sabotage itself SECRETLY, when you are using it intensively and may be on a deadline? That's what happened to me multiple times before I figured this thing out - I was working on a video project that was on a deadline and the harder and faster I worked, the slower and slower my system became. Imagine my surprise and fury when I discovered that for months, Dell had been surreptitiously and severely cutting into my productivity for reasons that they still will not explain to me. Imagine my outrage that I had spent over $2000 for Dell's top-of-the-line corporate-class dockable desktop replacement with a 2.2GHz processor that was inexplicably choking itself down to the equivalent of less than 100MHz (yes, you read that right - 0.1 Ghz).

And you might want to ask yourself - even if you know about the potential problem, would you really want to put yourself in a position where you (or whoever else uses your system) may not even know, months or years down the line, that it's slowing down and cutting into productivity due to this problem? Dell says they've sold millions of these systems - I wonder how many people's systems are slower than they should be and are cutting into user productivity without them even knowing it? Even one out of a hundred people would mean at least tens of thousands.

Why would you buy a laptop from a company with that combination of incompetence and disdain for their users? I know I wouldn't if I had the chance to do it over. I used to be a big Dell fan. I bought what I did based on their reputation from a few years back. Dell has since clearly gone downhill and has lost their way. Their global market share has been falling for some time (they just lost the #2 spot to Acer after giving up #1 to HP a few years back).

And as for whether a brand new system will be any better, let me just say that when Dell finally decided they wanted to exchange my system so they could do testing on it, they built me a brand-new system, as if I had just ordered it (this was back in August). But it had exactly the same problem.

Furthermore, it seems clear that the problem can be entirely addressed with an appropriate BIOS fix (though they've yet to come out with that and it stands to reason they may never do so, for some mysterious reason), so I doubt they'll make a hardware change in the E6400, especially after the system has been out for more than a year now. It's tough to see how making a hardware change would fix what is essentially a software problem. And as specialized and compact as laptop system designs are, it's also hard to see how they could make a hardware change to completely neutralize this problem without a radical redesign.

Has anyone tried this fix? http://www.xtremesystems.org/forums/showthread.php?p=4147775#post4147775

I have an E6500 running with a T9400. I've been having this problem pretty much since Winter of 2008. I had done so much testing to figure this stupid thing out, but it's nice to finally see a definite answer. Tinkerdude, you are awesome. Your dedication to this problem is inspiring. Dell should hire you so you can teach their engineers a lesson on thermal management.

My symptom is pretty straight forward. At some point in the day (usually multiple times) my processor spikes at 100%. This happens even with no general applications running. I have to completely shut off my computer, a reboot will not do. I even pull the battery out. Then it goes back to normal for a time. I've noticed in my computer's System event log a Kernal-Processor-Power error. Here is a screenshot of it. I get that error every time my computer spikes it's processor. I've been running Real Temp for some time and I notice that if the temp stays around 60C for a minute or two that it causes the problem. 

My office temp is usually around 74 to 80 F. I only notice it when it is docked. I ran with my laptop unlocked for a day or two and it didn't happen once. My laptop is 98% of the time docked.

Thinking that it was heat related, I created a sticky-note heat duct. Here is a picture of my heat duct. Ever since I started using it, my problems have been reduced to maybe only once or twice a week instead of once or twice a day. I have cleaned out the duct as well since adding the heat sink but it wasn't really dusty.

I do not want to install BIOS A18. It says they've removed the HDD Accoustic settings which I am guessing means the HDD will be in "quiet" mode instead of performance to reduce heat. I wish someone could confirm if this is the case.

Although some people have said this is not heat related, that is not entirely true. It is heat related to the extent that the buggy software reacts to the heat improperly. Therefore, keeping the laptop cooler definitely helps.

Tinkerdude, thanks again for all your work on this, I know a lot of people stand to benefit from it if Dell changes their ways.

Has anyone tried this fix? http://www.xtremesystems.org/forums/showthread.php?p=4147775#post4147775

The 59 page tinkerdude bible made this problem very easy to understand.  I could have helped you guys out a year ago if I knew what was going on but I just found out about this issue the same day I wrote that tool.

Anything a Dell bios can do to slow your computer down can easily be reversed with software.  I am well aware of and have worked with the registers in the CPU that tinkerdude documented.  His testing and the Intel documentation shows what value these registers should be set to so that the CPU can run at full speed.  Monitoring these registers and making sure that they always contain the "full speed" values and not the "slow throttled" values is simple to do with software.  I already had a similar utility available for the DFI X58 motherboards that had a similar problem when Windows 7 was in beta testing so changing it slightly to handle this problem was a trivial task.

I've talked with tinkerdude via e-mail and he's suggested that I should make my tool as simple as possible so anyone can use it.  I agree.

The tool I posted should work just fine and won't consume a lot of your system resources.  Start it up and let it run for a couple of hours and then check the Task Manager and you will see how little CPU time it uses while monitoring these registers.  Even if it takes 0.01% of your CPU time, that's a small price to pay to get back 90% of the performance that's gone missing on some throttled down Dell laptops.

Edit: Head to Google and type in, Asus turbo throttling, and you will quickly see that I'm no stranger to CPU throttling issues.  Getting these Dells back up to full speed was like another day at the office. :emotion-1:

With some user feedback, hopefully I can make this tool even better.

Hi all,

I have been suffering with this problem for a long time now on an E6400 (with Nvidia graphics).  It went away for a long time, and came back last week worse than ever.  I think it was about the time our office switched from AC to heat (I am in California so it just happened recently.)

Anyway, I think I know what is causing it, and maybe why some are seeing it and others are not.  I took the bottom off the laptop and blew the crud out of it with compressed air.  This made the problem go away last time a long time (~3-6 months).  In the past 2 weeks I had been having this problem every single day (even with firmware A19).  Again, I took the back off, blew out the fan and all the heat syncs, and the problem has not come back since even though the ambient temp at my desk has been going up a degree or two every day.

It is also worth noting that I only have this problem at the office in a doc hooked up to two monitors (with the lid down 100% of the time).  I have NEVER had the problem at home (no doc) even though I use it at home very often. Also,  since I saw in another thread it might be the Power Supply, I got a brand new PS out of new stock and hooked it up to the doc.  This made no difference at all.  And yes, I am using the big dock power supply.

This may not solve it for everyone, but it seems to have solved it for me (at least for a little while).

W

http://www.fileden.com/files/2008/3/3/1794507/RealTempBeta.zip



I added a feature to RealTemp so you can monitor for Clock Modulation.  This is a "feature" in Intel CPUs that quite a few Dell laptops seem to be using and the tinkerdude paper clearly showed this.  When the bios starts using Clock Modulation to control the thermal performance of your laptop, CPU performance can drop like a rock without your knowledge.  At least now you have a tool available so you can monitor this.

The big problem is the bios that doesn't restore full power soon enough.  It takes way too long for performance to return to full power after these laptops get even slightly warm.  Intel says these CPUs can run reliably at 100C but the bios is cutting power to them a long time before that to control heat output.

As soon as Clock Modulation is detected, the Thermal Status area in RealTemp will immediately change and report how severe it is.  87.5% is mild modulation and 12.5% is severe modulation and a severe reduction in performance.

If you see signs of Clock Modulation then you can do a quick (or not so quick) XS Bench benchmark and you'll be able to see how much performance you have lost.  This benchmark scales very linearly with CPU performance so when your benchmark score drops in half, that's a sign that your overall CPU performance has dropped in half.

At least now when users feel that their laptop is running like a slug, they'll be able to document it and will be able to see if Clock Modulation is part of the problem or not.

The tool I posted above has breathed some new life in the Core i7 mobile laptops that have similar issues.

Unclewebb,

Initial test results seem positive! 

For whatever reason, I get this problem when hooked up to a dock using an external monitor with resolution higher than the native panel (1440x900) at home only.  At work I dock with a 19" (1440x900) and don't see the problem.  Very weird.

Usually within a few hours at home my CPU will peg at 100% (with FF and Outlook open) and the system becomes unusable.  I let this happen as usual, waited about 15 minutes for throttling to subside, and fired up the app.  Haven't seen a problem while docked since.

Question: is going to eat battery life if left on?

Thanks for the efforts!  If this is a real fix for the problem I will be grateful (even if it should really be a bios fix).

John

unclewebb: thanks so much for creating this program!! it has truly been helpful in tracking this problem.

I just downloaded the beta but it crashes on me. The GT version runs ok though. The 3.0 non-beta works fine (which I've been using for some time now).

I'm running Vista SP2 32bit.

Problem signature: 

  Problem Event Name: BEX 

  Application Name: RealTemp.exe 

  Application Version: 3.5.0.0 

  Application Timestamp: 4b25596b 

  Fault Module Name: StackHash_8c1b 

  Fault Module Version: 0.0.0.0 

  Fault Module Timestamp: 00000000 

  Exception Offset: 0012f114 

  Exception Code: c0000005 

  Exception Data: badc0de1 

  OS Version: 6.0.6002.2.2.0.256.1 

  Locale ID: 1033 

  Additional Information 1: 8c1b 

  Additional Information 2: 4fdc91ca2e8b0d38236992b4c6740617 

  Additional Information 3: 45f7 

  Additional Information 4: 36594a51604567d89c15ca6c871eef05

I downloaded MaxCore2 and it doesn't appear to be doing anything. RealTemp still shows the MHz fluxuating from 1.8GHZ to 2.2GHz, when it should be at 2.53GHz (I'm running a T9400).

MaxCore2 is designed to eliminate Clock Modulation and to keep your multiplier high at full load.  It probably won't do anything at idle.

If you don't want your multiplier fluctuating at idle then you need to go into your Control Panel -> Power options and you need to set the Minimum processor state to 100%.  Also disable C1E.  That's what ultimately controls your multiplier at idle.

Run i7 Turbo which is included in the RealTemp download and then run a testing program like SuperPI mod to put some load on your CPU.  Post a screen shot of i7 Turbo when it is about half way through a calculation so I can see how your CPU is doing.

http://www.fileden.com/files/2008/3/3/1794507/RealTempBeta.zip

I've recently added ATI GPU temperature reporting to RealTemp but there is still a bug or two.  I'm not sure if that caused your problem.  I got a little side tracked from RealTemp when I found out about all these throttling issues that effect the Core 2 and the newer Core i7 laptops.  If that is what is causing the crashing problem in RealTemp then you could try adding this to the RealTemp.INI file:

NoATIGPU=1

That might help.  Too many projects, never enough time to finish them all.