random BSoD minidump included

[PMildren]

I have been in computers for many years and have had very good luck in building systems that are for Win based systems, rock solid.

Recently I upgraded my system from an AMD 3000+ s939 w/ eVGA 7800GT OC and oem audigy sound card to the current system show in sig. Once I had everything installed it started getting random BSoDs. I reinstalled all drivers, over and over again. Used CCleaner, Det RIP and various other programs each time to try and remove driver issues. I tried full software installs of SB packages, then stripped them to just drivers. I have tried various NVidia drivers, I have reinstalled Win XP twice. After the Win XP home reinstall, things got a lil better. Fewer BSoDs, but still getting them. I have run AdAware, SpyBot S&D, am running ZoneAlarm firewall and AVG software and all showing nothing.

I am at wits end and figured I would finally get some professional help on reading the minidumps. Obiwan Kenobi, your my only hope. Help me Obiwan Kenobi.

http://www.filefactory.com/file/0cd2c9/

I know its kind of a pain in the ass...if someone can refer me to a free web based file hosting site, I am open

 

[mikeblas]

I can give it a twirl, but since I quit I don't have access to all the resources I used to...

 

[PMildren]

Thanks all help is appreciated

 

[der_jackal]

You've got a rag tag collection of BSODs that's for sure.

1. BugCheck 100000C5, {2e88cd84, 2, 1, 8054a16a}
2. BugCheck 10000050, {c19c0000, 8, c19c0000, 0}
3. BugCheck 100000D1, {1, 2, 8, 1}
4. BugCheck 1000000A, {0, 2, 1, 805217e0}
5. BugCheck 24, {1902fe, af64b878, af64b574, 8051a3d3}
6. BugCheck 7A, {3, c0000005, 0, 0}
7. BugCheck 100000C5, {6421de6f, 2, 1, 8054afd2}
8. BugCheck 4E, {8f, 4f712, 471b2, 0}


I don't have time to dig really deep to these right now, but so much variety is usually an indicator of something wonky going on w/ your hardware.

You can parouse the values yourself here.

Software related issues tend to be repeated, and narrow in BSOD values. You've only managed to hit the same bug check code once in 2+ days. That and the rest of those codes aren't terribly common for pure driver issues. The 0x24 and the 0x4E thrown in here could be indicators of either bad RAM or an overheating / faulty CPU. My favorite is the 100000D1, w/ the single digit parameters though.


Arg1: 00000001, memory referenced
Arg2: 00000002, IRQL
Arg3: 00000008, value 0 = read operation, 1 = write operation
Arg4: 00000001, address which referenced memory


DANGER Will Robinson, DANGER.

Compound all those w/ the fact you've performed a reinstall and it's still happening, I'd say run some Memory Test utilities, check your load / idle CPU temps to start.

If there is a driver / kernel mode binary causing this (by stack trashing, etc.), it's not likely we'll be able to pinpoint it from a minidump sadly.

Didn't know you quit Mike.





Maybe that's why the stock price has been going up lately? /joke

 

[iSkylla]

That's about the best responses to one of these BSOD threads that I've seen.

Apps to use:

memtest86+
Prime95
3DMark05, 06

 

[mikeblas]

I don't have time to dig really deep to these right now, but so much variety is usually an indicator of something wonky going on w/ your hardware.[...] Software related issues tend to be repeated, and narrow in BSOD values. You've only managed to hit the same bug check code once in 2+ days. That and the rest of those codes aren't terribly common for pure driver issues.

Indeed! I was about to post with way more details (and can if ya'll want) but this is a great summary. You trap in NV4_Something twice, but other than that, they're all different. Something's wrong; I'd guess that you're either overclocking and not nearly as stable as you think; or that you have been over clocking and you burned something marginally; or that you've got bad memory. The two traps with the same code were in different modules, IIRC.

Didn't know you quit Mike.
Maybe that's why the stock price has been going up lately? /joke

Yep, I'm done, and my time off has been great, neglecting random and unfounded pot shots at my career. Fact is, the stock tanked after I announced my decsion, then rose on news from the Fed and because the price of crude oil is going down.

 

[der_jackal]

Yep, I'm done, and my time off has been great, neglecting random and unfounded pot shots at my career. Fact is, the stock tanked after I announced my decsion, then rose on news from the Fed and because the price of crude oil is going down.

Oh let me live in my delusional world a little longer please. Besides, you're only replying here for the post count increase.

If it makes you feel any better, the stock price halved in my first year as an FTE.

Back to the Topic

One other thing I forgot to mention as a troubleshooting suggestion; It sounds to me like you built you system, one thing you should probably check is your Heatsink - CPU connection. Ensure the entire face of the CPU has thermal paste coverage and the Heatsink isn't making metal to metal connection anywhere.

 

[mikeblas]

Oh let me live in my delusional world a little longer please. Besides, you're only replying here for the post count increase.

Absolutely not. That said, I'm glad you've given me the opportunity to reply.

If it makes you feel any better, the stock price halved in my first year as an FTE.

Nothing about the stock price going down makes me feel good: during those years, I was "all in".

One other thing I forgot to mention as a troubleshooting suggestion; It sounds to me like you built you system, one thing you should probably check is your Heatsink - CPU connection. Ensure the entire face of the CPU has thermal paste coverage and the Heatsink isn't making metal to metal connection anywhere.

Do you think thermal throttling works? That is, modern processors are supposed to throttle themselves to deal with thermal problems... to the point of shutting down if they're about to burn up. Do you think a processor with heavy throttling is stable under the OS? If it's slow to respond to interrupts, having high latency in executing code, and so on -- won't this necessarily create lots of odd timing issues, and on?

That is, I wonder if thermal throttling actually causes lots of these wide-ranging bug checks.

 

[der_jackal]

Do you think thermal throttling works? That is, modern processors are supposed to throttle themselves to deal with thermal problems... to the point of shutting down if they're about to burn up. Do you think a processor with heavy throttling is stable under the OS? If it's slow to respond to interrupts, having high latency in executing code, and so on -- won't this necessarily create lots of odd timing issues, and on?

That is, I wonder if thermal throttling actually causes lots of these wide-ranging bug checks.

HHHmmmm....An increase bug checks, I don't think so, but only because I can't think of a lot of issues you'd see which would be exlclusive to a fully down throttled CPU versus a fully loaded CPU (stress testing) or a low memory system. And because of that I think the OS code (*nix, Windows, Apple) is robust enough not to commit seppuku there, but you would be more likely to see those non-fatal, nuisance issues (dropped frame in video, stuttering audio, ! code devices when trying to enumerate, etc.) as the lower level code failed out under that scenario.

That being said, the majority of desktop CPUs (AMD / Intel), I'm pretty sure (never have read the full white papers), all have a heuristic which throttles them back up to the highest required clock speed for operation as it sees CPU usage increase. So I'm not sure we'd ever be in a situation where a CPU would be under enough load and throttled down to its lowest clock speed for a long enough period of time as to present a problem.

Now, I'm not 100% sure here either (again,never read the full papers), but speed stepping lap tops under battery power should remain throttled down until the unit either has a full battery or is plugged into an external power supply so IMO that's where we would be more likely to see a rise in those nuisance failures or possibily an increase in fatal failures. And from my previous experience testing USB / 1394, I can't recall we ran in to any issuse which weren't also hit in those CPU stress / low memory conditions.

For thermal throttling, I don't think it really matters in the majority of overheating situations really. If a CPU gets out of its Goldilock temperature zone it generally "barfs" anyway. So unless the user sets the upper limit below that Papa Bear red zone, I think we'd see a CPU generated error which caused the OS to bug check and force the system to be rebooted before we'd see any negative impact from thermal down throttling on that CPU.

Yes I threw a bunch of caveats and some opinion in there because there are some "what if" type corner and low level cases that I'm leaving out and some other things I'm probably not thinking of right now. Add to that the only thing I can be certain of right now is I'm out of beer and there's always somebody smarter.

But I think I'll conctact one of the crash analysis guys Monday and see if they've seen any correlation to throttled down CPUs and fatal failures. It'd definitely be interesting to see.

 

[mikeblas]

So I'm not sure we'd ever be in a situation where a CPU would be under enough load and throttled down to its lowest clock speed for a long enough period of time as to present a problem.

Sure: when the HSF isn't attached corectly. From an OEM, this might be because it got knocked loose during shipping. Or for a DIY, it's because it's not installed correctly. As the CPU heats up, it'll throttle more and more to try and cool itself down.

So unless the user sets the upper limit below that Papa Bear red zone, I think we'd see a CPU generated error which caused the OS to bug check and force the system to be rebooted before we'd see any negative impact from thermal down throttling on that CPU.

Do you get BC's for this? I've never seen one; the system just shuts down.

Add to that the only thing I can be certain of right now is I'm out of beer and there's always somebody smarter.

Is there a bucket for beer-caused bug checks?

 

[der_jackal]

Sure: when the HSF isn't attached corectly. From an OEM, this might be because it got knocked loose during shipping. Or for a DIY, it's because it's not installed correctly. As the CPU heats up, it'll throttle more and more to try and cool itself down.

Sure, but it should throttle back up to the safest higher speed to complete operations as the load increases. If it can't, it really wouldn't be any different than running on an old P3.

Do you get BC's for this? I've never seen one; the system just shuts down.

Not one specific to overheating (I did kind of imply that), just something like the thread starter ran in to, lots of random errors where the registers values can be corrupted under those upper thermal operating limits.

I have seen some rare heat generated triple faults that cause the CPU scram and it reboots the system all w/o generating a debugger break / bug check., but those were only when running Virtual PC and really only because it accesses the CPU directly....long story.

So here again from personal experience in overheating CPUs, they all have caused some form of BSOD prior to crashing hard. I'd get up to about 65-70 degrees and poof BSOD.

Is there a bucket for beer-caused bug checks?

Sure, amazingly enough, all those bugs are assigned to me.

 

[PMildren]

I havent been overclocking either the CPU or GPU, mostly due to the issues I have been getting. Now the memory coiuld be at fault, as it is older and has been moved around alot. Wouldnt put it past the memory. Will run Memtest and let you know.

As far the the HSF issue, I lapped it and knew that it wasnt making "full" contact, but temps have never been very high even under higher loads. I have been monitoring with Riva TUner and have logs if anyone wants to see them.

 

[der_jackal]

I havent been overclocking either the CPU or GPU, mostly due to the issues I have been getting. Now the memory coiuld be at fault, as it is older and has been moved around alot. Wouldnt put it past the memory. Will run Memtest and let you know.

As far the the HSF issue, I lapped it and knew that it wasnt making "full" contact, but temps have never been very high even under higher loads. I have been monitoring with Riva TUner and have logs if anyone wants to see them.

Well the next step should be to try and build a repro scenario of this; Is there any pattern to when it happens? Is it within 10 mins of booting or does it take a while? Are you doing anything CPU intensive? Does it survive a 2+ hour Prime95 run? Those sorts of things can help narrow the failure down as well.

It never hurts when you're experiencing these random failures to make sure your HSF and CPU aren't making even a small bit of metal to metal contact. That's something that won't show in temperature values as it's not going to really cause a spike in temperatures.

 

[PMildren]

I will try and reseat them both to make sure. I do have a question regarding HSF. I installed the Thermaltake Big Typhoon on this build and wasnt real sure on how to make sure I didnt put too much pressure with out under tighttening it. I just tried to make sure there was even pressure by looking at the length of thread either left showing or thread showing on back side.

Will prbably take the HSF to a machine shop and have the face resurfaced for a flat surface.

 

[djnes]

I apologize for the unrelated post in advance:

Do you realize this is one of the very few places where the words "Dump Included" is actually a good thing?

 

[mikeblas]

"Dump included" is always a good thing! I've met lots of people who bitch about Microsoft software, but can't be bothered to send their crash reports in. They're invaluable in diagnosing problems and improving the products. I wish everyone would do it, all the time.

 

[der_jackal]

"Dump included" is always a good thing! I've met lots of people who bitch about Microsoft software, but can't be bothered to send their crash reports in. They're invaluable in diagnosing problems and improving the products. I wish everyone would do it, all the time.

Especially when you find that ~75% of submitted crashes are caused by 3rd party applications / drivers. It really helps Microsoft hammer on those companies to fix their software (glares at ATI).

 

[djnes]

Especially when you find that ~75% of submitted crashes are caused by 3rd party applications / drivers.

Not that I have any insider info...but purely on guessing, I'd put that percentage much higher.

 

[der_jackal]

Not that I have any insider info...but purely on guessing, I'd put that percentage much higher.

Well based on the top 100 issues reported (by WinHEC 2004)


Top 100 Reported Crashing Issues*
~5% Caused by Microsoft (we’re working hard to fix these!)
~10% Caused by Hardware Issues
~15% Caused by Unknown (too corrupted to tell)
Pool Corruptions
Memory Corruptions
~70% Caused by 3rd party code
Documentation issues (what can Microsoft do to improve these?)
Tools issues (what can Microsoft do to improve these?)
Just plain old bugs

 

[der_jackal]

I will try and reseat them both to make sure. I do have a question regarding HSF. I installed the Thermaltake Big Typhoon on this build and wasnt real sure on how to make sure I didnt put too much pressure with out under tighttening it. I just tried to make sure there was even pressure by looking at the length of thread either left showing or thread showing on back side.

Will prbably take the HSF to a machine shop and have the face resurfaced for a flat surface.

Determing how much "down force" to use is such a tough one. I remember getting a lttle wigged out when my MB bowed slightly on my old P4 3.2 when I dumped on the stock HSF. But in the end you need to crank it down pretty good to ensure you've got a good surface connection across the whole CPU face.

But here again, if you didn't it would likely show up as temperature spikes / overheating. Since we're not seeing that, it's really likely something else going on at the hardware level.