Win7 crashes, BSOD, startup failure...the works?

[Ebden~SPARTA~]

It all seems to be on a slippery downward slope. Errors abound.

 

I reset everything to stock earlier today, anticipating this time an intermittent heat issue with the memory and my OC.

 

At default BIOS settings it wouldn't post. that's stock.

 

I reset the OC to boot up, but ARMA2 crashed again. This time my bsod was 'bad_pool_handler'. From what little I googled, that's a really bad thing, in so far that it represents permanent damage somewhere in the machine. Tried a second time, it maxed the CPU and froze. I only barely got into task manager before the mouse quit responding.

 

Why don't I satisfy myself with a stick and a wooden hoop? Ungh.

[custard~SPARTA~]

Gigabyte may be on the money with RAM voltage, my Assus board will not set enough volts for my RAM and the RAM timings are way out at stock bios settings I have to do it manually.

 

Find out (via MOBO or RAM manufacturers forums) the voltage the RAM needs and set it manually along with the timings.

 

edit: timings (via your link) 5 5 5 15 and voltage 2.1

[Athlon64~SPARTA~]

Oh I saw that at one time on mine, The guy a asus said those were memory related error codes. pop out the ram on the secondary slots and see if problem goes away, if it does then its the ram.

Edited January 16, 2010 by Athlon64~SPARTA~

[Ebden~SPARTA~]

I have manually set the RAM v to 2.04 (Range of 2.0-2.1 for my mem). On my previous OC, it was 2.06 (all to run at 1066). I went down a step to help cut down on heat. BIOS properly identifies the memory timings.

 

I'm going to spend some time today anyway looking over different forums for why I have instability at stock settings.

 

FYI. It's running right now. As of yesterday, the moment I start ARMA2 everything goes to shit.

[Batwing~SPARTA~]

Ok Ebden, I really understand your frustration !

 

However, don t let your brain get frozen trying to solve issues are beyond your possibilities. Analize the situation.

 

Actually you said when you did set on Stock on your BIOS you had also serious issues. That should let you think for a moment. At the state of your situation I do not advice to play around with ANY OC setting. You need to have your machine running flawlessy at Stock. If that is not happening you need to get to the understanding something hardware is failing and there are no works around. You are just contribuing to get a huge headache to yourself.

 

You had this PC working fine in OC for a while I do understand. Remember while OC give us better performances, it also stress components way more than what is supposed to be. An hardware failure after OC stress is very possible to happen. It may happen any moment with no advice. It is possible you are suffering just that, unfortunate but real.

 

It is not clear if when you where using your OC PC you ever monitored your temps. It has no sense to try working on temps now, that is something you had to monitor earlier. If a damage is done, is done.

 

There are few tools you HAVE to use when you decide OC a system, they are:

1) CORE TEMP - Monitoring core temps (duh !)

2) CPU -Z - monitoring Speed and other important info about CPU and MB and RAM

3) GPU-Z - Monitoring info about your videocard (if you need)

4) Prime95 - Stress test for CPU and RAM. It put your CPU and RAM under maximum stress and test for errors.

 

While overclocking you have to use all of these tools together and monitor all of the system reaction. You open the temps monitors and then run Prime95 for 1 hour. Everytime it reports erros you get back on your BIOS and scale down your overclock. Your final objective is having a steady rock system in Prime95 at VERY acceptable temps.

 

More and more to discuss about overclocking, but this is a basic. Now going back to your problem.

 

Your PC has to work at stock. If it doesnt you are seriously compromised. Culprit will be between MB, RAM sticks and CPU itself (it could be power supply also, but these symptoms are not leading to that).

 

Again, if your MB is in warranty, do not waste other time and go immediately for an RMA. When you get the new one, do not even try to OC, check EVERYTHING is working fine at stock first. I really think you RAM is ok, but that could be the next step. However do not underestimate the possibility your CPU is damaged. Although I guess if CPU was damaged it was probably unable to work at all, you cant say for sure.

 

Try to save yourself some headache and follow my suggestions, you should be able to be up and running as soon as u get a new MB.

 

Good luck.

 

[Cylawyer~SPARTA~]

Well, horrendous times on my part here, so you have my sympathies.

 

Christmas present ASRock P55 Extreme MB, was DOA on 12/26/09, so I sent it back to New Egg for RMA. Just got RMA Thursday, 1/14/10, and new MB had no USB power and caused the CPU to have increasing temperatures until it would freeze at about 80 degrees Celsius approximately 30 seconds after loading windows each time. After third time, I shut it down so it would not damage CPU (It and the ram were also new Christmas presents).

 

Calls to New Egg result in, they won't RMA because past 30 days since original invoice. I think this defeats the purpose of the 30 day replacement/RMA terms because I just got the replacement and it was already non-returnable according to their policy. They could easily hold a stock of known bad parts and send them out for RMA replacements, knowing that they will be bad, but past the 30 days in all likelihood, leaving the customer's only resort to warranty RMA with manufacturer. ASRock won't advance RMA, nor guarantee a replacement RMA, instead they will put it on the bench and test it first, leaving customer with many more untold days or weeks of waiting for something that should have worked right in the first place or been advanced RMA'd with a replacement. Anyway, I suppose if this is a regular practice of New Egg it certainly answers how they keep from losing their collective shirts and offload bad boards coming back. ASRock techsupport and customer service suck btw. New Egg was nicer, but more rigid in their position.

 

I have not given up on nagging New Egg into a refund though.

 

 

So, I scraped the last of the loose change around the house and drove to CompUSA retail outlet and bought an ASUS P7P55D Pro. Slapped it in last night and it worked fine, but problems with OS then arose. Win XP would not boot, nor would it load a copy in another partition, so I pulled out Win 7 and tried to load it into an empty partition on another hard drive in the system, but it didn't like it because it was dynamic. Therefore I had to repartition the extra drive, where I had just saved a lot of my "do not lose" files and patches and stuff I really didn't want to get rid of and no way to recover. Then a clean install on a smaller partition on same drive succeeded in loading win 7, but buh-bye files!

 

Hmmm, many things designed to work well in XP don't work without a lot of work in Win 7. Took all day to figure out my bluetooth keyboard will never connect, even if the mouse did just fine. Then there was the issue that drivers for D-Link USB adapter DWA-130 are incompatible with Win 7. 2 hours of research online finally found that if I relocated my entire system to the end table next to the router and hard LAN it, I could update driver in Device Manager and fix issue (evidently 100's of other poor saps out there didn't figure that out and are languishing in cyber pergatory, or so it would seem from my forum search). Once done, I was able to relocate to normal computer location and USB adapter working fine since then. The other 12 hours today have been loading drivers and programs. Lots of fun.

 

At least I am up and running. It boggles my mind that after putting in the ASRock MB and the same CPU/RAM I have in this ASUS MB, the system had no issue in loading up Windows XP (other than MB causing overheat CPU and dying 30 secods in as mentioned above), and yet the ASUS, which is working wonderfully otherwise, somehow gummed up the works and the original OS, which should have been safe and secure on the harddrive it had always been in, somehow fekked up and crapped the bed.

 

Oh well, so long to all my years of files...

 

At least everything works now, but it is just like a completely vanilla build. Lots of drudgery installing and downloading stuff. My first build two years ago was smooth as butter in comparison.

[Athlon64~SPARTA~]

Thats bad. Anouther reasion I don't deal with new egg. Tiger direct has 1 year return policy and extended warranty.

Edited January 17, 2010 by Athlon64~SPARTA~

[Ebden~SPARTA~]

Ya' think with Win7 and ARMA2 we've(as a computer-using public) have finally pushed past the usefulness category for home computing?

 

I redid all my settings yesterday, had it running steady at stock, Prime95, monitored temps, etc, then began turning on the OC so I could run ARMA2. I left the CPUv a little higher than I had previously to help keep it smooth.

 

NOW she drops internet after a few minutes,while playing ARMA2. Win7 dropping web with ARMA2 is a known problem on the Bohemia forums (No solutions suggested from my quick browse). Once I could force quit, the other time forced reboot. At least I haven't seen a BSOD-yet.

 

I thought it might be related to me keeping STEAM and ARMA2 on a second partition separate from the OS, but a few other STEAM games run fine (Osmos, World of Goo-admittedly simpler than ARMA2).

 

Laundry now.

 

How about checkers? Anyone?

 

 

[Ebden~SPARTA~]

Ha, finally some preliminary good news!

 

All those bad ping readings seem to have gone. I've made several changes to my settings outlined below. Not sure which fixed it. Also, I haven't tried a game yet, so it could be superficial.

 

Setup Static IP, as I had done for BF2

 

Disabled RSS

 

Disabled IPv6

 

Disabled Nagle's Algorithm

[Cylawyer~SPARTA~]

That's good to hear.

 

[Batwing~SPARTA~]

Ebden,

 

glad you r getting somewhere after many hours of hard work and settings.

 

Just continue to keep an eye on temps, just to be safe.

 

I have Arma2 on a different partition but is retail, not Steam. don t know about limits with the steam version, but it would sound weird to me any incompatibility related to different partitioning.

 

Good Luck.

[Ebden~SPARTA~]

Ebden,

 

glad you r getting somewhere after many hours of hard work and settings.

 

Just continue to keep an eye on temps, just to be safe.

 

I have Arma2 on a different partition but is retail, not Steam. don t know about limits with the steam version, but it would sound weird to me any incompatibility related to different partitioning.

 

Good Luck.

 

I spoke too soon. The game seemed to run fine, up until my Bad_Boot_Header BSOD. I've read a little, and it means I either have toasted hardware, or I need to remove and reinstall all of my drivers.

 

Where would I find a list of installed drivers? I assume, once i find this list, that I'd just begin googling their names for replacements.

[Batwing~SPARTA~]

Well, after all of this , with several error messages in your history saying thhat most probably is a hardware failure (plus all the symptoms u got so far showing an hardware failure) you are still thinking is about drivers ????

 

I have to tell you, u r stubborn like a mule I don t know what is wrong to send that MB for RMA, but iif u did it 2 weeks ago when we started this thread, today you would have a new MB already

 

However, if u happily wanna mess with drivers, Ii would suggest you a BIG FORMAT and a clean installl of Windows. 7 would be the best choice because installl ready to work in 20 minutes and is able to set the best drivers for your machine out of the box. If you don t have it, good luck with XP or Vista.

 

Fresh install is the best optiion in case of drivers. It also guarrantee you r not missiing anything. Then, if iis not working, i hope u get to the point either u write a letter to the Pope to have a miracle or u RMA that MB

 

EDIT:

I made a quick search about the new error you are getting, because at first sight iit was reminding me something about "hard drive". In fact i found that usually that error message is related to the hard drive.

 

This now open a quite new window to your problem:

Either your MB controoller is so screwed up now is throwiing random error messages ( and I saw that on othher PC I fixed, at certan point, error message are just not relevant anymore, the machine crash for any possible reason...)

OR

Really your hardrive is hhaving an hhard time, which happens sometimes. Files can get corrupted for various reasons and HDD are subject to failures in time, they are the most stressed componennt on your PC and mechanical components under stress are subject to failure with time.

 

So, U have really nothing to lose at this point, giive a nice FORMAT to ure drive and reistall nice and clean, U could end up with a fortunate solution

 

Edited January 18, 2010 by Batwing

[Ebden~SPARTA~]

Well, after all of this , with several error messages in your history saying thhat most probably is a hardware failure (plus all the symptoms u got so far showing an hardware failure) you are still thinking is about drivers ????

 

I have to tell you, u r stubborn like a mule I don t know what is wrong to send that MB for RMA, but iif u did it 2 weeks ago when we started this thread, today you would have a new MB already

 

However, if u happily wanna mess with drivers, Ii would suggest you a BIG FORMAT and a clean installl of Windows. 7 would be the best choice because installl ready to work in 20 minutes and is able to set the best drivers for your machine out of the box. If you don t have it, good luck with XP or Vista.

 

Fresh install is the best optiion in case of drivers. It also guarrantee you r not missiing anything. Then, if iis not working, i hope u get to the point either u write a letter to the Pope to have a miracle or u RMA that MB

 

I haven't considered the RMA yet because I can't confirm if it's the Mobo or memory. The computer fails solely when running ARMA2, other than those two botched startups last week. I think my most stubborn activity is insisting to play ARMA2 over and over again.

 

BTW, I live in Newfoundland, getting anything here by mail takes twice as long, so if/when I do RMA that mobo, I'll be waiting for a good long spell.

 

I'll do the reinstall and reformat first, and I'll run Memtest on each stick of RAM overnight as well. How could I determine if the CPU is at fault. Could it be?

[Rooster]

I'll do the reinstall and reformat first, and I'll run Memtest on each stick of RAM overnight as well. How could I determine if the CPU is at fault. Could it be?

Could be your PSU also. When this happens one might experience the symptoms you're describing (random reboots, BSOD, untraceable memory errors). The capacitors in PSUs age over time and their max output will droop. What brand, wattage, and how old is your powersupply?

 

Edit: I realize no one really asked for your PC details. Could you list the following? I know some of your parts already, but having them posted here will help us all t-shoot.

 

Case (interested in airflow so post any added fans)

CPU (any overclock?)

motherboard (gigabyte p45 series (UDP3?))

RAM (DDR2, speed, volts... PS ddr2 is supposed to be compatible with 1.8vdimm)

Video card

CPU cooler (stock?)

Powersupply (this might just be the problem!)

Hard disks / optical drives

Edited January 18, 2010 by Rooster

[Ebden~SPARTA~]

Thanks to the thread posters helping me sort through this. It's frustrating to use technology beyond my understanding which is also so very addictive.

 

Case (interested in airflow so post any added fans)

Standard-ish full size steel case. One 80mm fan on the access panel running as exhaust.

 

CPU (any overclock?)

Intel E8400 (series E0), 3.0G, OC to 3.6 @1.2175v (9x400)

 

Mobo

Gigabyte EP45-UD3L v1

BIOS version F4

 

RAM

G.Skill PC2-8500 2Gx2

F2-8500CL5D-4GBPK

Manually set 5-5-5-15, v@2.1

 

Video card

Palit Nvidia 9800GT No OC. I bump the fan speed to 100% using VTune when I play ARMA2

 

CPU cooler

Cooler Master TX2 (link to TX3)

 

PSU

SPI 450W ATX V2.2 (I think, writing this from work)

 

HD

WD Caviar Blue 320G 16mb Cache

 

Other

1x CD-R

1x DVD

1x 3.5 floppy

1x LaCie external 500G HD firewire (or 250?)

1x PCI card w/2x firewire

[Rooster]

Case (interested in airflow so post any added fans)

Standard-ish full size steel case. One 80mm fan on the access panel running as exhaust.

Okay, lets start by flipping that fan around. Typically it's better to have positive pressure in your case. Meaning, more air blowing in... Otherwise you will have a lot of dust and uncontrolled airflow in your case.

 

CPU (any overclock?)

Intel E8400 (series E0), 3.0G, OC to 3.6 @1.2175v (9x400)

Normally I would respectfully request that you turn this back to stock speed and vcore until we get a handle in this. We have an oppertunity to test this though... So if you can pass 20 loops of linx, then I would consider your OC stable. As Bat touched on, running coretemp or realtemp is very important when stress testing. I actively help uncleWebb test real temp so give that a go and report your temps back (high/low) after 20 loops of linx.

 

Mobo

Gigabyte EP45-UD3L v1

BIOS version F4

I know this bios revision to be stable as I have a fileserver running windows 7 with this board and an e8400 cpu.

 

 

RAM

G.Skill PC2-8500 2Gx2

F2-8500CL5D-4GBPK

Manually set 5-5-5-15, v@2.1

The website reports this RAM as rated for 1.8vdimm. Is there a reason you've bumped it up to 2.1vdimm? Normally this would be pretty safe, but I think your case has pretty restricted airflow in the configuration it is in now. So this may be an issue.

 

 

Video card

Palit Nvidia 9800GT No OC. I bump the fan speed to 100% using VTune when I play ARMA2

This should be pretty reasonable. My thoughts are how loaded your PSU is at this point. Could ramping up the speed, and the increase in case temps, all be related to the PSU starting to get less efficient? The hotter your PSU is, the less efficient it is at outputting its max wattage. Most PSU's are rated to output max @ 40C.

 

 

CPU cooler

Cooler Master TX2 (link to TX3)

Due to the size of this cooler, it could be a little bit stressing on your board. Could you check the mount to make sure it's not stressing your board too much? Even simply pulling up on the cooler, feeling the board flex a bit toward you, can sometimes revive flaky motherboard paths that are under a lot of stress near the CPU socket.

 

PSU

SPI 450W ATX V2.2 (I think, writing this from work)

Looks like you have two, 18amp rails on your PSU. This should be pretty good for your setup. A couple things to check would be the CPU 4/8pin connector and any PCIe connectors. Sometimes people connect the wrong plug into the motherboard since it is keyed the same. If this were the case then it would be plausible that the system is using just one rail. I think this is a little less likely... Over time a 450watt PSU can act more like a 300watt PSU due to capacitor aging. If you have a way to test a new PSU, this may prove to be a worth while test.

 

HD

WD Caviar Blue 320G 16mb Cache

Other

1x CD-R

1x DVD

1x 3.5 floppy

1x LaCie external 500G HD firewire (or 250?)

1x PCI card w/2x firewire

I'm looking at my favorite PSU calculator and it seems to think that your system needs about 260watts of power. If I up your PSU age to one year old (20% cap aging) and calculate 100% load, it says you could use about 322watts. Still in the clear in terms of wattage. In theory your amps could go beyond its output it the rails are not split up properly, but this isn't really a typical problem. My hunch is that there could be a temp + overclock stability issue going on here.

 

hope this helps... if not I'll run this by my other engineer colleagues. We will get this solved.

 

 

[Ebden~SPARTA~]

Okay, lets start by flipping that fan around. Typically it's better to have positive pressure in your case. Meaning, more air blowing in... Otherwise you will have a lot of dust and uncontrolled airflow in your case.

 

Done.

 

Not sure what you saw to get the memory v figs. I pulled mine from here.

Normally I would respectfully request that you turn this back to stock speed and vcore until we get a handle in this. We have an oppertunity to test this though... So if you can pass 20 loops of linx, then I would consider your OC stable. [/url].

 

Run first test:

 

For this test, I had set Priority class to real-time. Data Align (KiB) was 4, Linpack32max prob was 15500(does it matter for 64win7?), Memory for OS (MiB) was 5.

 

Currently reset settings of test to Priority class normal, Memory (MiB) on the front to 2048. Hoping I did something wrong with the first set. Pending the second failure, I'll reset BIOS, and set the memory v to 1.8 or 2.1, depending on what it resets to. (according to CPU-Z, I can run at 1.8v for anything less than 1066, but 2.1 for the max rated speed)

[Batwing~SPARTA~]

Good job Rooster, very good insight and analisis.

 

Actually, what jump to my eyes right away is that you are pushing an overclock on CPU and RAM (RAM just changing voltage, but just that generate way more heat in fact) using a Case not qualified for.

 

You do not have enough airflow to overclock and having a dedicated third party CPU cooler does not mean it can make a miracle. Juste there is not enough airflow inside to keep your stuff in a safe temp under stress.

 

However, observing this fact is not leading to a solution per se. If working for long time overheating has already compromized your MB North bridge (responsible for the Memory RAM exchange with your CPU) and / or your CPU itself, there is nothing you can do now.

 

Recapping all the symptoms really I would not think is a PSU issue. Although is probably on a wattage limit, your system is not too demanding and PSU failure brings to several random shut downs, you don t suffer by this at all.

 

From your previous tests your RAM souned working fine as sticks, problem raised regarding some MB slots.

 

You reported issues to have your system running at stock config, which is Very Bad. Yiou shoudl be able to run at stock flawlessy, CPU and RAM stock, no voltage or other modificators.

 

You ask how to check if your CPU could be compromized, unfortunately in your situation is not a simple task. If you continue to test your CPU on a compromized MB you will continue to have misleading error messages.

Theoretically you should get a brand new MB, install your CPU, keep it stock and run tests on it. If prime95 doesn t fail or any other CPU stresser you want to use, it means the CPU is fine (and I do belive so..)

Your CPU will not stay fine forever if you continue to use it improperly on a damaged MB because all of this is connected. This kind of stress and improper shut down procedures for sure are not going to help any guest of your case, naming CPU, RAM, HardDrive.

 

Also, every time you have an improper shut down you are playing "russian roulette" with your hard drive, risking to compromize sectors, FAT etc, till one day you might have a failure from your HD simply because of that.

 

It is like driving a car going with one cylinder less... do not expect that is good for the rest of your mechanical components.

[Rooster]

Looks like I pulled up the wrong values for your RAM (was looking at the tridents). So you should be OK there. What was your CPU temp like? Could you post a screen or just look at the min and max values?

 

As for linx settings, you can accept the defaults and just run a 20 pass test. At a minimum you should be able to pass all twenty at stock speeds.

 

Edit: linx uses the same linpack algorithm that intel engineers use to spec out the max stress load possible on your CPU. A typical 5 loop test usually indicates moderately stable systems. 20 is pretty much regarded and accepted as fully stable. Like Bat said running at stock speed you should be able to run linpack/linx indefinitely. The temps are important because if the CPU is touching above 70c then it's getting pretty close to the danger zone while overclocked. At about 83c your e8400 will start to thermal throttle, in which case realtemp will let you know that the cpu is throttling. Technically you can run at stock speeds and throttle but it's so hot that really no one recommends doing it for very long.

 

R

Edited January 18, 2010 by Rooster

[Ebden~SPARTA~]

 

Thanks again for the extra detail. My max temp after this test was 68c, min 33c. After the test, the temp fell back to 37c. (GPU never above 45ish).

 

I'm about to shut down, reset CMOS, set my memory voltage as appropriate, and try the test again with everything else at default (minus anything else that identifies the OS as 64, there was something I remember).

 

When I'm back up, check in TS tech help channel. I'll be in there until about 6pm East coast time, then I may be back up an hour later.

[Ebden~SPARTA~]

I've reset CMOS and made the following changes to default.

 

Changed CPU clock from 8.5x333 (2.83ghz) to rated CPU 3.0Ghz @ 9x333, auto v to 1.1875

Set DRAMv to 2.1 (Auto settings placed memory at 1066)

Enabled CPU warn temp 80c

Enabled CPU fan fail warn

 

I restarted machine. Wouldn't boot. Went auto to Win 7 startup diagnostic. Ran for 40 minutes. Concluded that it was an unknown error.

 

Restart.

 

Played a little ARMA2, because I needed my fix. One crash-a CPU-max freeze. Managed to force-quit. Max temp while playing was 46c.

 

Ran LinX. Failed before 5 min. Max temp 60c. Remember, the following photo is for a test at default settings, except for the changes above.

 

Going to bed now, considering if I can RMA everything in the computer and play minesweeper for a month on my laptop.

[Rooster]

Well, at this point it looks like there is some unstable/bad hardware. I wish you lived closer because I'd swap you out a s775 mboad, ram, PSU to try out first.

 

Last question, if you underclock your ram to ddr 400 does your stability increase? What about when you underclock your CPU? If your CPU became stable at a lower speed, then we could guess that the CPU was bad. Same for RAM...

[Athlon64~SPARTA~]

Oh ya, you need at least 4 120mm fans on that case. 1 on the cpu, 1 on the vid card, 1 in the front, 1 in the back, i would put in a 80mm chimmney fan on top as well to remove the heat as it collects at the top of the case.

 

Get out the 4" hole saw and

put 2 in the side

1 in front

1 rear

80mm on top center of case intween the p/s and cd/rom drives

Get some kind of cooler on those hard drives as well, or put it in the air stream from the one in the front to blow across it.

 

Edited January 19, 2010 by Athlon64~SPARTA~

[Ebden~SPARTA~]

Get out the 4" hole saw and

 

Dude, she's a steel case! There's no hole saw of mine going to be ruined with that chore. Also, at max use, 67 is still not wretched. I'll start testing today by UC on the PSU, and if that works then UC to 800 on the memory (at 1.8v).