Diagnose my ****ty PC

[godly_music]

Hi folks,

Lately I've been getting crashes and weird hang-ups. It all started with BSOD's referencing dxgmms1.sys, SYSTEM_SERVICE_EXCEPTION and SYSTEM_THREAD something, 0x3B, 0x1E, 0x7E.. all graphics related. Once it even BSOD'd right as I was uninstalling the graphics drivers.

I removed the GPU's (stable and well-cooled) overclock, and lo and behold, the system started to freeze seconds to minutes after I logged onto Windows. Just a hard lock-up with looping sound. The same thing happened on a brand-new install of Win 7 moments after logging on by the way, so it's definitely hardware related. Leaving the graphics card disabled and using software mode seemed to have fixed that for now.

And once, last night, I got a DRIVER_IRQL_NOT_LESS_OR_EQUAL tcpip.sys 0xD1 BSOD. Just letting me know it ain't over yet.

At first I was going to replace the graphics card, but after the last one I'm thinking this is a mainboard failure. Just days prior to all this I've been running demanding games for hours without crashing. It usually happened when I was watching youtube videos. Memtest ran fine, and the fact that it's not breaking down during peak times leads me to believe the PSU is fine as well.

CAN a failing graphics card, if it's still being used, create a tcpip.sys error? I doubt it, but I had to ask.

[tp4tissue]

Hi folks,

Lately I've been getting crashes and weird hang-ups. It all started with BSOD's referencing dxgmms1.sys, SYSTEM_SERVICE_EXCEPTION and SYSTEM_THREAD something, 0x3B, 0x1E, 0x7E.. all graphics related. Once it even BSOD'd right as I was uninstalling the graphics drivers.

I removed the GPU's (stable and well-cooled) overclock, and lo and behold, the system started to freeze seconds to minutes after I logged onto Windows. Just a hard lock-up with looping sound. The same thing happened on a brand-new install of Win 7 moments after logging on by the way, so it's definitely hardware related. Leaving the graphics card disabled and using software mode seemed to have fixed that for now.

And once, last night, I got a DRIVER_IRQL_NOT_LESS_OR_EQUAL tcpip.sys 0xD1 BSOD. Just letting me know it ain't over yet.

At first I was going to replace the graphics card, but after the last one I'm thinking this is a mainboard failure. Just days prior to all this I've been running demanding games for hours without crashing. It usually happened when I was watching youtube videos. Memtest ran fine, and the fact that it's not breaking down during peak times leads me to believe the PSU is fine as well.

CAN a failing graphics card, if it's still being used, create a tcpip.sys error? I doubt it, but I had to ask.

try a different graphics card?

[The Mad Professor]

It could be a BiOS issue. Have you tried flashing the BiOS?

[godly_music]

Trying a different GPU tonight, will see how that turns out.

No, I haven't flashed the BIOS in years - but it's also been running fine for years.

[IvanIvanovich]

What is the hardware specs? Do you have any overclock, specifically any overclock on the pci-e bus? If so remove the OC, set all to recommended default in bios but switch back to ahci if applicable. Run a hdd scan for whatever make you have... unless it's an SSD since there really aren't any such tools available still. Then I would suggest to run admin cmd, chkdsk /B on your system drive. That should prompt reboot so it can scan before OS loads... it might take some time. After that finishes I would run sfc /scannow just to make sure there is no corrupt system files.

If everything checks out I would then make sure all your hardware drivers are up to date, then run realbench on loop for awhile to see if you can get it to bsod again. If it does you can use bluescreenview to give you some human readable insight into what is the actual cause in your crash dumps.

[osi]

How's your hard drive or ssd health?

[godly_music]

Asus P5Q Pro
Q9550 @ 1.1 vcore
2x2 GB Corsair XMS2 DDR2-800
NVidia GTX 275
Cooler Master RealPower M620
2 Harddrives

The GPU was overclocked, I took it out. The CPU is undervolted and running fine, might bring that back up just to have it all on factory specs. Harddisks check out fine, sfc /scannow is unlikely since I had the same lock-ups immediately after reinstalling drivers on a fresh Win 7 install. The PC is from 2009 so drivers are all mature and have never caused issues. I'll give realbench a go.

Edit: Just had another one. Here's the BSODs I still have saved. Note that when I was still using hardware graphics, all BSODs hinted at graphics. Now that that is no longer present, they're all about the network. I'm thinking both my graphics and network are fine and I need to get a new mainboard.

==================================================
Dump File         : 091614-8346-01.dmp
Crash Time        : 16.09.2014 20:58:58
Bug Check String  : SYSTEM_SERVICE_EXCEPTION
Bug Check Code    : 0x0000003b
Parameter 1       : 00000000`c0000005
Parameter 2       : fffff800`02fe7efc
Parameter 3       : fffff880`051dae10
Parameter 4       : 00000000`00000000
Caused By Driver  : raspptp.sys
Caused By Address : raspptp.sys+2313e10
File Description  :
Product Name      :
Company           :
File Version      :
Processor         : x64
Crash Address     : ntoskrnl.exe+75bc0
Stack Address 1   :
Stack Address 2   :
Stack Address 3   :
Computer Name     :
Full Path         : C:\Windows\Minidump\091614-8346-01.dmp
Processors Count  : 4
Major Version     : 15
Minor Version     : 7601
Dump File Size    : 271.579
Dump File Time    : 16.09.2014 21:00:10
==================================================

==================================================
Dump File         : 091614-9843-01.dmp
Crash Time        : 16.09.2014 07:54:18
Bug Check String  : DRIVER_IRQL_NOT_LESS_OR_EQUAL
Bug Check Code    : 0x000000d1
Parameter 1       : 00000000`00000010
Parameter 2       : 00000000`00000002
Parameter 3       : 00000000`00000001
Parameter 4       : fffff880`01a766c0
Caused By Driver  : tcpip.sys
Caused By Address : tcpip.sys+756c0
File Description  :
Product Name      :
Company           :
File Version      :
Processor         : x64
Crash Address     : ntoskrnl.exe+75bc0
Stack Address 1   :
Stack Address 2   :
Stack Address 3   :
Computer Name     :
Full Path         : C:\Windows\Minidump\091614-9843-01.dmp
Processors Count  : 4
Major Version     : 15
Minor Version     : 7601
Dump File Size    : 271.579
Dump File Time    : 16.09.2014 07:55:17
==================================================

[dorkvader]

Graphics cards should NOT cause BSOD in TCPIP.sys driver.

The last two (thanks for posting more BSOD info) were both caused by networking stuff. Do you have a spare MB to try? Might want to reinstall the ethernet driver.

Try volting your CPU back up. Sometimes undervolted CPU will cause panic somewhere down the road.

I agree that SFC probbaly wont help, and HDD issue is unlikely.

Right now with all the TCP/IP and "raspptp" (what is this? network to rpi?) makes it look like either hardware issue on ethernet chip, ethernet driver failure or virus.

[tp4tissue]

at this point... you might consider jumping from p5q.. anything around 7 years isn't worth using, UNLESS you're uber strap for cash..

[godly_music]

That's what I was thinking also. I'm grabbing a spare motherboard tomorrow. The last crash happened after upping VCore back to factory setting. Raspptp.sys has something to do with peer-to-peer, I've never seen it before today. With the PC being as old as it is, and having used the same drivers for years, I'm gonna assume it's hardware.

Thanks for the help folks!

As far as ancient hardware goes.. Some people still use the E6600 and pair it up with a beastly graphics card. I love my C2Q and think it's pretty speedy for its age, certainly enough to play most games still.

[dorkvader]

That's what I was thinking also. I'm grabbing a spare motherboard tomorrow. The last crash happened after upping VCore back to factory setting. Raspptp.sys has something to do with peer-to-peer, I've never seen it before today. With the PC being as old as it is, and having used the same drivers for years, I'm gonna assume it's hardware.

Thanks for the help folks!

As far as ancient hardware goes.. Some people still use the E6600 and pair it up with a beastly graphics card. I love my C2Q and think it's pretty speedy for its age, certainly enough to play most games still.

i had a macpro 1,1 with dual xeons and 4gb of extremely expensive server ram. Add a modern GPU and it games great, especially in the winter when the rest of the apartment is 50 degrees F. I was thinking about upping the RAM to 8GB but that's expensive.

Anyway, notbad for a computer from 2006. It's all about the GPU nowadays if the rest of your hardware is solid. Hard to beat workstation-quality stuff, though it took me forever to scrape that hideous apple logo off the side.

[godly_music]

Urgh I'm so tempted to grab another 4 gigs of RAM also just to play those juice hungry (often horribly optimized) games, but then comes the realization that years later DDR2 is still damn expensive and I'm putting a bigger tank on a sinking ship.

Server hardware sounds pretty sweet if you happen to come across it. The only one I was ever granted the luxury to have was a 800 MHz Pentium with a bunch of weird harddisks, tape storage and the sound profile of an elephant. Not bad for a Linux box you put in the basement, far away from where people live.

[tp4tissue]

That's what I was thinking also. I'm grabbing a spare motherboard tomorrow. The last crash happened after upping VCore back to factory setting. Raspptp.sys has something to do with peer-to-peer, I've never seen it before today. With the PC being as old as it is, and having used the same drivers for years, I'm gonna assume it's hardware.

Thanks for the help folks!

As far as ancient hardware goes.. Some people still use the E6600 and pair it up with a beastly graphics card. I love my C2Q and think it's pretty speedy for its age, certainly enough to play most games still.

I honestly doubt, those people are that involved with their PCs..   

I can accept that c2d is fine for internet/pron/ counterstrike/ dota


if that's all you do... then yea.. c2d..


But try converting a video on c2d... I just want to kill myself..   got a q66 doing 3.9ghz, converting a 1080p video took 9 hours..   it only took ~1 hr on 4.8ghz 3770k..

[dorkvader]

That's what I was thinking also. I'm grabbing a spare motherboard tomorrow. The last crash happened after upping VCore back to factory setting. Raspptp.sys has something to do with peer-to-peer, I've never seen it before today. With the PC being as old as it is, and having used the same drivers for years, I'm gonna assume it's hardware.

Thanks for the help folks!

As far as ancient hardware goes.. Some people still use the E6600 and pair it up with a beastly graphics card. I love my C2Q and think it's pretty speedy for its age, certainly enough to play most games still.

I honestly doubt, those people are that involved with their PCs..   

I can accept that c2d is fine for internet/pron/ counterstrike/ dota


if that's all you do... then yea.. c2d..


But try converting a video on c2d... I just want to kill myself..   got a q66 doing 3.9ghz, converting a 1080p video took 9 hours..   it only took ~1 hr on 4.8ghz 3770k..

You should have seen Ron's video editing machine at keycon. It has RAID 0 SSDs and a beast of an overclocked CPU with watercooling.

[tp4tissue]

That's what I was thinking also. I'm grabbing a spare motherboard tomorrow. The last crash happened after upping VCore back to factory setting. Raspptp.sys has something to do with peer-to-peer, I've never seen it before today. With the PC being as old as it is, and having used the same drivers for years, I'm gonna assume it's hardware.

Thanks for the help folks!

As far as ancient hardware goes.. Some people still use the E6600 and pair it up with a beastly graphics card. I love my C2Q and think it's pretty speedy for its age, certainly enough to play most games still.

I honestly doubt, those people are that involved with their PCs..   

I can accept that c2d is fine for internet/pron/ counterstrike/ dota


if that's all you do... then yea.. c2d..


But try converting a video on c2d... I just want to kill myself..   got a q66 doing 3.9ghz, converting a 1080p video took 9 hours..   it only took ~1 hr on 4.8ghz 3770k..

You should have seen Ron's video editing machine at keycon. It has RAID 0 SSDs and a beast of an overclocked CPU with watercooling.

Video editing with only 1 cpu?  was it the 6 core?

[dorkvader]

i dunno but he packed a ton of power into such a small case. I liked it.

[IvanIvanovich]

I suppose what I would do next is disable the onboard NIC in bios and see if that resolves anything. If so you could always pick up a used intel gig-e nic off ebay for a few bucks. Don't discount the possibility of file corruption even on a fresh instal. I've seen it happen, and it's really not outside the realm of possibility especially if you used optical media.

If you want some older server type stuff as an upgrade it's definitely not out of reach. You can pick up some dual xeon 1366 stuff for less than $250 for a 2xcpu+motherboard combo. Even if you're not interested in rackmount it's sometimes cheaper to pick up a whole system when it comes with memory. But do be careful as use of non atx family motherboards is pretty common in rack servers. Great care is needed if there is no picture of the inside in the listing.

[dorkvader]

I suppose what I would do next is disable the onboard NIC in bios and see if that resolves anything. If so you could always pick up a used intel gig-e nic off ebay for a few bucks. Don't discount the possibility of file corruption even on a fresh instal. I've seen it happen, and it's really not outside the realm of possibility especially if you used optical media.

Very uncommon in my experience, but can happen even with USB install media. Usually a sign of failing sata/IO capabilities on the motherboard, but i can sometimes "just happen" on an otherwise good system. If you get this, doublecheck your install media (install fresh copy to USB, or burn a new optical disc and re-verify) and reinstall again.

[godly_music]

I just find it strange that at the same time, I start to get hard lock-ups (typical for a failing HDD motor among other things), dxgmms1.sys errors pointing at the graphics card and two different errors pointing at the ethernet.

After a bloody ordeal with getting the new mainboard flashed and setting it all up, I was greeted by yet another screen freeze half an hour or so after I installed all the drivers.

Since you've said (and I agree) that a graphics card issue should not bleed over and cause network issues, the next thing I'll be replacing is the PSU.

[dorkvader]

I just find it strange that at the same time, I start to get hard lock-ups (typical for a failing HDD motor among other things), dxgmms1.sys errors pointing at the graphics card and two different errors pointing at the ethernet.

After a bloody ordeal with getting the new mainboard flashed and setting it all up, I was greeted by yet another screen freeze half an hour or so after I installed all the drivers.

Since you've said (and I agree) that a graphics card issue should not bleed over and cause network issues, the next thing I'll be replacing is the PSU.

what psu? It might be okay.

Keep in mind that all the issues I saw were with TCP/IP stuff which is not the same as the ethernet.

edit:
Cooler Master RealPower M620

might want to upgrade that.

[godly_music]

Okay, here's the update:

It's not the PSU either.

Pretty much only leaves the graphics card. How in the hell does that cause these weird BSODs? Or did the mainboard just decide to break at the same time?

[tp4tissue]

Okay, here's the update:

It's not the PSU either.

Pretty much only leaves the graphics card. How in the hell does that cause these weird BSODs? Or did the mainboard just decide to break at the same time?

everything is plugged into PCIE..

Graphic card being pretty high current, could mess with your other stuff.

[godly_music]

Lookie here.. this is my PSU. Never had a busted cap before but that looks suspicious.

[IvanIvanovich]

Yes i would say that is clearly bulging. Time for a seasonic.

[godly_music]

I bought one. It's already plugged in. And a R9 270X on the way.

[tp4tissue]

You could probably just fix that..

[godly_music]

It would be prudent to replace all the caps, and my soldering skills aren't that good. Also.. the PSU outlived its warranty by barely a week or two before this whole ordeal started, that tells me it's probably engineered to blow up in my face soon.

[tp4tissue]

It would be prudent to replace all the caps, and my soldering skills aren't that good. Also.. the PSU outlived its warranty by barely a week or two before this whole ordeal started, that tells me it's probably engineered to blow up in my face soon.

well,  if that cap is an "important" cap.. they usually have more than one for safety.  so swapping just-it out will probably be ok.. though if the rest of your setup is expensive.. i suppose the risk is not worth it

[dorkvader]

It would be prudent to replace all the caps, and my soldering skills aren't that good. Also.. the PSU outlived its warranty by barely a week or two before this whole ordeal started, that tells me it's probably engineered to blow up in my face soon.

well,  if that cap is an "important" cap.. they usually have more than one for safety.  so swapping just-it out will probably be ok.. though if the rest of your setup is expensive.. i suppose the risk is not worth it

if that one failed then the rest will likely fail sometime soon. It's best practise to replace all the "low end" chinese capacitors while you're in there for safety.

Also putting a second cap there for redundancy is not any safer and a huge waste of resources. I recommend studying the following diagram to get a better idea of how these things work.

description: http://www.pavouk.org/hw/en_atxps.html

if there is one or two main filter caps means something else entirely.

[godly_music]

Thanks for all the help, guys. Bottom line: Aging PSU managed to kill the graphics card. Just glad I took it out before anything else blew up.