My new PC is cursed!

[Npl]

Besides the SSD dying in 1 month, taking the company with it (just OCZ just time, before I killed Commodore the same way) I still have troubles with stability.
Not that it crashes but the Graphics driver sometimes bogs down, rendering everything with ~4FPS. It shows in the event-log as Event 14 from "nvlddmkm".

Intel i7-4771
16GB Geil DDR3 (forgot exact name)
Gigabyte GA-Z87MX-D3H
MSI GeForce GTX 760, N760 TF 4GD5
Seasonic X-460

* The error is not exactly reproduceable, once it appears it requires a shutdown to fix. Doesnt seem to depend on heat - most of the time there after the bootup.
* No overclocking whatsoever
* Fresh & clean windows install.

Its either some obscure driver bug (using newest drivers) or some weird hardware problem. Maybe graphics card, maybe pcie.. no idea.
Heck, I though it was gone for good after disabling some bios option for faster boot, but its still lurking around.
Only think left is to change the gfx card and/or change the mobo.

Anyone know any tools to eg. drop down pcie speed for testing?
This bug still common with actual hardware and drivers?

 

[homerdog]

Have you tried the lastest WHQL driver?

Are both of the GTX760 4GB models?

 

[karlotta]

GPUz has a helpfull setting for seeing your PCI-e lanes in use. Tho not a tool to reduce on the fly.

Memtest the RAM for Seven or more full runs? 16GB should take a couple of days...But/only way to know on RAM. More volts may help here.

When it does happen is there a huge amount of IO activity? H/W faults? Cant you just shut down the offending processe(s) and go about your biz. That is if you can get to the taskmanager.

Then there is the "USB" is possessed issues, a funky controller, or Device can hose a session.

Do use Forceware?

 

[Npl]

drivers are 331.82, just plain drivers, no 3D or Gforce experience stuff. I went through 2 revisions I think, all exhibiting the same problem.

And its just one GF 760 with 4GB.

No, nothing running, no usb device attached (appart from logitech keyboard + mouse), I can even spot the issue when dragging windows around, no need to even start a game.

Can try downclocking the RAM, but the problem is, the issue can be gone for seamingly no reason and reappear the same

 

[Grall]

My new haswell PC bluescreened a lot, randomly, when running my DIMMs anywhere near full clip (they're rated at 2400MHz) and then after updating the BIOS it worked. Almost, anyway. I still managed to bluescreen the damn thing by running 7Zip's benchmark mode so maybe it needs a few more BIOS revisions, I dunno.

For now, I stay away from 7Zip, and the system is stable.

I'd suggest that if you get log events from your video card that you replace your video card with something else. Try your video card in another PC. Things of that nature.

 

[Arwin]

My last system was unstable because the power chord connector to the external HDD drive bay was loose ... replacing that cable solved everything.

In your case, I would definitely start with taking out the GPU and putting in an old (or just really cheap) one to test if that makes a difference (I bought at a store which offers decent support, so they did that with me in the store, but I have a spare cheap GPU card around myself for these things if necessary).

Next, I would take out half of the RAM, see if that helps, and then take out the other half of the RAM and see if that works. That's a lot faster than testing each RAM bit ...

 

[digitalwanderer]

Does the GPU's fan do anything when it does the fps dip? Like rev up or rev down?

 

[Bludd]

Have you installed the Intel chipset drivers?

 

[Florin]

Although nvlddmkm issues can be indicative of a hardware issue, I recently ran into another PC that had regular nvlddmkm events and they turned out to be a side effect of a rogue bitcoin miner.
Much like a virus, this miner ran invisibly in the background, and it would reinstall and reactivate itself at some random interval between about 10-60 minutes after every reboot.

In any case, as a first step you might want to try a GPU activity monitor (like GPUShark from the GPU-Z guys) to check which programs are actively using the GPU. See if there are any unexpected processes there. If not, it may still highlight some correlation between specific programs using the GPU and the events popping up.

 

[Npl]

(Intel) Drivers are fine and no viruses (Pretty much 3 times a clean install thanks to having to RMA the hdd).
dig: nothing of the sort and fps remain stable until reset or powerdown - its basically a machine with two personalities and you never know which one you get when you boot.

It managed to reproduce the issue in the bios bootmenu once so its very likely to be a gpu, motherboard or pcie issue (thanks to that beeing integrated into the cpu leaving the cpu as issue aswell).

Narrowing down is hard because the thing might work fine for weeks between 2 sluggish encounters - without changing a single thing.

 

[Mize]

Whenever I've had ghost in the machine oddities it's invariably been power related. Sagging rails on the PSU, droop on the motherboard Vcpu or dirty lines (such as being on the same circuit as a refrigerator that pulls the circuit low every time the compressor kicks on).

 

[Sinistar]

This sounds like the exact problem I was having after installing an 290X. Sometimes it would boot and run fine, but most of the time it would boot and be a stutter fest, it even started happening in the BIOS. I switched to the second bios on the motherboard, and all was fine, so knowing this, I reset the original BIOS, and have not had a problem since.

Edit: I think it might have something to do with a UEFI setting in the BIOS.

 

[Grall]

Whenever I've had ghost in the machine oddities it's invariably been power related.

My new(ish, by now, since it's roughly half a year old by now) behaves much better after flashing the BIOS, and therefore handling my RAM sticks in a more compatible way most likely. May have been other behind-the-scenes compatibility stuff changing as well, I dunno.

PCs, especially high-end DYI boxes, can be quirky when they're brand new and untested. Everything's rushed to market, development costs must be recouped, etc. *sigh*

Consoles do have a certain appeal from this respect... Not that I'll ever give up on PC gaming just because of some crashes every now and then!

 

[Mize]

Absolutely one should be on the latest BIOS.

 

[Grall]

ASUS built-in "EZ" flasher is a funny character indeed. It can read the directory structure of NTFS-formatted devices, such as harddrives, but it can't actually read the FILES THEMSELVES (and it doesn't reveal that in any way, lol!), so when you select your BIOS file you downloaded to your harddrive it tells you that the file is not a valid UEFI BIOS...

Jesus effing christ.

No, you need to search for the error message online, find some god damned forum posts somewhere which explains all this, and tells you to FAT-format a flash stick and put the file there instead. Ackpth!

 

[Mize]

ASUS built-in "EZ" flasher is a funny character indeed. It can read the directory structure of NTFS-formatted devices, such as harddrives, but it can't actually read the FILES THEMSELVES (and it doesn't reveal that in any way, lol!), so when you select your BIOS file you downloaded to your harddrive it tells you that the file is not a valid UEFI BIOS...

Jesus effing christ.

No, you need to search for the error message online, find some god damned forum posts somewhere which explains all this, and tells you to FAT-format a flash stick and put the file there instead. Ackpth!

LOL, so THAT'S why it never works. I've just been using their in-windows flash utility instead.

 

[Bludd]

LOL, so THAT'S why it never works. I've just been using their in-windows flash utility instead.

Same here. Asus Update has worked fine for me for over 10 years now with Asus mobos.

 

[Npl]

Hmm, now my pc crashed midgame and i couldnt power it on again.
I disassembled till it worked again and it was the gfx card that caused the issue - it worked after reinstalling!

I begin to believe that the pcie-connection is the issue, and possibly the vibration from the hdd causes the connection to move around.

 

[Grall]

Does your HDD vibrate a lot? ...Coz, it shouldn't. PCIe connectors really are quite firm and snug - when manufactured properly anyway. Maybe you have a foreign body (like piece of lint, hair or such) interfering. Remove the GFX card again and blow out the socket with canned air.

Then again, re-seating the card may have been enough, let us know how things work out for you!

 

[Npl]

its a western digital and it vibrates more than I would like and alot more than the ~10 WD drives I had before.
Though probably not more than the average Seagate or Samsung I encountered.

Anyway, issue still persists

Practically the only thing I would be interested in playing right now would be PA, so Im pretty close on trying to get my money back for the gfx-card and run on intels IGP. Really dont wanna piss around further for just one game.

Thanks for all your feedback.