25 replies to this topic

[p8ragon]

Hi everyone, I recently purchased a new video card (gtx 780 ti), and it's causing MWO to crash to the desktop at random intervals. The thing is- when I do crash, I don't get any error messages, etc...the game seamlessly transitions from mech to microsoft, and it's infuriating, to say the least. Originally, I thought that the problem lay with a faulty graphics card, so I got an RMA. Nope. Just plugged in the 'new' graphics card yesterday, and I have had the same problem. I'm quite certain that the card's overclock has nothing to do with it, since I have seen 0 artifacts, and temps remain @ a constant ~80C

My specs are:

i5 3570k @ 4.79ghz
MSI Z77a-GD65 Gaming
850 watt PSU
gtx 780 ti
1tb caviar black
2tb seagate
128gb vertex IV
8gb of kingston RAM (dual channel-4gb each stick)
windows 8.1 64bit

Any help would be welcome,

Thanks! (btw I game @ 1080p)

PS: when I purchased the video card, i formatted my pc, and reinstalled windows. My previous video card was a 7970 dc2 (all other specs identical) and I had no problems with that.

[NIRVAN]

Reduce the frequency of CPU.

[Goose]

Anything informative in the games' log files?

[Loc Nar]

[p8ragon]

Goose, on 31 May 2014 - 11:37 AM, said:

Anything informative in the games' log files?

I'm not sure...is it possible for me to upload them here?

[Catamount]

Are you certain the new GPU is causing MWO to CTD?

Could game files simply be corrupt? OTOH, if it's GPU-related, could it be a bad driver install or version?

[p8ragon]

Catamount, on 31 May 2014 - 01:35 PM, said:

Are you certain the new GPU is causing MWO to CTD?

Could game files simply be corrupt? OTOH, if it's GPU-related, could it be a bad driver install or version?

I'm not sure...I downloaded the repair tool for MWO a while ago, but couldn't get that to run. I have the latest nvidia drivers, and when I installed them, I checked the "clean installation" option.

MWO isn't the only game that crashes on me (tomb raider, bf4, and far cry also crash...however those give me "*game name*.exe has stopped working" error messages when they crash. MWO on the other hand, just transitions.

[Dragonsyoung]

sounds like a GFX driver issue. I have had this issue. try older verions.

[Catamount]

There are two likely possibilities:

1.) Your graphics driver is corrupted, or the new version just doesn't work well

2.) Your PSU can't handle the new graphics card, either because it's a low quality unit, or a defective unit that your previous GPU didn't stress enough to reveal problems with

You only say that your PSU is an "850W" unit, but units that claim to be 850W span an entire spectrum, from this Enermax, which would probably just about last a lifetime, to this Diablotek, which despite being called an "850W unit", would probably explode if you pulled 400W from it.

It would help to know exactly what kind of PSU you have.

In the meantime, if you want to test the PSU hypothesis at least partly, just run Furmark and see how quickly and severely it crashes. Also, run GPUZ and see what it says about the 12v rail when you load it up, with anything. Software voltage solutions aren't always perfectly accurate, but it'll be good enough to see if you get massive variance between idle and load.

[p8ragon]

Dragonsyoung, on 31 May 2014 - 02:12 PM, said:

sounds like a GFX driver issue. I have had this issue. try older verions.

Is there a particular version you suggest?

Catamount, on 31 May 2014 - 03:58 PM, said:

There are two likely possibilities:

1.) Your graphics driver is corrupted, or the new version just doesn't work well

2.) Your PSU can't handle the new graphics card, either because it's a low quality unit, or a defective unit that your previous GPU didn't stress enough to reveal problems with

You only say that your PSU is an "850W" unit, but units that claim to be 850W span an entire spectrum, from this Enermax, which would probably just about last a lifetime, to this Diablotek, which despite being called an "850W unit", would probably explode if you pulled 400W from it.

It would help to know exactly what kind of PSU you have.

In the meantime, if you want to test the PSU hypothesis at least partly, just run Furmark and see how quickly and severely it crashes. Also, run GPUZ and see what it says about the 12v rail when you load it up, with anything. Software voltage solutions aren't always perfectly accurate, but it'll be good enough to see if you get massive variance between idle and load.

I am using a cougar 850 watt (http://www.newegg.ca...N82E16817553006). It's 3 years old. I will load up furmark and report back on those results. Am i looking for anything in particular on the 12v rail?

Thanks!

Update: I ran furmark for 15 minutes...and there were no problems. However, I crashed after running tomb raider for 5 minutes.

Update 2: Weird bug I noticed....after crashing with an error message (IE Tomb Raider.exe has stopped working)...no other games will boot up. (They all crash upon startup with x.exe has stopped working). I downgraded my drivers to March's release. Unfortunately, this does not seem to have alleviated anything.

Edited by p8ragon, 31 May 2014 - 06:56 PM.

[Frosty Bits]

Maybe check too see if there is motherboard bios upgrade for your new graphic card to work? Just a thought

[Redshift2k5]

I had tons of trouble with my new GPU(Radeon HD 7770) until I underclocked it by a moderate margin. Running it at it's factory defaults caused a video driver crash several times per hour.

Look for some windows driver error logs, or at least the MWO crash dump (Omnicron) file for clues.

Edited by Redshift2k5, 31 May 2014 - 07:45 PM.

[Catamount]

Underclocking should never be necessary unless you have a borderline PSU or insufficient cooling (on the card or in the case).

P8ragon, it's a long shot but you say other games won't start. What if DirectX is corrupted? Or some part of the OS?
What you're describing is NOT indicative of a power supply issue (that model is good to go anyways). It could still be drivers. Whatever it is, it sounds like software. Is just nuking the computer and reinstalling Windows an unreasonable inconvenience? I know it's easier for some people to do than others (I could do it tonight after 5 minutes of backing up and be up and running completely again within like an hour, but I also have bad luck, so I've learned to be that way)

Edited by Catamount, 31 May 2014 - 08:30 PM.

[p8ragon]

Catamount, on 31 May 2014 - 08:29 PM, said:

Underclocking should never be necessary unless you have a borderline PSU or insufficient cooling (on the card or in the case).

P8ragon, it's a long shot but you say other games won't start. What if DirectX is corrupted? Or some part of the OS?
What you're describing is NOT indicative of a power supply issue (that model is good to go anyways). It could still be drivers. Whatever it is, it sounds like software. Is just nuking the computer and reinstalling Windows an unreasonable inconvenience? I know it's easier for some people to do than others (I could do it tonight after 5 minutes of backing up and be up and running completely again within like an hour, but I also have bad luck, so I've learned to be that way)

Just reinstalled windows. Hope things turn out well. Will report back tomorrow. (it's nearly 2am my time)

[Golrar]

Have you tried downloading the Directx 9 complete pack from Microsoft and see if that fixes the issue? It won't overide any DX10 or 11 files. However, some people report issues with DX9c on Windows 8.1

http://social.techne...8itproappcompat

if you can get it to install, try using DX9 in MWO and see if the game works.

Edited by Golrar, 01 June 2014 - 03:13 AM.

[Barbaric Soul]

The GTX 780ti requires 42 amps, the OP's PSU has the following voltage specs-

Quote

+3.3V@25A, +5V@25A, +12V1@22A, +12V2@22A, +12V3@24A, +12V4@24A, -12V@0.6A, +5VSB@3.0A

I'm wondering if one of the rails may be starting to get a little week or of the OP does not have the 780 connected to more than one rail. Taken from one of the Newegg reviews (this is why I ONLY buy single rail PSUs)-

Quote

This PS has two 22A 12V rails and two 24A 12V rails. There is already two integrated PCI-E 8(6+2) pin connectors connected to one of the 22A 12V rails. Two modular cables plug into one of the 24A 12V rails.

If you're still having problem after reformatting, try using only one of the permanently connected 6+2 connectors and one of the modular 6+2 connectors to power the 780ti.

I know you ran Furmark without a issue, but reading a review of your PSU, saw this-

Quote

DYNAMIC FOUR-12V RAIL
Automatic dynamic load distribution on Four-12V rail provide power separately to the GPU and the CPU. If you are not using all output lines, the PSU automatically reroutes needed power from unused lines. This improves the performance of the 12V lines considerably for systems with high-end graphic cards in SLI® or CrossFire® mode and provides high voltage stability to keep components safe.

So with that, since Furmark loads your GPU about as hard as it can be loaded, I'm thinking this is the reason why Furmark is running fine, but games are not. Furmark is putting enough load for that feature of your PSU to kick-in, but the games might not be quite enough for that feature to kick-in, which is resulting in your 780 being starved for electricity.

DISCLAIMER- I AM NOT A PSU EXPERT, AND THIS IS ALL JUST SPECULATION, BUT IT IS WHERE I WOULD START IF THIS WAS MY SYSTEM AND THE REFORMAT DOES NOT HELP.

Edited by Barbaric Soul, 01 June 2014 - 05:10 AM.

[Catamount]

Huh, that's definitely a possibility.

I'm not sure if that would explain subsequent refusal of programs to boot, but drawing from two rails would most definitely be advisable one way or the other, and yes, while multirail PSUs have their minor advantages, I say single-rail all the way. Of course, I don't bother with any brand but Rosewill anymore anyways unless there's a REALLY good sale on something else, and all their newer lines are basically single rail, so that about settles that anyways (even if they do almost certainly do it for cost reasons)

Edited by Catamount, 01 June 2014 - 06:04 AM.

[9erRed]

Greetings all,

- Ensure that your running something that allows your to 'ramp up' the video card fans speed.
(originally Nvidia used 'system tools' for this but it's discontinued. Now they recommend the manufactures specific tools, Asus GPU Tweak, EVGA Precision, MSI Afterburner, Zotac FireStorm. Whatever you use set up a profile for MWO that ramps the fan speed to the card temperature.)
- The latest certified drivers for your card are here: (337.88 WHQL, May26,2014)
http://www.geforce.c...s/results/75991

- Additionally make sure that the 2 plugs for the Video card power are directly tapped from the PSU, and not split between other peripherals. Your PSU has 4 red plugs at 8 pins for this, make sure that the vid cards 8 pin plug and 6pin plugs are actually plugged in.

- And as others have stated reload the Dx9c drivers, and do a windows update as your system has now changed.

Good luck,
9erRed

[p8ragon]

Update: reformatted PC. Same problem. Will try out Barbaric Soul's suggestion and report back.

Edit1: Tried using one modular and one permanent...didn't change anything

I'm half tempted to go out and buy another 850 watt right now.

Edited by p8ragon, 01 June 2014 - 11:05 AM.

[Goose]

http://www.memtest.org/

http://img571.images...elshortcut1.png