Unreadable wall-of-text is unreadable.
If you edit and make the paragraphs shorter it will really help people help you.
If you edit and make the paragraphs shorter it will really help people help you.
GTX 470 Crash Problem
Alright. I'm pretty highly experienced with computer hardware and software, but this problem has me somewhat stumped. I've got an EVGA GTX 470 running in a machine with a Gateway OEM G31 motherboard and Core 2 Quad Q6600. The mobo and CPU were a pull from a Gateway FX7024 with a bad video card (9800GTX+) and I've been using them for quite a bit of time now. In Windows 7, just about any application or game that uses the video card seems to have a chance to cause the process to hang, or sometimes crash the entire system. On rare occasions I get a message stating the the NVIDIA display driver crashed (I have 266.58, the latest according to nVidia System Update). Now newer and more graphics-intensive games certainly make the problem occur more often. I'm at the point were Bad Company 2 will run for less than 10 minutes before the entire system locks up and I have to reset. However, even my 2D Baldur's Gate 2 game will cause the crash. In the less severe instances, the screen just goes black and the sound stutters. When this happens, I can usually Ctrl-Alt-Delete and kill the game. This is only in Windows. I dual-boot Mac OS X with a custom bootloader and the Mac Fermi drivers released by nVidia and I have absolutely zero problems with games there. I've tried things like the Mac port of Call of Duty 4 and Starcraft 2 and they both run fine, while their Windows counterparts kill the system. I've tried reinstalling Windows, installing nothing but the nVidia drivers and a test game. I ran memtest86+, once on the whole 8 gigs of RAM I have (DDR2 800 MHz) and once on each 2 gig stick separately. No errors reported. I tried reseating the card in the PCI slot. I have a 630W PSU. It may seem a little low, but according to nVidia specs the minimum is 550W.
The fact that more graphically intense games make it worse made me think it was a cooling issue first. I ran a GPU temp monitor and saw that even with default settings the temp never exceeded 85ºC. I tried ramping up the fan manually and kept the temp under 70ºC but it still crashed. While I still had Aero (or whatever Win7 uses) on, it crashed once or twice just sitting at the desktop. I tried cooling the entire case more efficiently in case it was a motherboard component overheating but that did not help.
Right now I'm at wit's end with this issue. Immediately before getting this card I had purchased a Gigabyte 465. It exhibited what seemed to be the exact same issue, so I RMAed it. Before receiving the replacement, I noticed the EVGA GTX 470 on sale for less than I paid for the 465. I bought the 470 and sold my replacement 465. The fact that the issue happens across two OSes and two different cards suggests I have a driver problem, but I haven't found anyone else on the net with the same issue. It could also be a motherboard or power supply problem, but I can't afford a replacement for either unless I can determine that's the case. I do have an identical model motherboard from a different system but that's four hours away. I don't have access to any other components for testing. Does anyone have any diagnostic suggestions? Perhaps a way to test if the PSU is being overdrawn (it's not overheating) or some logs somewhere I can dig up? Like I said, I'm at my wit's end. After the new graphics card, I had hoped to keep the system as-is for at least another year and a half.
The fact that more graphically intense games make it worse made me think it was a cooling issue first. I ran a GPU temp monitor and saw that even with default settings the temp never exceeded 85ºC. I tried ramping up the fan manually and kept the temp under 70ºC but it still crashed. While I still had Aero (or whatever Win7 uses) on, it crashed once or twice just sitting at the desktop. I tried cooling the entire case more efficiently in case it was a motherboard component overheating but that did not help.
Right now I'm at wit's end with this issue. Immediately before getting this card I had purchased a Gigabyte 465. It exhibited what seemed to be the exact same issue, so I RMAed it. Before receiving the replacement, I noticed the EVGA GTX 470 on sale for less than I paid for the 465. I bought the 470 and sold my replacement 465. The fact that the issue happens across two OSes and two different cards suggests I have a driver problem, but I haven't found anyone else on the net with the same issue. It could also be a motherboard or power supply problem, but I can't afford a replacement for either unless I can determine that's the case. I do have an identical model motherboard from a different system but that's four hours away. I don't have access to any other components for testing. Does anyone have any diagnostic suggestions? Perhaps a way to test if the PSU is being overdrawn (it's not overheating) or some logs somewhere I can dig up? Like I said, I'm at my wit's end. After the new graphics card, I had hoped to keep the system as-is for at least another year and a half.
1 - 11 of 11 Posts
Bullet points:
Graphics hang in games/ gfx-intense programs
Hang is sometimes followed by driver crash
Happens in win7, but not on OSX
Happens on two different cards
Has crashed with only windows Aero running.
630 watt PSU.
I believe it's one of two things, your "630" watt PSU is lower quality and not actually delivering that. Driver problems are the other possible issue, I'm not overly versed with OSX, but is it possible that the OSX drivers could have had any effect on their windows counterparts? Try uninstalling the OSX drivers and rebooting into Windows, do a clean install of drivers, and see what happens.
My gut tells me the PSU isn't up to scratch.
Graphics hang in games/ gfx-intense programs
Hang is sometimes followed by driver crash
Happens in win7, but not on OSX
Happens on two different cards
Has crashed with only windows Aero running.
630 watt PSU.
I believe it's one of two things, your "630" watt PSU is lower quality and not actually delivering that. Driver problems are the other possible issue, I'm not overly versed with OSX, but is it possible that the OSX drivers could have had any effect on their windows counterparts? Try uninstalling the OSX drivers and rebooting into Windows, do a clean install of drivers, and see what happens.
My gut tells me the PSU isn't up to scratch.
Apologies for the wall-of-text-ness. I was trying to provide as much information as possible.
Yeah I've been thinking about the PSU but I don't understand why if it was an overdraw problem, it would only occur in Windows. As far as I can tell the card performs the same in both OSes and there have been totally zero graphics problems in Mac OS X. I'll see if I can put a meter on the 12V rail and look for voltage drops under load. Theoretically if I downclock the card it's draw would decrease as well. Lower clocks means less resistance and thus less draw.
The OS X drivers don't have any effect on the card, they're not adjusting settings or anything.
Yeah I've been thinking about the PSU but I don't understand why if it was an overdraw problem, it would only occur in Windows. As far as I can tell the card performs the same in both OSes and there have been totally zero graphics problems in Mac OS X. I'll see if I can put a meter on the 12V rail and look for voltage drops under load. Theoretically if I downclock the card it's draw would decrease as well. Lower clocks means less resistance and thus less draw.
The OS X drivers don't have any effect on the card, they're not adjusting settings or anything.
What PSU are we talking about here?
You may not actually have to break out the voltmeter ... first thing I'd try is monitoring 12V stability with Speedfan ... it usually will show your 12V rail and even has handy graphing and logging utilities (the logging is sometimes extra useful if you're having crashing issues since it can do real-time logging of voltages to a text file that will persist after a crash).
Some other seemingly random questions: Do you have a different physical HD you could install Windows on? Any chance you forgot to install the standoffs between the board and the motherboard tray? Are you using any adapters for the powering of the GPU? Has your machine ever just straight-up shut off when you fired up a 3d app? What exactly do you mean by 'crash the system'?
You may not actually have to break out the voltmeter ... first thing I'd try is monitoring 12V stability with Speedfan ... it usually will show your 12V rail and even has handy graphing and logging utilities (the logging is sometimes extra useful if you're having crashing issues since it can do real-time logging of voltages to a text file that will persist after a crash).
Some other seemingly random questions: Do you have a different physical HD you could install Windows on? Any chance you forgot to install the standoffs between the board and the motherboard tray? Are you using any adapters for the powering of the GPU? Has your machine ever just straight-up shut off when you fired up a 3d app? What exactly do you mean by 'crash the system'?
It's not the PSU. What motherboard do you have? Specifically what chipset does your NIC and integrated sound use? It should be Realtek or something like that.
If you have Realtek audio, try disabling it through the BIOS(mine shows up as "Azalia audio codec" in the BIOS) and then wipe the drivers from your install. If the game runs fine then it was your audio drivers.
I had to buy an Asus Xonar because the Realtek drivers were conflicting with the new 2xx drivers for the 470. I had problems for months until I put that new card in... never had a problem since.
When I had PSU problems the whole computer would just cut out(case lights, everything).
If you have Realtek audio, try disabling it through the BIOS(mine shows up as "Azalia audio codec" in the BIOS) and then wipe the drivers from your install. If the game runs fine then it was your audio drivers.
I had to buy an Asus Xonar because the Realtek drivers were conflicting with the new 2xx drivers for the 470. I had problems for months until I put that new card in... never had a problem since.
When I had PSU problems the whole computer would just cut out(case lights, everything).
Well for the record it's a 630W modular RAIDMAX PSU. I did break out the meter before I saw your post. At POST the 12v rail going directly into the card (no adapter, dedicated rail) read 11.79v. At windows desktop it held steady at 11.71v. The lowest it reached in Bad Company 2 was 11.40v, though when it crashed it was at around 11.50v. Those drops aren't bad at all, and seemed to coincide with revving the fan.
According to the Gateway specs on this motherboard (Schroeder Town G33), the audio chipset is "Sigmatel / IDT-STAC9271". I didn't install the OEM drivers as I presently use an external USB sound device (which Windows 7 seems to support natively). The new sound device is relatively recent however, it crashed just as much before I had it. Nevertheless I'll first try disabling the integrated audio in the BIOS. I do still use it for my mic, but that can be rearranged. If that doesn't help I'll try disconnecting my sound unit and disabling the system drivers.
When I reinstalled Windows a couple weeks ago, I put it on a different (larger) drive than the one I had been using previously. Both Western Digital. Definitely didn't forget the standoffs. No shorting as far as I can tell. No power adapters on the GPU, there's a video card power line out of the supply with two 6+2 connectors on it. The card uses both, although just the 6 pins on each. The machine has never suddenly shut down, while using 3D or otherwise.
I use the ambiguous word "crash" simply because I don't know how else to describe it. In less severe instances (older game/simpler graphics), the screen freezes and sometimes goes black a second later. Back when I used the integrated audio, the sound would stutter and loop the last moments of audio. With the USB sound, it just stops. I can CTRL-ALT-DELETE to bring up the management screen and then open task manager. The mouse is usually nowhere to be found so I use the keyboard to select the game process and kill it. The system then seemingly returns to normal. I can immediately relaunch and resume the game. (There is one very odd exception to this - any Call of Duty game (4, MW2, BO) will sometimes go back to working if instead of opening the task manager I just hit escape the close the management pane.)
In more severe cases (I've observed this mostly with Bad Company 2, Far Cry 2, Black Ops, and other really "pretty" games), the same symptoms occur but pressing CTRL-ALT-DELETE doesn't seem to do anything. I can't do anything to end the task and I must hit the reset button on the machine. Sometimes the screen has even gone black. Very rarely, the full system crash has happened with less intense games and other applications. I've had my 3D CAD (engineering modeling) software completely lock up the system.
Thanks for the prompt replies, by the way.
According to the Gateway specs on this motherboard (Schroeder Town G33), the audio chipset is "Sigmatel / IDT-STAC9271". I didn't install the OEM drivers as I presently use an external USB sound device (which Windows 7 seems to support natively). The new sound device is relatively recent however, it crashed just as much before I had it. Nevertheless I'll first try disabling the integrated audio in the BIOS. I do still use it for my mic, but that can be rearranged. If that doesn't help I'll try disconnecting my sound unit and disabling the system drivers.
When I reinstalled Windows a couple weeks ago, I put it on a different (larger) drive than the one I had been using previously. Both Western Digital. Definitely didn't forget the standoffs. No shorting as far as I can tell. No power adapters on the GPU, there's a video card power line out of the supply with two 6+2 connectors on it. The card uses both, although just the 6 pins on each. The machine has never suddenly shut down, while using 3D or otherwise.
I use the ambiguous word "crash" simply because I don't know how else to describe it. In less severe instances (older game/simpler graphics), the screen freezes and sometimes goes black a second later. Back when I used the integrated audio, the sound would stutter and loop the last moments of audio. With the USB sound, it just stops. I can CTRL-ALT-DELETE to bring up the management screen and then open task manager. The mouse is usually nowhere to be found so I use the keyboard to select the game process and kill it. The system then seemingly returns to normal. I can immediately relaunch and resume the game. (There is one very odd exception to this - any Call of Duty game (4, MW2, BO) will sometimes go back to working if instead of opening the task manager I just hit escape the close the management pane.)
In more severe cases (I've observed this mostly with Bad Company 2, Far Cry 2, Black Ops, and other really "pretty" games), the same symptoms occur but pressing CTRL-ALT-DELETE doesn't seem to do anything. I can't do anything to end the task and I must hit the reset button on the machine. Sometimes the screen has even gone black. Very rarely, the full system crash has happened with less intense games and other applications. I've had my 3D CAD (engineering modeling) software completely lock up the system.
Thanks for the prompt replies, by the way.
Yeah then it is definitely something to do with the audio drivers. With my integrated sound it will loop as you said. Mine was more severe and I needed to press the reset button because the whole computer locked up. When I did that it would make a loud "CHCKKK" sound out of the speakers.
My hunch is its the PSU not giving the card enough juice.
Just because it says 630w doesn't mean that's what its getting.
The amperage is the most important part though & I don't think
that lower quality PSU has enough for the GTX 470.
![]()
From the RAIDMAX 630w PSU specs I've looked at (the best one) they only have 22A,
so you are 16A under...
The other 630w model only had 16A...
GTX 470 Power Requirements:
A minimum of a 550 Watt power supply.
(Minimum recommended power supply with +12 Volt current rating of 38 Amps)
My 800w PSU has 98A on a single +12v rail.
Just because it says 630w doesn't mean that's what its getting.
The amperage is the most important part though & I don't think
that lower quality PSU has enough for the GTX 470.
From the RAIDMAX 630w PSU specs I've looked at (the best one) they only have 22A,
so you are 16A under...
The other 630w model only had 16A...
GTX 470 Power Requirements:
A minimum of a 550 Watt power supply.
(Minimum recommended power supply with +12 Volt current rating of 38 Amps)
My 800w PSU has 98A on a single +12v rail.
I'm having a somewhat similar issue with my 460, games will sometimes "crash" and become unresponsive. Normally if I alt tab and go back in things return to normal, except my framerates are cut in half, and that's because when I open MSI Afterburner I see my clocks are at 405mhz and not my overclocked 850mhz. That only happens because I'm still trying to find the right voltage for this OC.
Are you OCing at all OP? The card may be reverting to 2D clocks because it detects instability or something.
Are you OCing at all OP? The card may be reverting to 2D clocks because it detects instability or something.
The PSU has two 22A rails at 12V. If the card alone is drawing 38A, that means it's pulling 456W (12V * 38A = 456W). Now the maximum that a card with two 6-pin PCIe power connectors can draw is 225W. 75W comes through the motherboard slot and the 6-pin molex connectors are each rated for 75W. (Offhand knowledge verified here) Furthermore, the raw output of a power supply is best measured in Watts. The two +12V rails in my unit each theoretically provide 264W. I don't have the right meter to verify what the rail is actually putting out, that's not easy to do from an electrical standpoint. Anyway, when a load (gfx card, cpu, whatever) starts to pull more Amps than the supply can provide, the voltage drops suddenly. Remember that Volts * Amps = Watts. The supply only has a certain number of Watts to give so to be able to increase Amps beyond it's rating, it has to decrease Volts. I put a voltmeter on the +12V connectors going right into the card and watched the voltage while the card was operating. Even under extreme load the voltage never dropped below 11.4V, which shouldn't come close to being problematic. Are you sure that's 98A? 98A * 12V = 1176W, which means that one rail exceeds the entire rated output of the supply...
I'm not overclocking, though while running a GPU monitor I have noticed that the clocks immediately switch back to the 2D when a non-fatal crash occurs. I figured that was because the 3D pipeline had terminated and it reverted to just rendering the desktop.
EDIT:
Also, isn't 850 MHz a bit of a stretch from a baseline 405 MHz? I'm not an overclocking expert (I'm sure you'll get better replies), but that might be why you've got an unstable setup.
I'm not overclocking, though while running a GPU monitor I have noticed that the clocks immediately switch back to the 2D when a non-fatal crash occurs. I figured that was because the 3D pipeline had terminated and it reverted to just rendering the desktop.
EDIT:
Also, isn't 850 MHz a bit of a stretch from a baseline 405 MHz? I'm not an overclocking expert (I'm sure you'll get better replies), but that might be why you've got an unstable setup.
1 - 11 of 11 Posts
- This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Overclock.net
A forum community dedicated to overclocking enthusiasts and testing the limits of computing. Come join the discussion about computing, builds, collections, displays, models, styles, scales, specifications, reviews, accessories, classifieds, and more!
Full Forum Listing
Explore Our Forums
Top Contributors this Month
View All
HappySheep
351 Replies
Ichirou
199 Replies
GRABibus
194 Replies
Recommended Communities
