Replacement Video Card - nv4disp

T

TurboDean

n00b
Joined
Dec 27, 2007
Messages
52
I would like some feedback on a video card recommendation, ATI, not Nvidia. This is not used for gaming, so trailblazing graphics isn't needed, but it is used for some personal photography/video editing, when I get the time.

Why not Nvidia? My computer freezes, giving the nv4_disp error, several time per week. Sometimes during use, but typically when "waking up". This is a relatively new system with internals selected based on recommendations from members here. This has happened since I got it running. Innards:

Corsair HX520W PSU
Gigabyte GA-P35-DS3R and Intel Q6600
4GB Patriot Extreme Performance DDR2 800 (2 x 2GB)
MSI NX8500GT-TD256EH GeForce 8500GT 256MB 128-bit video card
3 x Seagate Barracuda 7200.11 500GB HDs (2x RAID1)
LG GSA-H55NK DVD burner
Windows XP Home SP2 O/S

Latest drivers make no difference, same with reinstalling old drivers. Driver Cleaner Pro was used. I have also tried re-installing Direct X, reducing the Hardware Acceleration, and turned off "fast write". No diff. I have not tries this patch: http://forums.nvidia.com/index.php?showtopic=4432&st=0

There are recommendations to try swapping out the PSU, and RAM, and xyz. I don't have any laying around, so I would be buying this stuff, just to "try". Seems silly. So, poking around Newegg.com I see this:
HIS Hightech H260XTP512DDN-R Radeon HD 2600XT 512MB 128-bit GDDR3 PCI Express x16. I have an Antec 182 case, so the size of the heat sink is not a concern.

Look reasonable?


Thanks,

Dean
 
Its strange you consider this to be an NVidia problem.
I dont think its anything of the sort.
I'd offer some help getting it going but that isnt what you want.
Good luck :)
 
Oh heck, I'll take help! I would love to get it running correctly with intelligence rather than parts, but so far that has been fruitless. Other than "the patch" that I read about, what have I overlooked?

I don't believe it is solely a Nvidia problem, but there is an issue with Nvidia drivers with XP on *some* combinations. I'm not alone. A "nv4disp windows xp" search on Google gets enough hits to confirm this, and I appear to be one of the lucky ones. :)

Dean
 
I've really only see actual GPU driver errors (from EITHER team) when I'm overclocking. Let's just say that last time I overclocked was with the 9800Pro or the 6800NU (16/6 of course). Running either of them at stock, you should never see this error.

I would say probably the best solution would be to try another nVidia card. This would rule out hardware as the problem (and then confirm it as a software issue). E.g. if you can find an el-cheapo nVidia card, of any kind, and then obviously fresh install a stable release of drivers (I'm still on 178's).
 
i had this issue with n4 issue many many many times under vista, overclocked or not, do some google searches and you see SO many hits on this very same problem! yet few every seem to find any resolution to the problem
 
You said you've done a little google research so you've probably already done this but using driver cleaner pro has in the past fixed driver errors for me. Maybe you could try uninstalling the driver and then use DCP?
 
MrGuvernment said:
i had this issue with n4 issue many many many times under vista, overclocked or not, do some google searches and you see SO many hits on this very same problem! yet few every seem to find any resolution to the problem
Click to expand...

That's nice to know, but the OP is on XP SP2, not Vista.
 
I had the same problem when I built my new system. RMA'd my graphics card, used another old graphics card I had, replaced ram, PSU, CPU, etc., I also tried reinstalling OS, using a different OS, 4 different versions of drivers, etc. So what ended up being MY problem? The Motherboard. Hope you get lucky and change the right component the first time! gl!
 
Try taking out one of the RAM chips. You're running a 32-bit operating system with only 4096 mb of adressable memory space, 2048 of them are reserved for system. Only 2048mb are usually available for programs and games. In most cases, your system doesn't utilize the last 2 gb. Gives more problems then gains, even for graphical use. Taking out 2 increases chances of better system stability and I don't think you'll notice any performance penalty even while working with photo's.
 
An nv4disp failure is simply a graphics card crash, it doesnt point to anything specific being the problem.
Possible things that could affect it:

Display driver corrupt/not properly installed
Motherboard drivers corrupt/not properly installed (related to PCI-E operation)
Corrupt or badly configured CMOS
Out of date BIOS
Sharing interrupt conflict from device on PCI bus
Bad PSU
Bad power connection
Badly fitted gfx card
Gfx card core, memory or voltage regulators overheating
Faulty or badly configured/fitted memory
PCI-E speed incorrectly set
Faulty / overheating motherboard
Short circuit on pretty much any component
Unstable clock speeds
Overheating PSU

To start make sure the PC is clean of dust and there is no dust in the memory and PCI-E slots.
Clean the gfx card and system memory contacts.
Take the case side off and blow a housefan at the PC to see if it helps. If so you likely have a component overheating.
Try resetting the CMOS (make a note of any settings you have changed first).
Install the latest motherboard drivers / BIOS.
As mentioned above, try just one ramstick in different slots to see if a slot or memory module is faulty.
Up the memory voltage to 2.0 volts (up to 2.1V is safe) and see if things improve (I have some memory here that will not even boot at 1.8V).
Make sure that there are no extra metal motherboard pegs underneath the motherboard in the wrong place as this will cause a short.
If you can borrow a known good PSU that is powerful enough, you can rule out a PSU issue.

That will do for starters.
Try things one at a time.
 
Like others said, your videocard doesn't have to be faulty to get that error.
I used to get that error all the time and in the end traced it down to my ram. I significatly underclocked my cpu/ram and then ran benchmarks in Crysis. By upping all the details/AA/AF to the max, I was ensuring that the underclocked cpu/ram wasn't the bottleneck and the gpu was still working it's harddest. Even after a whole night the thing didn't die so I was pretty sure it wasn't the gpu.
Long story short, it seems my motherboard didn't like Corsair memory. It killed 3 seperate sets of them... Each time I would get nv4disp errors after about a month... So I ended up getting some OCZ and it runs fine now.
If you can, try installing a fresh OS and test for the errors. You don't have to wipe your current OS, just get another HDD and do it on there. Once you found the problem, pop the old one back in.
 
Sorry folks, I'm not ignoring the responses, but about an hour after posting this we got a call that a friend died and we have been occupied with that. I'll get back to this next week.

Dean
 
Must be the week for it, my Nan went today.
Good luck getting things in order.
 
Man, what a month. OK...


alg7_munif said:
http://www.newegg.com/Product/Product.aspx?Item=N82E16814131135
Click to expand...
Thanks, though I have read of issues running PCIe 2.0 card on the Gigabyte DS3L mobos, and am concerned of similar issues with my DS3R, so I would look into 1.x cards first.


Lender said:
...Hope you get lucky and change the right component the first time! gl!
Click to expand...
Pfft. Me too. I can't afford to literally build another computer by buying and swapping out parts.


Tamlin_WSGF said:
Try taking out one of the RAM chips.
Click to expand...
I can do that with dual channel? I thought I had to have two chips in.

Nenu said:
...Try things one at a time.
Click to expand...

First, sorry for your loss.


I shut it down and cleaned it all out, though it wasn't very dusty. I ran it with the case cover loose for a week so I could interrupt my use to feel the CPU and VC heat sinks, which were only warm, so I don't believe heat is an issue.

I am not entirely computer savvy, so I will need to read my manuals to see how to reset the CMOS and increase the memory voltage. Should I try adjusting the memory voltage before removing one RAM stick, or pull and move around 1 stick first?

The driver updates listed at http://www.gigabyte.com.tw/Support/Motherboard/Driver_Model.aspx?ProductID=2626 that would be relevant, considering my purchase date of Jan 2008, would be the following:

  • BIOS F13c 2008/11/17
  • Realtek Function driver for Realtek Azalia audio chip (Including Microsoft UAA Driver in English edition)
  • Realtek 8111/8169 LAN Driver
  • GIGABYTE SATA2 RAID Driver
  • GIGABYTE SATA2 RAID Driver (Preinstall driver)
  • Intel ICH9R/ICH10R SATA RAID Driver
  • Intel ICH9R/ICH10R Driver (Preinstall driver)

Except for the F13c, I don't see how they would matter to what is happening, but should I do 'em anyhow? (I believe I am running the Intel RAID Controllers, though I forget how to verify that. I'll look into it.)

No mobo pegs where they shouldn't be. I know no one off hand to even ask to borrow a PSU from, though I'll ask around.

hl3395 said:
If you can, try installing a fresh OS and test for the errors.
Click to expand...
The OS was a PITA to get up and running. It was loaded off an old Dell XP disk from '02, and it nearly took an act of God to get the "Key Code" which I was told was a one-time deal since I was ditching the Dell box. (My argument was that nothing in the Dell documentation stated I was only licensed for use.) Something about MS licensing not allowing it, though it was MS that finally gave me the code, and they had no problem doing it since I could prove I was the original owner. Anyhoot, then I had to get the SPs installed, which didn't go well. It took several tries to get the OS up.

If I must reload, I suppose I could go buy a new full version, but that would REALLY irritate me to have to do that. :)
 
I had that driver problem with my nvidia 8800 gtx and a abit mainboard. I never found out what was wrong, blamed the nvidia card and bought a ati card last summer. Sorry I can't be any help, but nothing I tried got it working. :/
 
i've experienced the nv4 bluescreen in the past when an overclock went bad, the card overheated, the power supply was inadequate for the system, or the card was a non reference design that the official drivers didn't support. looking over the initial post, i don't think the op is suffering from any of these problems, but i must ask what drivers are being used? lastly, its entirely possible that the video card is defective, and a replacement will solve the problem.
 
rennyf77 said:
...but i must ask what drivers are being used? lastly, its entirely possible that the video card is defective, and a replacement will solve the problem.
Click to expand...
Current 178.24, but now downloading 181.22, which apparently became available last Thursday. I too wonder if it is just a defective card, as I have read it happens, much like anything else. I'll go through the freebie checks first, though for $15 plus shipping, I could try this:http://www.newegg.com/Product/Product.aspx?Item=N82E16814127294. I could put this in and see if it continues to run.


alg7_munif said:
If he is using an ATi GPU, he won't get the nv4disp error. The error is only on nVidia cards.
Click to expand...
I more meant the memory dump / infinite loop issue. :)

Does that not occur with ATIs??
 
TurboDean said:
Does that not occur with ATIs??
Click to expand...

nv4disp.dll is the main nVidia driver file. You can find it somewhere in the Windows system folder. When you get this error code, it's difficult to interpret because it just means that something went wrong with the video card or its drivers. This could mean a bad video card; it could mean a bad driver install (especially if old driver code is left over from previous drivers); it could be a motherboard gone bad or with bad chipset drivers; it could even mean bad RAM corrupted the video card driver while in memory.

ATi cards wouldn't get that specific error because they don't use that file. You might get a different error or behavior if using an ATi card doesn't solve the problem.

So, you can do a few things that cost nothing to rule them out:

1. To check for bad RAM, run Memtest86+ 2.11. I always do this when experiencing problems on a PC, because bad RAM can cause all manner of odd behavior. It is important to run this test through more than one pass - preferably overnight for 10 hours. I have personally seen several computers fail Memtest86+ only on every third pass.

2. Make sure your motherboard's BIOS and chipset drivers (two different and very important things) are updated to the latest version; check the motherboard maker's website for them. Doing this fixed one of the computers I witnessed failing every third Memtest86+ pass.

3. Use Driver Sweeper and/or the manual driver removal instructions on the Tweakguides nVidia driver installation and removal page to ensure you have a clean install of the nVidia drivers.
 
Every time I go to wake up the computer, without touching the mouse, I turn on the monitor and wait a few seconds. Then I immediately begin the shutdown process as soon as the monitor begins shows that first log in screen, up to the point of that last click. Any oddities and I immediately do that last click to shut down in an attempt to avoid the RAID to have to rebuild, yet again, but sometimes the mouse locks and it is too late.

Today is the day to start working on this, but of course it had been 3 days since the last crash, so as soon as it wakes up I can tell it is time. The evidence was a flickering screen with some "stuck" pixels, a portion of a program (Thunderbird) still visible on the Windows user log in screen, and a jerky mouse pointer. I shut down in time.

I was going to start the day by loading all the updates, but figured as long as the computer was powered down I would go ahead and install the spare pair of RAM sticks I bought when I originally built this computer. I got them to maintain continuity in case I went to XP-64 or Vista. At the same time I was going to put them in slots 3 & 4, but followed Nenu's recommendation (and the protocol I typically to follow when tuning or troubleshooting cars) and decided to do one change at a time.

I booted up and noticed an interesting and welcomed effect: It booted much faster. I never timed it, but I can say that boot-up would take ~2 minutes. This time it was under 1 minute. I did it again, timing it, to make sure I wasn't bonkers. Now I suppose I should wait a few day to see what happens. Even if it doesn't crash, I'll still do all the updating, just because I should.

Now, I wait...
 
MocciBoo said:
nv4disp
errors
are
a
problem
with
nvidia
cards, not ati cards.
Click to expand...

Are you saying that ATI's drivers have some strange ability to not crash when faced with faulty RAM/Mobo/<insert component here>?

Yes, the nv4disp driver is specific to video cards, but the fact that the video driver is crashing can be related to any number of problems. One of them could be a faulty card, yes. But there's a long list of things that can cause such a crash. What these people are trying to do is rule out any other system issues before having the OP go and spend money on a solution that may not work.

What good would it be for him to buy a new card, only to have the ATI driver start crashing, too? There are steps to be taken to narrow down the list of possible causes, and continually spouting the incredibly obvious fact that "nv4disp is not an ATI issue" is NOT one of them.

To the OP, good luck on finding the problem. Hopefully you can, as you said, fix it through inteliigence rather than through new parts.
 
Un4given, thanks, I'll check that out.

MocciBoo, yes, I get that, as noted earlier.

Pyroja, I'm trying. Hopefully the ol' gray matter is up to it. :)


UPDATE JAN 29, ~9:00 a.m.: Well, this cycle between crashes lasted all of about 24 hours. During use, the courser disappeared, then the mouse froze, then the screen started going speckly (sp?), then BSOD, all in a matter of seconds. No prior indication. This time while it was off, I put the RAM in slots 3 & 4. Presuming there is a next time, I'll pull one... or should I do all the updates first, memtest, etc.?

UPDATE JAN 31, at ~11:00 a.m.: This cycle between crashes was about 48 hours. Wouldn't wake up. Performing updates today.
 
Nvidia driver is updated.

I downloaded but did not install the F13c BIOS update from .../Support/Motherboard/BIOS.... Why? The warnings scared me: .../Support/Motherboard/HowToReflashBIOS.... Considering the notes for the update ( 1. beta BIOS; 2. update CPU microcode; 3. Support RTL 8111C ), should this be my next step or wait, and do other, safer things?

I found the RAID1 controller information and it is the Intel ICH9R/ICH10R SATA RAID driver, v7.6.0.1011 dated 07/12/2007. The new driver is v8.5.0.1032 dated 09/15/2008 so I'll get that, but two are listed, an "Intel ICH9R/ICH10R SATA RAID Driver" and an "Intel ICH9R/ICH10R Driver (Preinstall driver)". Is the Preinstall Driver for a new installation and should I ignore that one?

I plan on running the Memtest tonight.

UPDATE Feb 1, ~1:00 p.m.: Crashed again when trying to "wake up". Last used last night at ~ 10p.
 
Cannot run Memtest. The machine crashes within seconds of the test starting. Repeatedly.
 
TurboDean said:
Cannot run Memtest. The machine crashes within seconds of the test starting. Repeatedly.
Click to expand...

In the motherboard BIOS, turn off USB Legacy Support if it is turned on. On some motherboards, if USB Legacy Support is turned on, Memtest will crash when you move the mouse or use the keyboard.
 
Thanks, evilsofa, I'll keep that in mind, but I didn't touch a thing. It started running, I thought, "cool", and the the computer shut down.
 
Crashing Memtest is probably a hardware or motherboard BIOS issue, because it's bypassing Windows and everything else on your hard drive (including the chipset drivers). At this point I'd update the motherboard BIOS to F13c. Don't worry too much about the warnings; they have to say that because while it's flashing, it is vulnerable to power blackouts and user stupidity, neither of which is likely to happen here. If it fails to flash, then the motherboard was dying anyway.
 
Any recommended backups before I do this, or would the HDs be fine? I do have a spare 500g HD in the box, and I presume I can make a rather complete copy of the entire RAID, couldn't I?

Am I worrying about this too much?
 
Man that sucks, sry to hear about your friend.

Does memtest run at all or you crash before it even starts running? I think the only thing that can crash memtest repeatedly is cpu/ram/mobo and since you've tried new ram that kind of points at the cpu/mobo. You can turn off some of the cores on the cpu to see if that helps. There should be an option for that somewhere in the bios. Something along the lines of disable multicore or some mobo you can disable specific cores.

Also you don't need your harddrives connected to run memtest or flash the bios. You can run them from just a usb stick (or cd/floppy). You can disconnect the hdds if you don't feel comfortable. After you flash your bios, make sure to "Load Optimized Defaults" to set everything to default.

gl
 
Thanks.

Memtest runs for ~5 seconds, then the computer begins to reboot.

I'm looking at procedures to update the BIOS and will give it a shot. I see a couple of update options, Q-Flash & @BIOS.
 
Just for the heck of it I tried Memtest one more time, with what I'll call Module #2 in Slot 1. Failed test 6 with 5 errors, but it also locked up the test with a counter for "errors per mem" or something like counting counting up, up, up. I manually rebooted when it hit 80 million.

Then I put Module #1 in Slot 1. Failed test 6 again, but this time with 47 errors, however it did complete the test and started to run again.

Module #3 in Slot 1 PASSED! It is late, though, so I stopped the test for now. Tomorrow when I get the opportunity I will continue, starting with Module #4 in Slot 1. If/when it fails, Module #3 goes back in and I'll test it at each slot, but if it passes the first test I'll let it continue to run a few times.

I'll be ticked if 3/4 of this Patriot memory is cr@p. Sorta pricey when I bought it @ $115/pr. At least RAM is cheaper now, but still...
 
You should still update the motherboard BIOS, and retest the RAM that seems to be bad. It could be that the older BIOS versions were incompatible with the Patriot memory, and that the latest version of the BIOS could fix that. RAM incompatibilities is one of the main reasons for BIOS updates.
 
When I started getting this on my 8800GTS, it turned out that the video card wasn't getting enough power. Apparently the plug that went to the video card would short out occasionally, so bam, blue-screen, and that's it. So, check your cords and the power supply. Make sure your power supply will supply the correct amount of amps.

Another option you have would be to get it replaced through MSI.. I'm not sure how good their warranty is, I personally can't stand them after they had the nerve to tell me that on a previous build it was everything but their video card. However, when I bought the exact same model through Asus, it worked just fine. Since then, no MSI for me.
 
You must log in or register to reply here.
Back
Top