9800GX2 nvlddmkm Whose responsible?

[owcraftsman]

Although I got the replacement GPU from EVGA RMA a while back it's only now that I've had a chance to get her back in to a system for testing.

The good news is both cores show x16 which one core in the previous card showed x8. See www.evga.com/forums/tm.asp?m=647374 thread.

The bad news is this card will not run 3dmVantage or FarCry2 benchmarks when it is set to SLI enabled although will if disabled. What good is that? :frown:

http://valid.canardpc.com/show_oc.php?id=553520
http://www.techpowerup.com/gpuz/9fenb/
http://service.futuremark.com/compare?3dmv=1007163

I have read the many threads associated with this error all over the net and have tried the following to resolve it.

First the errors:

"Display driver nvlddmkm stopped responding and has successfully recovered."

"std:execption
IDXGISwapChain:resent failed: Device removed.
(DXGI_ERROR_DEVICE_REMOVED)"

http://img259.imageshack.us/im.../160/vantageerror4.jpg

Sadly many other have received the same errors.

What I tried to correct the trouble.

I adjust mem voltages manually
I have run memory diagnostics zero errors.
Tried two other pairs of memory 2x2 and 2x1 GB configs
I'm using FM approved gpu drivers.
I installed and reinstalled gpu drivers.
I'm using VM hotfix 1.01. for 3DMV
I've reinstalled 3DMV and hot fix.
I've installed reinstalled Adobe Flash and JavaRE.
UAC is disabled on my machine.
Tried disabling Aero using Vista Basic
I have run as administrator.
I've run dxdiag 32 & 64 no problems found.
I run game-o-meter system check OK ready to run.
I have all available windows updates.
Verified KB960715 is installed
I have the latest chipset drivers 15.25.
I have stopped all but essential programs one at a time from running in notification area.
I've no start programs running other than windows essential.
I've disabled anti V and firewall.
I use on board audio again latest HD driver version 6.0.1.5821 driver date 03/30/09
Windows & programs on Raid-0 C:\ Data on F:\
I tried older drivers that I knew worked with this card.
I used default clocks and stable OC clocks
GX2 fan at 100% temps ranged between 49-60c idle/load
All other system temp well within reason.
Tried all available 6x2 pwr on Corsair TX850.

Note: All the above are fixes that helped some overcome this issue and failed for others. So it's just possible if you are getting these errors one of them may help you

What I know/have eliminated at this point.

PSU is fine
Motherboard/chipset drivers/bios are fine.
OS,installed software and drivers are fine.
All memory module pairs and dimm slots are fine.

How I know this.

The system as currently configured runs both apps fine. Problem is I had to go the the heap and dig out a pair of 8800GTS 640's to SLI and test. I retested all memory and drivers (GPU Chipset included) with this config and ran 3DMV and FarCry2 benchmarks no problems. This effectively eliminates all the above and points to another failed GX2.

http://valid.canardpc.com/show_oc.php?id=556221
http://service.futuremark.com/compare?3dmv=1014612
http://service.futuremark.com/compare?3dmv=1014773
http://service.futuremark.com/compare?3dmv=1014914



Okay, you?re thinking. Why bore me with the details? just go ahead and RMA the product....
For me it's not that simple. First off I paid a premium price for a high end product that has never really worked right from the get go. I don't believe it's fair that I should have to pull more money out of my pocket again to send it back. When you start adding up the total cost not to mention the aggravation it would be easy to swear off Nvidia/SLI all together.

Plus it's obvious to me some folks aren't having trouble or the rash of complaints would be much higher which indicates the answer is out there. It would be nice to get to the bottom of this issue because even though the magnitude of complaints don't support a recall there is still a large volume of peeps with the problem as evidenced by the number of threads here and with Google search.

So what really is the problem here?

Note: It is not my goal here with this thread to trash anybody or entertain rants. I want facts and resolutions.

Is it simply a defective GPU?
Is it Vista no playing well with Nvidia driver?
Is it Nvidia drivers not playing well with Vista?
Is it Nvidia drivers not playing well with GX2 or for that matter 295 GPU?
Is it a GPU design flaw?
Is it DX10?
Is it Nvidia chipset?

One thing is clear it's simply not fair to pay a premium price and not get premium hassle free performance on your investment. We deserve better than this! Someone needs to step up to the plate and accept responsibility and then make it right.</

 

[taltamir]

nvlddmkm = nvidia driver. thats it, its just the driver's kernel.

 

[owcraftsman]

Yes I'm well aware of that taltamir.
I don't think the error itself is the problem.

 

[Mem]

Originally posted by: taltamir
nvlddmkm = nvidia driver. thats it, its just the driver's kernel.

I have that happen on factory overclocked video card(simple underclock fixed it in my case),so is anything overclocked at all?

 

[masteryoda34]

Have you tried reinstalling windows?

 

[vhx]

That error is basically like XP's nv4_disp error. In other words it can be anything from system memory to drivers to the graphics card.

 

[owcraftsman]

Originally posted by: masteryoda34
Have you tried reinstalling windows?

Yes I did a format and reload and the initial setup was on a clean system also.

 

[mhouck]

I've never had that problem with my card. I don't think you should have to pay for another RMA. The customer support people I've talked to have always been very understanding and have never given me a problem if I was having any conflict, especially when I went through and told them about how I went through and eliminated other variables.

The only problems I've ever had with the GX2 that was program specific was with GRID and Fallout3. I don't remember the exact name of the error because it was last year but the conflict had to do with a conflict from an overactive antivirus that deleted the specific dll files that were used by those programs.

That's the only advice I can provide. It looks like you've tried everything. I would bitch and complain until you get them to upgrade you to a 285 or reinstate your step up for whatever card you want.

 

[owcraftsman]

Regarding Overclocking the GPU at this point all is stock.
Underclocking is not an option IMO even if it works.

@mhouck your comments are welcome and underscores my Q Whose responsible? Yours works fine as I suspect do many others. My bad luck aside this tells me what I concluded in the all fine list above and points to a hardware issue. Still when you google this problem it's obvious there is a much bigger story here.

I have started this same thread at couple other forums also and collecting ammunition so to speak. I have every intentions of approaching tech support and at this point the last thing I want is another GX2 in return. I had not thought of it but your idea about reinstating my upgrade path is a good one but this card was an upgrade it self. Certainly worth a shot though. Thanks for the input.

 

[CoinOperatedBoy]

Provide other specs?

I just went through a very similar problem with a recertified eVGA 9800 GT that was sent to me as part of an 8800GTS RMA. I've been getting nvlddmkm bluescreens with the 9800 GT and when I could boot, my card was running at x1 link width. There are many, many reported issues like ours. The problem, as previously mentioned, is that almost anything along the chain could be responsible. Simply popping in another video card or two is not an exhaustive test that exonerates all your other components. (For example, your mobo may have issues with PCIe 2.0 cards like the GX2, whereas the PCIe 1.0 8800GTS 640 is fine.) You should:

- Try reinstalling the drivers again, but make sure you use Driver Sweeper / Cleaner after uninstalling to clear out everything Nvidia-related. Then reinstall.
- Clear your mobo's CMOS and reset to defaults.
- Update to the latest BIOS.
- Lock the PCIe bus frequency at 100MHz. If that doesn't work, try increasing it a little at a time up to 110. You can also try lowering it incrementally to 90.
- Increase the voltage to your northbridge.
- Increase the PCIe voltage.
- I know you ran a "memory test", but run memtest86+ for several hours if that's not what you did.
- Check your PSU's 12V output with a multimeter. There are instructions in the power supply forum.

Your tests thus far certainly point to the card being the issue, and that seems to be the case with me as well. I just got my second RMA (free cross-shipping and return label) and the card booted up at 16x with no bluescreen. All you have to do is call up eVGA and tell them your situation, ask nicely for free, expedited shipping since this will be your second replacement, and they will probably accomodate you. I have never had anything but great service from them, even if their product reliability has left me a little underwhelmed.

 

[owcraftsman]

PSU = TX850 Corsair rails multimeter OK
P08 bios
all air antec 1200 + True 120 X
Vista x64 SP1

Not sure if there is anything else you interested in knowing. LMK

That's a very good point about the PCIe spec 1.0 vs 2.0 my other 2.0 GPU's are on Water so sadly I can't try on this system. So as you point out it could still be the mobo. It's back from RMA as well and ditto on the EVGA comment.

I've tried all the above you mentioned with the exception of raising the PCIe voltage and bus freq.. However I doubt this would make a difference. My initial setup was at stock speeds for all components therefore no bumps should be required. If it was required I think it still points to a hardware issue.

At this point I wonder if that would be worth a try considering I'd have to remove the 8800's to do so.

Great Input COB thanks for chiming in.

 

[CoinOperatedBoy]

Originally posted by: owcraftsman
I've tried all the above you mentioned with the exception of raising the PCIe voltage and bus freq.. However I doubt this would make a difference. My initial setup was at stock speeds for all components therefore no bumps should be required. If it was required I think it still points to a hardware issue.

I think you should try that. Some people have reported success with adjusting the PCIe frequency by even 1MHz. Others needed a more significant bump. (I don't recommend exceeding 115MHz... this setting can affect other components in your system, and it is an overclock.) Gigabyte mobos had issues before with PCIe cards, and the temp fix was to lock the bus at 101MHz. And yes, that could indicate a hardware issue, but if it's the mobo it could also be something a BIOS update could alleviate (as was the case with Gigabyte).

It could also point to a power problem. Reports of this issue seem to be more common with factory- (or otherwise) overclocked cards eventually needing more juice, or failing altogether in this fashion, even though no configuration changes had been made.

Personally, I would just push for the free RMA. You can really tear your hair out on this one, and the only real option is to get an exact replacement part to try.

 

[dmh1167]

Hi all, owcraftsman hope I can be of help had your problem before, if you installed Vista with more than one stick of ram installed you can get alot of little bugs appearing everything from install crash to god only knows what after install, second do you have the Vista 64 sp2 beta installed if ya dont get to it man it saved my bacon.

http://service.futuremark.com/compare?3dm06=10809504.

I have room to go further but I like the stability and cool temps on air, hummm you have tried everything accept choking the guy that made it , this may make you scream but your psu might have just enough of a hiccup to pull this,

got rid of my thermaltake 700w modular for 1000hx but your psu is nothing to sneeze at. Try xp on seperate drive if you have a 32 bit copy see if this issue is present in 32 bit enviroment I have xp pro 32 on one drive and vista 64 ultimate on the other it helps me find OS or hardware issues pretty quick. Hope this helps if not hope some one has an answer to your issue. Good Luck PS check your sli bridge who knows it might be sick.

 

[CoinOperatedBoy]

Originally posted by: dmh1167
Hi all, owcraftsman hope I can be of help had your problem before, if you installed Vista with more than one stick of ram installed you can get alot of little bugs appearing everything from install crash to god only knows what after install, second do you have the Vista 64 sp2 beta installed if ya dont get to it man it saved my bacon.

http://service.futuremark.com/compare?3dm06=10809504.

I have room to go further but I like the stability and cool temps on air, hummm you have tried everything accept choking the guy that made it , this may make you scream but your psu might have just enough of a hiccup to pull this,

got rid of my thermaltake 700w modular for 1000hx but your psu is nothing to sneeze at. Try xp on seperate drive if you have a 32 bit copy see if this issue is present in 32 bit enviroment I have xp pro 32 on one drive and vista 64 ultimate on the other it helps me find OS or hardware issues pretty quick. Hope this helps if not hope some one has an answer to your issue. Good Luck PS check your sli bridge who knows it might be sick.

I've heard of the Vista installer bluescreening with certain RAM configurations, but not that users may have issues later if they manage to successfully install with more than one stick. Power is still a possibility, but I kind of doubt that his PSU could handle two 8800 GTS 640s but not a single GX2. On that note, I don't think he's using an SLI bridge. The GX2 is a single-slot, dual-GPU card.

Testing in XP isn't a terrible idea if that's possible.

 

[CoinOperatedBoy]

Any update, OP? Did you go through an RMA, or identify any other problems?

Personally, I received and installed my second replacement 9800 GT from EVGA and so far, so good. No more nvlddmkm bluescreens and I'm running at 16x.

I want to personally commend EVGA again for their level of support. I never had to wait more than one business day to get an email reply from tech support and their troubleshooting assistance was detailed and exhaustive. They gave me no hassle at all when I requested my second RMA be cross-shipped and they paid for return shipping.

My only minor quibble is that they didn't include my apartment number on the shipment, which delayed the delivery and ended up forcing me to go pick up the package myself. Might be worth verifying with them on the phone that they use the correct, complete address if you're going through an RMA.

Regardless, despite my issues with the hardware, EVGA has scored points with me and I'm definitely inclined to continue using their products.

 

[owcraftsman]

I have thoroughly tested with memtest 3 different pairs of memory on the setup at stock speeds and then tight timings underclocked and loose timings overclock. Memtest minimum 6 passes zero errors with all clocks.

At this point I'm ready to RMA again but, here and a couple other places I posted this thread I was recommended to Format and reload by more than one individual so I intend to give that a shot ASAP. If that doesn't do it I will call tech support to see what they can do for me. Frankly I've had it with GX2 and very dissatisfied from head to toe. This has been anything but a plug n play device. I plan to beg them to reinstate my upgrade path and get a 285 or 275 in it's place. I know they have done this for others, because I been told by others as a result of my threads on this subject. So if the Format and reload doesn't work wish me luck.

 

[CoinOperatedBoy]

Originally posted by: owcraftsman
I have thoroughly tested with memtest 3 different pairs of memory on the setup at stock speeds and then tight timings underclocked and loose timings overclock. Memtest minimum 6 passes zero errors with all clocks.

At this point I'm ready to RMA again but, here and a couple other places I posted this thread I was recommended to Format and reload by more than one individual so I intend to give that a shot ASAP. If that doesn't do it I will call tech support to see what they can do for me. Frankly I've had it with GX2 and very dissatisfied from head to toe. This has been anything but a plug n play device. I plan to beg them to reinstate my upgrade path and get a 285 or 275 in it's place. I know they have done this for others, because I been told by others as a result of my threads on this subject. So if the Format and reload doesn't work wish me luck.

Format & reinstall is pretty much the catchall solution for unclear issues. Based on your description, I'm still leaning toward it being a hardware problem, but here's hoping the reformat will work for you.

Asking for a step-up or a card with similar performance seems like a good idea to me. Looks like this particular model has rampant problems and I don't blame you for not wanting to get another one.

 

[owcraftsman]

Format and reload is complete
Current system Specs
Mobo: EVGA 132-CK-NF78-A1
CPU: E8400
Mem: PVS24G8500ELK
GPU: 01G-P3-N891-AR
HDD: WDC WD740GD-00FLA1 (3xRaid-5 C:\ Vista/programs 69GB, D:\ Windows7 69GB )
HDD: ST640032 3AS SCSI (PageFile F:\ 10GB /Data G:\ 590GB
OPT: HL-DT-ST-DVD-RAM GH22NS30 SCSI CdRom E:\
PSU: Corsair TX850w
OS: Vista HP x64 SP1 oem
Drivers:
Chipset: 15.25 whql
GPU: 185.85 whql
Audio: R2.20 (on board)
System Clocks: 9x, 100,Auto, 5x, 5x, Dis, Dis, Dis, Dis,
FSB & Mem Config: Expert, Linked, Sync, 1600,
Mem Timings: Expert, Cas=4 all else auto or 4 4 4 12 CR2
CPU Config: Dis, Dis, Dis, Dis, Ena, Ena, Ena,
Sys Volts: 1.34375 vCore, 1.2 vFSB, 2.00 vMem, 1.35 vSPP, 1.525 vMCP, 1.25 vSPP<->MCP GTLvRef all auto
Dynamic Fan Control= All set manual 100% except nForce fan @ 60% (custom ACRyan 60mm)

After loading drivers>AV>Windows Updates I loaded minimal software:
Everest Ultimate, CoreTemp, CPUz, GPUz EVGA Precision (to control Fan only @ 100%), IntelBurn Test and 3DmarkVantage AE w/ Hotfix 1.0.1 all are latest versions

UAC was disabled before any drives or software at first boot. (super admin not available w/ Home Premium.
All exe's were run as admin
For 3DMV I changed permissions to FULL Control user name: xxxx
I used "Take Ownership" (<<< a good toot here)of C:\ProgramFiles (x86)\Futuremark\3DMark Vantage.

Before running any benchmarks:

* I ran CCleaner,
* Defragged C:\
* I ran memtest for 6 hours w/zero errors
* I ran Windows Memory Diagnostic Tool 5 passes set to Extended with default cache which took about 5.5 hrs to complete. No errors were found.
* I ran IntelBurn Test > error detection ON, Maximum Stress level 1, 3 then 5 passes with no errors.
* I let system complete idle time task for 20 min.
* Last thing I did before 1st run was turn off Vista Aero.

http://img4.imageshack.us/img4/4518/stable3600specs.jpg

I know this CPU will do 4.0+ stable but I choose these setting because ''m on Air at the moment and wanted to keep temps in check.

Results 3DMV AE:

EPIC FAIL

same error as before

http://img7.imageshack.us/img7.../vantageerrorfinal.jpg

http://img16.imageshack.us/img16/883/vantageerror.jpg

I ca''t believe I wasted 12 hours on this card again 3rd format and reload.
Still it could be the motherboard PCIe 2.0 slot since I know the 8800GTS 64''s will run in sli 3DMV AE. As someone pointed out to me they are 1.0 devices so I have one more last ditch effort to make.

You can see in my sig "SysI" a 790 Ultra board running a 280FTW. I want to try the 185.85 drivers on it so when I uninstall the drivers I intend to install the 9800GX2 first to test on that system. In hopes this will eliminate the 780i as the possible cause. The big problem is this system has a triple loop on it CPU/GPU/NB,SB,PW''s so ther''s not much room to squeeze the GX2 in there just for a test but I think I have it figured out so I can put it back together when ''m done.

Unless someone can point out the error of my ways stay tuned for the results of the next test hopefully the final test before I pitch a fit with TECH.

*********************************************************************************************************************



''m SOL I tried loading several other games Crysis, COD5 etc and all failed and would not run same error. I then loaded several GPU stress and performance test if interested in what LMK.
before I moved the GX2 to my Asus Striker II Extreme and again fail fail fail.

It was tuff getting it in there but I managed.

http://img21.imageshack.us/img...474/10019574982504.jpg


I loaded the 185.85 driver then dbl checked it was running in SLI as evidenced by GPUz in the screenie below and again Epic fail.

http://img11.imageshack.us/img11/2413/sossmo.jpg

This tells me that it is not the PCIe 2.0 slot on the 780i. So I uninstalled drivers then removed the GX2 piece of junk and reinstalled the 280FTW and drivers to run a 3DMV bench.

http://service.futuremark.com/compare?3dmv=1050700

As you can see the system is fully capable of running a 2.0 card and 3DMV.

At this point I believe I can safely say i''s the GPU all else seems to be fine.

Your input is welcome especially if you disagree.
<b

 

[CoinOperatedBoy]

Did you ever try adjusting the PCIe bus frequency?

Have you not been working with EVGA tech support the whole time? They might have recommended another RMA long ago and saved you a lot of time and effort. You're bound to get some concession out of them as long as you're polite.

 

[owcraftsman]

Yes I went up to 105 and tested again with same result. I guess I forgot to mention it if you can believe that with all the above verbiage. Honestly I never have had to touch it before below a 4GHz clock and most my test were at stock speed all components.

I lost faith with EVGA tech when they tried to tell me that my first GX2 was normal running one core x8 and the other at x16. This would not be so bad if it were the regular tech and not elevated tech support who transferred me to the head tech for 9800GX2 wo told me this.

I had turned to the forums then for answers. Fortunately all I had to do was ask and I had several people post two instances of GPUz one for each core to show that they all had both cores running at x16. I really didn't have to call that tech back but I did not to get a pound of flesh, rather to enlighten. Did I have to go through that exercise? no. Unfortunately that's what it took to get them (tech suppport) to give me an RMA. I could have had an RMA easily going to there web site but IMHO people abuse the privilege. I read a thread the other day where a guy had trouble and RMA'ed everything in his case. WTF is that all about. We pay for that Sh#t. I'm not made that way man. I would hate nothing more than to find out the GPU was not the problem 2-3 wks down the road when I have a new one in hand. Actually I'd feel pretty stupid. I would think at some point these companies will have to tighten up there policies. I for one will be prepared. Besides I've fixed thousands of problem following the same rituals. Nothing errks me more that to read a guy saying, "I don't know what I did but now it's working". and the thread ends and no body benefits from that. Sadly this thread will end with no better than another Failed part but maybe someone will benefit from the steps I used to defeat it another way. At least some of the methods worked for others without an RMA.

 

[CoinOperatedBoy]

Originally posted by: owcraftsman
Yes I went up to 105 and tested again with same result. I guess I forgot to mention it if you can believe that with all the above verbiage. Honestly I never have had to touch it before below a 4GHz clock and most my test were at stock speed all components.

Of course you shouldn't need to bump PCIe frequency in normal operation. But it can be a temporary measure, and can help further identify the problem. Some report having to go up to 110.

I could have had an RMA easily going to there web site but IMHO people abuse the privilege. I read a thread the other day where a guy had trouble and RMA'ed everything in his case. WTF is that all about. We pay for that Sh#t. I'm not made that way man. I would hate nothing more than to find out the GPU was not the problem 2-3 wks down the road when I have a new one in hand. Actually I'd feel pretty stupid.

I commend your effort, but there is a reasonable point up to which the onus is on the customer to perform some troubleshooting steps, and I think you went way beyond that and just got yourself worked up in the process. That's why it's a good idea to work closely with tech support -- even if your faith in them has been damaged by past experience -- because they should have a good idea of when to tell you to stop and just pack it up. You have no obligation to make up for people who RMA at the first sign of trouble; the manufacturer has an obligation to you to provide a working product and rectify the situation when they have failed to do that. It's not a privilege; you paid for it!

All I'm saying is I wish you could have avoided some of the irritation. If you're personally satisfied by exhausting all conceivable possibilities on your own, then congratulations... I think you just about managed to do it.

Best of luck with the RMA. I expect you'll be asking for a different card.

 

[owcraftsman]

Well I called tech Thursday 05/14 those guys are really friendly and very helpful. As expected I had no trouble getting an RMA. As suggested by others I did way more than was required but I would hate getting a new card having the same issue and it's not the GPU. I would feel pretty stupid at that point.

Someone had suggest I ask to have them reinstate my step up or upgrade path again and I tried. I really do not want another GX2 at this point but no dice that is not available to me. Dayum I had my hope up too.

Apparently there is no guarantee I will get a 9800GX2 either he said, "I would get another GX2 or something of equal or greater value". So I'll cross my finger for a 285 even a 1792MB 275 would make me very happy again.

The good news is they are shipping what ever out to me with a prepaid UPS box to return the GX2. So there is no expense to me at all. Woot!


I know there's been a lot said, about EVGA failure rate but I really have only had serious problems with the GX2 and when something does happen they are simply FANTASTIC.

I'll update again when I get the replacement to let you know what they sent.

wish me luck

 

[CoinOperatedBoy]

Originally posted by: owcraftsman
GX2 or something of equal or greater value". So I'll cross my finger for a 285 even a 1792MB 275 would make me very happy again.

The good news is they are shipping what ever out to me with a prepaid UPS box to return the GX2. So there is no expense to me at all. Woot!


I know there's been a lot said, about EVGA failure rate but I really have only had serious problems with the GX2 and when something does happen they are simply FANTASTIC.

I'll update again when I get the replacement to let you know what they sent.

Nice. Good that EVGA is taking care of you. You should be able to see what product they're shipping to you if you log into the website and go to Member Home > View/Edit My Products. It shows the current RMA status and the replacement part number. Note that the part might change between the time you start the RMA and the actual time of shipment. That happened to me when I RMA'd my 8800 GTS 320; I ended up getting a 9800 GT as a replacement.

 

[PCTC2]

Yeah. eVGA CS is one of the best. I've had many cards and motherboards needing replacement (like needing a new motherboard revision when the Q6600 came out, eVGA sent me one).

As for equal or greater value, it usually is NOT a new card. Usually it's a generation old refurb, which would be the GTX 280. I sent in a 7950GT that was artifacting at idle temperatures and got a 8800GT. My GTX 260 Core 216 SC was DOA and I got a new one next day (I always buy their cross-shipping).

Good luck on your replacement card.

 

[owcraftsman]

Well if I said I was pleased I'd be lying.

I bought this back in August of last year 08/14/08 to be exact. $459.99 total paid, but what the heck, this is what you pay for emerging tech, hugh? Almost one year later I finely get a working unit. About the time I would normally upgrade. Upgrade you say?, well yes, About every year I upgrade to new stuff and I suspect I'm not alone. The GX2 is an enthusiast GPU and as with most enthusiast products they don't come cheap. In there lies the rub. I'm not saying I deserve an updated GPU but the facts are I haven't had the use of one for almost a year and paid premium bucks for it. The facts are most new offerings single card and sli setups out perform what I finally have to use but never got a chance to enjoy when it was at the top of the heap. My 280 sgl out performs it a pair of 260's out perform it. So if I want to have a decent card I have to go buy one, sell the GX2 and swallow my losses. I know this all sounds trite but it's better to vent here then take it out on something else. I know I bought and paid for a GX2 and that's what I deserve, but I'm left with an awful taste in my mouth and feel I could have been treated more fairly. You say, What do I mean by that? I mean something of greater value, given the loss of use and progress of advancements. I'm not saying it should have been a 295, all though I did ask for an upgrade path to be reinstated, which is another source of the bitterness here. In essence I was ignored. I buy emerging tech, I am an enthusiast, and I expect it to work out of the box, don't you? I would have been happy with a sgl. 280 or a 275, even up, and asked for that as well, again no respect. So what I'm left with 3 months shy of a year is a deactivated item at New Egg. http://www.newegg.com/Product/...x?Item=N82E16814130338 It?s a shame.

All that said, Still EVGA remains the best at RMA and support for it's products. I just wish somehow the rules were different. ie: no exact replacement after item is out of production instead an equal or greater valued production product or auto reinstate upgrade path after 2nd RMA, say for 30 days or some EVGA bucks, heck a fricken T shirt for God?s sake, at least more choices. Is that to much to ask? The only real conciliation is I don?t have to pay to send the 2nd card back.

Deep Breath..... ....... OK.... I'm better now...

Since everybody likes screenies (except AnandTech) and I'm no stranger to them. I'll post up the testing of the FINALLY working GX2.

Did I mention it took 9 months to get a working unit. lol

http://service.futuremark.com/compare?3dmv=1088039 (high performance test results oc'ed)

http://service.futuremark.com/compare?3dmv=1087963 (standard test results RMA replacement oc'ed)

Direct links add images



http://img199.imageshack.us/im...stockbenchgx2replm.jpg

http://img33.imageshack.us/img...2/gx2stabilitytest.jpg

http://img193.imageshack.us/im.../498/furmark000002.jpg

http://img30.imageshack.us/img30/6470/fc2brgx2dx9.jpg

http://img268.imageshack.us/img268/7135/stbtest.jpg

http://img30.imageshack.us/img.../fc2dx10defaultrun.jpg

http://img195.imageshack.us/img195/5838/furmarkgx2.jpg


after 9 months blaming Vista, EVGA Bios, EVGA motherboard, Nvidia Drivers, and User Malfunctions, all it was, was a defective GPU, a GX2, two of them?. Go Figure