Again: your problem doesn't seem to be driver related.
Maybe you don't have problems with previous drivers because they had a more relaxed timing, but the problem lies with your own hardware. I've never had such problems (error 43) with any released driver.

Back on topic
 
Update:
 
Well the first cold restart failed

 
This is the exact same failure mode, so the PS2 issue is NOT the cause
 
I am back on 285.79 with which there are no failures or issues of anykind, and right back where I started
 
 

@EVGA
 
To whomever is deleting posts in this thread this morning, I would like to be contacted by PM by you.
Perhaps you are unaware of user Zibri's history of repeated warnings, post edits, post deletions and a prior temporary fourm ban.

Excuse my ignorance, but I've never seen the big deal with drivers. I don't download new drivers unless a new game is out that doesn't work very well on the current ones and there is an updated driver for it. I basically take the "if it ain't broke don't fix it" attitude towards them. These new 295 drivers are apparently optimized for Skyrim, but that game plays more than fine on my machine, so I don't see any point to them. The 285's are working fine.
 
Also, I really don't get the point behind all the fancy ways of removing old drivers and installing new ones. In the last 3 years of owning nVidia cards I've downloaded the new drivers and installed right over the old ones with the installer that comes with the driver and I've never had a single issue.
 
I guess I've just never understood the craze about having the latest driver and swapping drivers in some fancy way. If someone would like to educate me on the issue I'd be appreciative of that, but in browsing message boards it just seems like these things cause people a lot more trouble than they're worth.

Others have reported GPU increase and heat etc with these drivers...possibly power changes to support DX11 more effectively??? just guessing
 
Could there be more of a power draw at cold boot that is somehow rendering one of your PCI lanes "dead" until windows boots...possibly a bad PSU rail or not enough power to support the SLI config at a higher load ? - did you check the connection - literally the plugs on every psu connector...
 
Have you taken out each card one by one and tested each one separately in the system to see if they will take a driver on their own and support multiple cold restarts
 
Just throwing ideas out
 
 

Man I'm stumped

"Man I'm stumped"...
You and me both
 
The most probable causes are a driver behavior change that is incompatible with my current MB BIOS, a driver conflict I cannot identify or a OS corruption in the driverstore or HAL
 
The most frustrating part, is this issue only occurs with drivers above 285.79 on my system.
 
And I am not alone is suffering from this issue, as others with different MB's/Chipsets/CPU configs, both SLI and non-SLI are also reporting failures that match my issues...
 
I have some things to still try on my end, and I do not give up.
I will continue to post updates and any progress.
Thanks for your suggestions and support

I still think your motherboard or your PSU are the culprits.
It's against statistics that both of your cards are faulty, unless you flashed the wrong bios in them, but even then, the so called "videocard bios"  is not code initself but mainly settings for ram and gpu, voltages and timings (and fan curves).

Zibri

I still think your motherboard or your PSU are the culprits.
It's against statistics that both of your cards are faulty, unless you flashed the wrong bios in them, but even then, the so called "videocard bios"  is not code initself but mainly settings for ram and gpu, voltages and timings (and fan curves).

Since you are being civil, please impress me and explain.
 
I am using a AX1200 (same PSU you have) what possible PSU fault/failure would allow SLI to work perfectly with one driver and not another? I have tested the voltage at both Idle and Load with software, and in addition have used a voltmeter to comfirm those readings. All voltages are showing approximately the same values and stablity that I wrote down when I tested my voltage well over 15 months ago, when I installed my PowerBoost module...
 
As for the MB, the e760 while a awesome board is not beyond physical failure.. but I am not aware of any "hardware" failure mode that would allow SLI with one driver and NOT another.
 
I am sticking with "The most probable causes are a driver behavior change that is incompatible with my current MB BIOS, a software/driver conflict I cannot identify or a OS corruption in the driverstore or HAL"
 
Unless you have information to suggest otherwise or another suggestion

No problem, hope it gets sorted out

silverdragon68

mainicvvv....I found this on tom's hardware forums.  Seems you've tried everything else might want to give it a try if nothing else seems to work.

http://www.tomshardware.com/forum/257021-33-code-error

number1ontheweb 02-28-2010 at 01:15:38 PM
Show message
-0+

Not to worry, I've fixed it now, solution is as follows.

I found a few differences in readings on the bottom rear section of the board, but it's literally filled with loads of tiny components it would be a very hard job to narrow it down to a single component since they're all connected on the circuit board and it's been a long time now since I was doing fault finding for my diploma in electronics.

Rather than replace all the components on that section of the board I followed someone else’s guidance that’s also experienced this problem:

1) Removed the heat sink and I/O shield so the card was naked
2) Made four metal balls of aluminium foil
3) Pre-heated my over to 385oF
4) Placed the card on a metal baking try using the balls of aluminium to support the card on four corners (not touching any components)
5) Waited for the over to reach the correct temperature then placed the card in the over for 9 minutes
6) Removed the card from the oven once my stop-watch had reached 9 minutes and placed on the side for 1 hour to cool down
7) After waiting about an hour I re-attached the heat sink and I/O shield then tossed the card back in my computer and voila problem solver, no more vertical green dots on boot and no more error code 43

Please note: This is not a guaranteed fix and although this has worked for me and many others this might not work for your card and could potentially cause more harm. If your card is under warranty then send it back to be properly repaired, but if you're not going to get your card repaired and you're only going to throw it away anyway then it might be worth giving this a go.

The second part of this post made me cringe, He isn't getting green dots he's just having problems with the drivers. 

thePoison

silverdragon68

mainicvvv....I found this on tom's hardware forums.  Seems you've tried everything else might want to give it a try if nothing else seems to work.

http://www.tomshardware.com/forum/257021-33-code-error

number1ontheweb 02-28-2010 at 01:15:38 PM
Show message
-0+

Not to worry, I've fixed it now, solution is as follows.

I found a few differences in readings on the bottom rear section of the board, but it's literally filled with loads of tiny components it would be a very hard job to narrow it down to a single component since they're all connected on the circuit board and it's been a long time now since I was doing fault finding for my diploma in electronics.

Rather than replace all the components on that section of the board I followed someone else’s guidance that’s also experienced this problem:

1) Removed the heat sink and I/O shield so the card was naked
2) Made four metal balls of aluminium foil
3) Pre-heated my over to 385oF
4) Placed the card on a metal baking try using the balls of aluminium to support the card on four corners (not touching any components)
5) Waited for the over to reach the correct temperature then placed the card in the over for 9 minutes
6) Removed the card from the oven once my stop-watch had reached 9 minutes and placed on the side for 1 hour to cool down
7) After waiting about an hour I re-attached the heat sink and I/O shield then tossed the card back in my computer and voila problem solver, no more vertical green dots on boot and no more error code 43

Please note: This is not a guaranteed fix and although this has worked for me and many others this might not work for your card and could potentially cause more harm. If your card is under warranty then send it back to be properly repaired, but if you're not going to get your card repaired and you're only going to throw it away anyway then it might be worth giving this a go.

The second part of this post made me cringe, He isn't getting green dots he's just having problems with the drivers. 

just a thought to try and help....btw there was another user in that same forum who was having similar problem as maniac (but no green dots) and he baked his cards and hadn't a problem since... I agree with outrage...update the mobo bios, reload windows from scratch, update the vid cards' bios one card at a time then try them in sli with the newest whql drivers.

silverdragon68

thePoison

silverdragon68

mainicvvv....I found this on tom's hardware forums.  Seems you've tried everything else might want to give it a try if nothing else seems to work.

http://www.tomshardware.com/forum/257021-33-code-error

number1ontheweb 02-28-2010 at 01:15:38 PM
Show message
-0+

Not to worry, I've fixed it now, solution is as follows.

I found a few differences in readings on the bottom rear section of the board, but it's literally filled with loads of tiny components it would be a very hard job to narrow it down to a single component since they're all connected on the circuit board and it's been a long time now since I was doing fault finding for my diploma in electronics.

Rather than replace all the components on that section of the board I followed someone else’s guidance that’s also experienced this problem:

1) Removed the heat sink and I/O shield so the card was naked
2) Made four metal balls of aluminium foil
3) Pre-heated my over to 385oF
4) Placed the card on a metal baking try using the balls of aluminium to support the card on four corners (not touching any components)
5) Waited for the over to reach the correct temperature then placed the card in the over for 9 minutes
6) Removed the card from the oven once my stop-watch had reached 9 minutes and placed on the side for 1 hour to cool down
7) After waiting about an hour I re-attached the heat sink and I/O shield then tossed the card back in my computer and voila problem solver, no more vertical green dots on boot and no more error code 43

Please note: This is not a guaranteed fix and although this has worked for me and many others this might not work for your card and could potentially cause more harm. If your card is under warranty then send it back to be properly repaired, but if you're not going to get your card repaired and you're only going to throw it away anyway then it might be worth giving this a go.

The second part of this post made me cringe, He isn't getting green dots he's just having problems with the drivers. 

just a thought to try and help....btw there was another user in that same forum who was having similar problem as maniac (but no green dots) and he baked his cards and hadn't a problem since... I agree with outrage...update the mobo bios, reload windows from scratch, update the vid cards' bios one card at a time then try them in sli with the newest whql drivers.

My cards have been babied and treated like treasure. They both run on stock 1.000v are are not overclocked.
They have never been overheated and my room is kept at a standard temp year round, which really cuts down on PCB expansion/contraction 
As such it is very doubtful my cards are suffering from PCB mirco cracks or weak dip points, if they were then my issues would be happening no what what driver I used. I dont not belive that baking is appropriate given the issues I have seen.
 
I do thank you for your suggestions and support bro, every little bit helps... Thank you

Zibri

maniacvvv

My cards have been babied and treated like treasure. They both run on stock 1.000v are are not overclocked.
They have never been overheated and my room is kept at a standard temp year round, which really cuts down on PCB expansion/contraction 
As such it is very doubtful my cards are suffering from PCB mirco cracks or weak dip points, if they were then my issues would be happening no what what driver I used. I dont not belive that baking is appropriate given the issues I have seen.

I do thank you for your suggestions and support bro, every little bit helps... Thank you

Mine have always been overclocked and overvolted to their fullest (on air) and I've never had any problems (not even with my previoius dual GTX 460 setup.

Again, maniacvvv, I'm confident your problem is in your mainboard.
I suggest you to try 2 things:

1) find another mainboard, same model/manufacturer and swap them. If the problem persists then it's something about the bios settings (as I suggested, raising the pci and qpi will greatly help stability on dual/triple sli setups).

2) try a different PSU (corsairs ax1200 is a good choice).

I'm not saying to ditch your hardware but just find a friend and do a swap just to test it.

Question:
what happens if you physically switch off the PSU and wait 10/30 seconds before restarting? Do you still get the unrecognized device error? I doubt that since I think it's an initialization problem.

As a workaround you can try to increase the PCI timings if your MB allows you to.

1) find another mainboard, same model/manufacturer and swap them. If the problem persists then it's something about the bios settings (as I suggested, raising the pci and qpi will greatly help stability on dual/triple sli setups).

Why on earth would I want to swap MB's without trying a flash and OS reinstall?
Given that many other are having the "same" issues with the newer drivers, but using different MB/ CPU configs I dont see it as a hardware issue....
As for my overclock settings (at 3.7ghz with a 3.9 turbo on core zero) there have not been any issues with stablity since 2009. I have never had any issues, and dont forget that 295.73 ran thru testing with Mark11 and heaven andat least 3 4hr sessions of BF3. The problem has nothing to do with stablity and everything to do with how the cards are initialized on a cold boot.
 
2) try a different PSU (corsairs ax1200 is a good choice).
I have a AX1200 and it is completely tested, its not a PSU issue.
 
3)...I think it's an initialization problem
I 100% agree with you on that point.
Nvidia began with the 285.xx series of drivers to implement a new OS hardware layer for Windows 8 devices that appears to involve changes to the HAL and a bypass of some of the "older" legacy Windows hardware device models. I do not pretend to understand these changes, only to note that when they occured-->my problems began.
Since my Nvidia drivers work perfectly and without any issues of anykind up to 285.79, what has happend is clearly due to Nvidia changing its driver initialization protocols...
It remains to be seen if this will be patched out, is because of a BIOS or a driver I am using. 
Because of the performance and testing levels of my system running 285.79, I am quite ceratin it is NOT a hardware issue, but a software based problem.
 
Thank you for being civil, and your suggestions and comments are welcome as long as you continue to be so. 

I don't know if it helps maniac, but there is 100% fact that these latest drivers have some sort of power initialization error/bug on the pci-e lanes dependent on what power state they are in. Some of us, myself included, have posted in the f@h section of these forums about this issue and that a new WU wouldn't start if the cards video signal was in a certain P state.
 
Im willing to bet your issues are somewhat related to this and its most likely a (poor) attempt by Nvidia to introduce some new power-saving code into these latest drivers.