graphics card memory issue

Fault finding has become a major headache over the last 10 years or so given the ever increasing complexity of the hardware and MS propensity to dumb things down for the end user. 

Sorry about the rant but I have had a ton of issues with TDRs myself, to the point I RMAd my GPU 3 times, changed motherboards, RAM, PSU and flashed to every possible VBIOS in combination with every driver since the release of my GPU. Eventually I went nuts and performed a "clean slate" installation of all my parts and software and have been mostly error free since. I still am not certain what fixed the issue and I have been in this business since the 90s..... It may be the driver I installed along with the fresh install as it was released around the same time. Even now however, some particular games still give me random TDRs, yes, even the games themselves can cause this issue.

This is such a well known problem there are literally hundreds of articles, threads and forum posts about it, all with different advice and potential "fixes". Part of my problem turned out to be a particular game I used to play a lot, a patch was released for it and the TDRs stopped. 

I think the most annoying thing about this error is the fact that it could be a hardware fault. If it was a given that it was only caused by software or drivers it would simply be a case of finding the right driver/software combination to get your system stable but knowing it could be an indicator of a bad GPU makes you constantly stress until you can resolve it. Fortunately, the likelihood of it being a bad GPU is remote and easy to ascertain. If underclocking/volting the GPU resolves it, the factory OC and or Boost of the card is not stable and the GPU should be RMAd. I wish I had learned this little tidbit before 3 RMAs and weeks of frustration.

I mostly resolved my problems with my GTX 960 by cleaning out my drivers with DDU (Display Driver Uninstaller) and trying each revision that supported my GPU till I found one that was stable. The problem is that sometimes the fault is with the GPU not being stable at the boost clock with the voltage set in the VBIOS and drivers. Based on your post however I am guessing your system did not have these problems until you upgraded to windows 10 which means it is almost certainly a driver issue from Nvidia with your particular configuration. 

One thing you can try to rule out the possibility that Nvidia are right about needing a BIOS update is to use MSI afterburner and set the "Power Limit" to about 70% and see if the card runs stable. If it is stable then all you likely need to do is wait until Nvidia release a more stable driver, you can run the card at the lower power limit till then without any risk, in fact, the GPU should run cooler and quieter. You will lose a few FPS here and there but not more than 4-5%.

If the GPU is still unstable with the lower power limit however it may be a compatibility issue with windows 10 and your setup. The problem is that there are a lot of things that can cause this particular problem, TDR (Timeout Detection and Recovery) is the name of the problem if you want to google it for more info. 

Just an FYI, don't bother with the TDR tools that you will find, they do not solve the problem, they only change its behavior and can cause damage to your file system if they cause lockups and BSODs.

I know this is not likely the answer you were looking for but I hope it helps none the less.

Good luck