Driver crash, nvldmkm Error 4101, 2x GTX670 SLI

The title is pretty much self explanatory and i need some help.. i've read through lots of forums and didnt really find something that could help, yet, although i'm still testing. The computer setup is as in my description but for everyones ease:
Rampage IV Extreme
i7 3930k
Vengeance Pro 2133MHz (run currently at 1333) 8GB -previously had a 16gb vengeance kit but gave memtest errors-
2x ASUS GTX 670 DCII

I really cant find what the problem is. Tried clean windows installation, a driver wipe after a few months amd reinstall, nothing. Applied a windows fix that increases polling delay (i think), didn't work. Tried setting tdrlevel to 0 as some said but that resulted simply in a total system breakdown.

Mainly this happens AFTER heavy load. For example games like Skyrim or Battlefield 3 will quite rarely give me any issues while playing but they will cause issues as soon as i close them, resulting in 1 or maybe up to a few consecutive crashes in a row in the next minute or so. Black screen for a couple of seconds and then back on, as before. Only exception is League Of Legends that only gives me crashes while playing and never afterwards.

I have ample power (ax1200) so it's probably not that, unless there's a problem in the wall outlet - which i highly doubt.

Cards are not overclocked, in fact this happens even when underclocked (although not after tonights efforts -windowsfix etc- although i dont think it would make any difference.

If there is someone who can read some log files maybe and give me a hint (for example i saw a forum when that cause of the issue was outdated wifi drivers) i'd really appreciate it, as i dont really have that kind of knowledge.

If my writing is somewhat fuzzy, im really sorry, it's just that it's 6AM, spent the night searching what to do (it's been around for quite a while but that's the point where i couldn't take it any more) and i really need my sleep

Lastest Beta Drivers (326.80) but the problem has persisted for several versions, including WHQL

I had done that at some point, but it's a few months ago so i dont really remember how far behind i went, the problem persisted though not matter hat version. That said, it was not a complete driver wipe, only uninstal.. Maybe i will try that later today, as well as taking out the cards and testing them seperately, or in different pci slots. What im worried about though is the probability of it being software corruption/conflict..

Something i think i've noticed is that the time after boot plays some kind of role -PROBABLY-. After boot i can run unigine Valley for 2-3 consecutive loops (maybe more, i just haven't seen yet) and everything is ok, but later on even 1 round can give me a crash. Again, i haven't properly tested that, it just hit me, i think i've noticed such a behavior, so take it awith a grain of salt. I will test further.

There is something i noticed, when windows automatically download display drivers they also download hdmi audio software (or that or somehow else these were installed on my pc, and i hadn't installed them)

When i found this i had already removed both cards and placed the first one (previously second) back in but hadnt started testing. While testin Unigine valley with a singhle card (after having removed hdmi audio) i had no crashes whatsoever. I let it run for 5 mins in DX9, came out clean and then 10-12 more minutes DX11, which also came clean.

Now i placed the remaining card back in and i am running test with sli disabled (which if i remember correctly also gave me crashes previously)

Will test further and report back