What is the crowd's opinion of UNIGINE Benchmarks?

From 3.5 billion comments

Download, Heaven, Valley, & Superposition from here for stability testing. Bump your core clock up 1st in small increments (I would say find what clocks boost to 1900Mhz then start bumping up in 10 to 20 Mhz increments from there) till you start to see artifacting in the benchmarks. Once you see find your limits of your card bring the core clock back down a little bit to get back to stable (15 to 30Mhz should get you stable again & note you can increase Vcore voltage to get stable on the core as well but that is not really recommended with Pascal cards since their clocks are very temp dependent). After you've found the highest core clock you can run do the same with the memory. Once you got both dialed in let one of the benchmarks run on a loop for a couple hrs & keep an eye out for any instability. If it passes that you are good to go.

> I enabled DRI3 as you said and restarted.. But this time, I am getting same fps for both GPUs i.e 60fps in glxgears, and glxgears window is now visible

It's mean that's working. Only thing you need now is actual benchmark.

PS: That's important because even if you disable VSync for glxgears you'll possibly get worse VPS on dGPU since PRIME will be always be bottleneck for too-high FPS.

DRI_PRIME=1 vblank_mode=0 glxgears

Assuming that you've also done some graphics drivers updates here and there, and that you're not running something in the background only while playing WoW and Lol (and not dying light), then I would assume yes.

Some games use different resources on the graphics card and could be the reason why its not crashing all the time.

I would run a couple synthetic tests like Unigne Heaven to see if they can reproduce a crash

http://unigine.com/en/products/benchmarks/heaven

well you can use the UNIGINE benchmarks and the games mentioned in this article