Michael

Sure, I get not liking cabliing.
 
But this is incorrect:
 
"the benefit of having fast read and write speeds of NVME SSD. "
 
NVMe does not produce that benefit in real world usage and stopwatch timing.  
 
But your idea should work, may need some BIOS tweaking or playing around to get it to see the addon card.  Best of luck.

If they can take advantage of CUDA and/or RTX then yes, you'll see a performance improvement.

If the programs are gpu accelerated and can have ray tracing acceleration then yes. Of course they'll still be held back by a i5 2400. It's not too uncommon to have a weaker cpu and really beefy gpu in 3d workstations when all is gpu accelerated as you don't have a bottleneck then but this i5 is well below that point so you will not see the full power of the gpu even when rendering or in viewport performance. Still massive improvement compared to your 560. However you need a new psu as that thing is not capable of driving any of those cards safely.

Nice to see this going on and the community pulling together. I've been reading up and looks the folks are working on bringing more WU's into the system as well as more back end architecture to support the surge!
 
Congrats and thanks to all for working and donating!
 
Cheers and love to all!
Whaler
 

Probably the PSU. When the GPU is powered, it is probably taking too much energy lowering the voltage

Sounds like PSU to me. Should try a different unit to be sure.

I've tried updating the bios and it worked! Now the card is running at full x16 and I've managed to squeezed couple of fps more on my games! Thanks bro

There are two PCI-E slots on that motherboard. One is Blue and one is Black. Make sure it is installed in the Blue one. 
 
If you have it installed in the Blue one already, try updating the Bios on the motherboard:
 
http://www.asus.com/Motherboards/P8Z68M_PRO/HelpDesk_Download/

@WhatTheQuack Neither mine does not have manual aperture. I've shot it on it's widest aperture so the DOF is so narrow. You probably won't see it on the image above because the subject(the LCD screen in that case) is totally flat.
 
@Weck I'm using canon. Aperture blades on the lens will engage ONLY during the length of exposure. What I did is I hold the shutter while in bulb mode so that the aperture blades are engaged and then detach the lens itself. BTW, I didn't think of that idea before! I'm shooting macros for a while now with reversed lens and the very shallow DOF always bothers me until now! Thanks!!

Guys, I don't think that they are really dead pixels because they respond on color changes. I took a short clip while running jscreenfix(see GIF below) and as you can see, you can barely see the "affected" pixels. I guess I just have to live with it. It's just quite unfortunate that it's almost on the middle of the screen.  
 

 
@JR1911 I hope thats my case. 
 
@WhatTheQuack I shot it with my dslr with "reversed" lens(to make the focusing distance close enough for macro shots)  

Wait and see if you can get someone with one on hand to give you an answer for sure but if no one can give you more info I would deduce:
On the molex pin 1 is 12V and pin 2 is ground then the fan headers wire will be sense, on the board side I would assume the contact point with the square surround is ground because it will likely be attached to the ground plane rather then a tracer, I cant see very well with that picture but you could confirm if that point was connected to a large area of copper rather then a small copper track. Then for the other two wires there are two possibilities (well no duh but I have two hypotheses too ) if it follows fan header standards the middle contact would be 12V and the final (top one in the picture) would be sense however it would probably be more likely that the 3 wires would be all soldered in order rather then twisting two at the last moment and by the looks of that picture the sense wire is in the middle? again you would be able to see better then me. If you are at all familiar with electronics too you could follow the tracers to see where they end up, the 12V one will likely lead to some sort of regulator while the sense is more likely to feed off a chip but may have a FET or similar in between, you can also check the chips datasheets to see if a 12V rail should be connected to said point or if its something else (hopefully a sense pin ).
To surmise;
Ground will be connected to the contact that is on a large copper plane.
12V and sense will either follow standard fan header layout or more likely be in order of the 3 pin cable once you've determined where ground connects.
I've got to admit I'm curious as to how you accidently detached that too lol

It's not brand new and no warranty when I bought it so I have to "somehow", ghetto-fix it's rattling noise (and thankfully, I've managed to fix it) so I have to dismantle the cpu-block assembly. And the only problem that I have now is refilling it. I've tried to refill it by forcing the liquid into the hole under the cold plate using syringe but I can't fully fill it up so it is so noisy because of air packet when I test it. So I guess, I have no choice but to fill it up by removing the tubes.

A stock intel hsf on my 2500k.
And
A pentium 4 - Willamette + mobo with ridiculously, stupidly rare type of ram, a RDRAM. Bought that several years ago and found myself stock with 64mb of ram. I wasnt able to upgrade it because I couldnt find one on sale.