My computer reboots on its own (no minidump even when set to output) when I have video playing, especially multiple videos at the same time.  I haven't been gaming much, so I can't really comment there.  I was running Windows 8.1 and had issues and now I've recently done a clean install of Server 2016.  The issues are currently still plaguing the system.  The only hang up with 2016 is the fact that it makes you enable audio devices manually upon boot due to the multipoint service I have running.  I didn't realize it at the time and thought it was drivers driving me crazy, and rolled back to 385.28.
 
Now Before I start messing with drivers,  how can I be sure the drivers are the issue and I'm not having an issue with the PSU?  I have a 1.2kW PSU from 2009 I believe.  I once had 4 GPUs running for folding at home and then BTC mining.  I wouldn't be surprised if the heavy load did a number on the PSU.  I'd love to just buy a new PSU, swap it out and see if there is a difference, but A) 1.2kw+ PSUs are expensive, and B)  cable management is a nightmare. 
 
current readings are:
172.4 Volts
3.62 Amps
620 volts
 
There seems to be plenty of headroom for a 1.2kw PSU, but I'm not an expert.  I can stress the system under a real load and see how it responds, but current crashes are happening with video playback and nothing too strenuous. 

I'm downloading fur mark.  I'll run that with an eye on E-LEET voltage monitoring.  I'll post results in a bit. 
 

Sajin
Put the system under a real load(cpu & gpu stressted at the same time), and monitor the 12v rail with a multimeter to see if it dips to 11.4v or lower before the crash occurs. 11.4v or lower with a crash will tell you it's the psu.

Steady 12v with a crash points to a gpu issue, or the 5v & 3v may be falling out of spec causing the system to reboot.

 
Currently running fur mark after disabling the multipoint service. 
3.3v @ 3.26
CPU VCORE  @ 1.15
DRAM @ 1.65
CPU VTT 1.17-1.18
NB @ 1.10V
5V @ 4.83
12V @ 11.96-12.04
VSB3V @ 3.17
VBAT @ 3.1
 
The only other thing that stands out is that this motherboard is an RMA board.  It was a pain to get the RAM working because I had to play musical slots with the ram until it would recognize all the ram available.  I forget how many GB it wasn't seeing (I think 2-4GB), but right now I am running it with about 8MB unallocated for whatever reason.  I mentioned it to tech support and he suggested swapping places with the RAM until it was recognized.  I figured 8MB wasn't enough to sweat and to move on for the time being.  All 6 RAM slots are occupied and the RAM is the same since the day I built the computer in 2009.  The only parts that really change are the HDDs and the GPU cards.

I've got the x58 758 3x SLI board
i7 920 D0 @ 2.667 stock
NB Frequency 2653.4 MHz
OCZ Platinum 12GB (6 x 2GB) 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800)
CAS7 running @ 9 current frequency 663.4 2:10 tRAS 21 tRFC 74 CR 2T
GTX 950 SC+
R9 290
BFG 1.2kw PSU
Antec 1200 with good air flow and a V10 TEC
 

Kanti
I'm downloading fur mark.  I'll run that with an eye on E-LEET voltage monitoring.  I'll post results in a bit. 

You need to use a multimeter to monitor voltages as software readings can be inaccurate. Running furmark & prime95 (small fft test) at the same time will load the system properly.

Does your MB have dedicated Voltage test points ?
 
Please list your full system specs

Cool GTX
Does your MB have dedicated Voltage test points ?
 
Please list your full system specs

EVGA 758 tri sli.  Guess I had a a typo above (750).  pretty sure it doesn't have the test pads.
 
Will testing the PSU outside the system show a faulty PSU if the PSU isn't under load, or does jumping it with a paperclip force it to max output?  I have multi meters, but I've never had to test a PC PSU before.  I used to repair PSUs for consumer electronics, so if possible I'd like to fix this PSU if it is faulty.  Swapping capacitors is easy enough.  I was on an assembly line though, so I never got to play with oscilloscopes etc. 

to the psu is like any hardware  if your is old and heats up it can throttle  / become unstable 
 
like I say BFG been out of business a long time  even as cheap as I am  I'm a firm believe  to have good fresh quality build psu  power  on my primary rigs  .   in the end that BFG maybe showing its age ??
 
thing about the paper clip test is that just turns on the psu  and no way to hold long term full load on the psu to see if it fails under stress [like the hot box test]   it may allways check out good sitting there doing nothing but being on   [opinion]   under load things heat up  and that's when things can start to fail
 
thing is if todays system pulls 45amps full load a 1200w at 100amps if single rail should not even flinch   your show 40amps max per rail  combined up to  1180w [ 98amps max total over all rails ]
 
just a 980ti card at 250w  is up to 20amps    250w / 12v = 20amps 
 
hard to say here  that age of your build to me factors in   11 year old psu ??   tough  call
 
good luck 
 
 
 
 

wmmills

Kanti
I'm downloading fur mark.  I'll run that with an eye on E-LEET voltage monitoring.  I'll post results in a bit. 
 

Sajin
Put the system under a real load(cpu & gpu stressted at the same time), and monitor the 12v rail with a multimeter to see if it dips to 11.4v or lower before the crash occurs. 11.4v or lower with a crash will tell you it's the psu.

Steady 12v with a crash points to a gpu issue, or the 5v & 3v may be falling out of spec causing the system to reboot.

 
Currently running fur mark after disabling the multipoint service. 
3.3v @ 3.26
CPU VCORE  @ 1.15
DRAM @ 1.65
CPU VTT 1.17-1.18
NB @ 1.10V
5V @ 4.83
12V @ 11.96-12.04
VSB3V @ 3.17
VBAT @ 3.1
 
The only other thing that stands out is that this motherboard is an RMA board.  It was a pain to get the RAM working because I had to play musical slots with the ram until it would recognize all the ram available.  I forget how many GB it wasn't seeing (I think 2-4GB), but right now I am running it with about 8MB unallocated for whatever reason.  I mentioned it to tech support and he suggested swapping places with the RAM until it was recognized.  I figured 8MB wasn't enough to sweat and to move on for the time being.  All 6 RAM slots are occupied and the RAM is the same since the day I built the computer in 2009.  The only parts that really change are the HDDs and the GPU cards.

Sometimes cleaning the gold fingers of the ram sticks themselves will get them functioning like normal, which that's all it might be since they have some age on them and will usually get residues and film on them from being handled etc... over the years. Just get a couple qtips and dip them in 91% or better rubbing alcohol and clean the ram stick fingers with it. Then use a papertowel, or a soft cloth, to make sure they are dry after cleaning. I use this trick on old stubborn parts all the time for people, but ram and gcards are the majority of the hardware I do it to. I don't see your pc specs so I cant comment further on those voltages looking right, but the psu rails all look to be in spec just running toward the lower side a little but there not bad.

I just ran MDSCHED and got zero errors.  System states that I have 9MB in hardware reserve though.  Before I pull the DIMMs, are there any other tests to perform?  I reckon I can run MDSCHED in extended mode. 

Kanti
Running furmark and prime95 at the moment with these numbers:
4.86 on the red
11.93 on the yellow
3.17 on the orange.
 
 
Total draw right now is apx 1010W and creeps up to 1090 on occasion. 
 
If I remember correctly my PSU is the BFG EX-1200 (1200W) Modular PSU
Specs off a review site (https://www.overclock3d.net/reviews/power_supply/bfg_ex-1200_1200w_modular_psu/1):
• ATX12V 2.2/EPS12V 2.91
• Quad +12V Rails
• Silent 135mm Intake
Fan
• Efficiency: > 80% Typical
• MTBF: > 100,000 Hours at 25°C,
> 20,000 Hours at 40°C
• 1200W Continuous Rated at 40°
• PCI Express
Ready
• Safety Approval: UL, CB, TUV, CE and FCC
• Dimensions: 15cm x 8.6cm x 16.5cm (5.9" x 3.4" x 6.5")
 
 

Yeah, that PSU is a dinosaur at this point, not that I don't own a few dinosaurs myself, lol.   Since it has 12v multirails it means you have to better distribute your loads and the 1200W Continuous Rated at 40° is a red flag cause most decent psu's are rated at 50. Also, a mtbf of 20,000 hours at 40c is not too great.... not just because of the temp, but most good units usually start at 40,000 hours I believe. Now if they said 100,000 hours at 40c I could probably forgive them some. For being as old as it is and used as long as it has been the voltages aren't horrible. They are in spec, but on the low end. Your 3 rail is the lowest since 3.13 is as low as it should go by spec. The 5v min is 4.75, so you can see how low they are going. They get much lower and you will most likely start seeing weird issues like hdd crashes or bsod's since the 3v and 5v rails control your hdd/ssd's, peripherals, etc...