R9 290 dieing? video inside

[chromic]

https://vid.me/bmM0 so basically, stock clocks/speed XFX R9 290 Core edition (reference cooler) does this when on heavy load, test can be replicated for the same results.
The temperatures look normal to me, for an stock cooler 290 atleast
Core 92c
Vrm1 67c
Vrm2 74c

Card around 2 years old.
specs:
MB: MSI Z97 Gaming 5
GPU: XFX R9 290 Core edition
CPU: I5-4670k 4.5ghz - 1.28v delidded
RAM: Team group vulcan 1600mhz 2x4gb ddr3 1600mhz
PSU: XFX 750W "XXX" edition
storage :Kingston 128gb ssd, Seagate 1tb hdd

 

[grimfox]

Sorry I don't have audio here, was the PC shutting down or just the display losing signal? Have you tried using a different output IE switching from HDMI to DP or DVI? or visa versa? That would eliminate connectivity issues.

It looks like the 12V is sagging it looks like it starts out and slowly falls which without the above information leads me to it's more of an issue with the PSU.

 

[chromic]

Shuts down, but all lights stay on (keyboard/mouse/power button)

Edit: actually it doesnt even shutdown, only black screen & gpu fan stops "screaming"
More edit: GPU-Z 12v isnt accurate, most of people it shows 11.75v & 11.63v under load for 290's (quickly googled it)

 

[chromic]

Anyone?
It seems to be somewhat of temperature related, i tried fiddling with fan profiles, under/overclocking/drivers etc.
Even put a support to the card so it doesnt "hang"
No matter what, when the temperature climbs to the 90c range, it fails.
900/1200 clocks - 91c gone
947/1250 - 92c gone
1000/1300 91c gone

I remember some time ago when i was playing i could see 94c (basically what these furnaces run at) regurarly with no problems.
Now i even cant run some higher end games, cause they will cause my GPU to throw up, games like battlefield etc

 

[grimfox]

90c is near boiling. It's very possible temps could be locally much hotter, 100c for instance, in places a little farther from the temperature probe. Over 100c any moisture in the system, even that related to humidity in the air, would expand and cause issues. It might be worth taking the GPU apart and replacing the thermal compound (use something non-conductive) and probably the thermal pads too. You may find that the compound has degraded over time or that something is not fitting as tight as it was originally. running at high temps for prolonged periods might be fine for a new GPU but as they age it's harder for them to maintain that load in those conditions. I unknowingly ran a CPU for a year with the HSF only making partial contact. It ran hot 90c+ but it did run. After a time it throttled more and more it just couldn't keep running that hard. Now I run a properly installed AIO cooler and it never goes over 60C but it doesn't throttle anymore either.