AMD Radeon R7970 crashing constantly


#1

Just installed this thing and after getting the drivers running smooth i put the card through it’s paces and downloaded + ran some games without any problems. Next morning when i booted up it barely got to my desktop before both screens went all scrambled and the machine rebooted. I have tried a number of things without any luck, at first i figured maybe it was overheating, so I pointed a fan at it and managed to get it to boot and run a bit longer, put MSI afterburner on it to monitor the heat and tweak the fan settings. again everything was looking great, the card was running at a cool 55C when bam same thing everything crashed.

any idea’s of what the heck is going on?


#2

I’m pretty new to pc building/upgrading, so my knowledge is limited, but what is your powers supply rated at?


#3

700W psu which is brand new.


#4

Try an older AMD driver, AMD loves to release butt drivers sometimes.


#5

I’ll give this a shot tonight, it’s been a while since i got a completely bad AMD driver… but it has happened to me before and i hadn’t thought of it this time around.


#6

so perhaps i’m just retarded… but am i going to the right place to find older drivers… or has AMD removed them? support.amd.com/en-us/download/desktop/previous?os=Windows%207%20-%2064


#7

did you buy it refurbished by chance? i’ve heard of people having to reapply thermal paste for used video cards. also, amd cards usually run on higher power draws so you could be overloading your PSU like camacho mentioned. 700w might be too low, especially if you’re running an amd cpu + gpu


#8

I always over estimate how much wattage I need for my PSU in my PC builds. I always take values of my parts, then round up. Then from my calculations I pick the PSU atleast 1 up higher than I need.
Some people argue I go too high, but there never any harm getting a higher capacity PSU.

Example if I build a PC where the typical PSU online calc says I need a lets say 431 watts, I would not even look at a 450 watt and I start with the 500 watts.
I also look at the efficiency of a PSU, if its not at least 80% I don’t bother.


#9

700w is overkill for that GPU


#10

Yeah, 700w is way more than enough, unless it’s a shitty Diablotek or something on that level.


#11

the card was purchased from a friend, but the thermal paste was re-applied. And i ran a

I did test other drivers, 1 version back. the newer version that comes with windows 10 and a couple other things. Looked deeper into the event ID’s and found that it’s either a mobo issue with power regulation or it’s a issue with the ram in the GPU. the event ID is “WHEA-logger event ID 20.” and it reports the issue as being AMD northbridge.

I did manage to rule out overheating by monitoring GPU temp with afterburner while it crashed, i saw crashes at 50c so… that’s out.

last note when i looked at the mobo specs i realized i am plugging this card into a old pcie 1.0 slot so maybe it just can’t handle the data?

Oh yeah, and while running tests i left it on in safe mode for a long ass time and eventually these messed up vertical bars show up on the display where the picture would like tear and colors would stay in those bars.

Currently my plan is to try to swap out a couple components starting with the mobo.


#12

is it possible to have too large a psu for your gpu?


#13

Nope, your PC will only draw as much power as needed from your PSU.


#14

No. Its better to have a larger than you need for a PSU than too small of a PSU.


#15

newer cards work in older mobo’s they just run at slower speeds. from your description of the problem it looks like your card may have been heat damaged. radeon cards are designed run at higher temps and alot of folks overclock them without changing the cooling system for the chip. depending on the conditions this can lower the lifespan of the card from 3-5 years to around 1-2 years but for people that upgrade them regularly this isn’t a big deal. run a google images search on ‘graphics cards artifacts’ and see if this matches up with the scrambles you saw before.

to add a personal anecdote, i ran a gta 4 mod on my pc about a year back and ran up the gpu temp to crashing a couple times. later on this caused the crashes to come more frequently with less graphics intensive games like dota 2. i also reapplied the thermal paste which held up for a little while but eventually started to crash again even at lower temps like 54c.

i’m hoping your friend didn’t try to rip you off or anything, that would be a jerk thing to do to someone trying to get into computer gaming. but it is for this reason that i never buy used or refurbished cards. there is no telling what kind of conditions it was run under and until you run a couple of furmark benchmarks yourself you can’t know if it is heat damaged or not. as far as fixes go you can check and see if it is still under warranty (it probably isn’t). if not, you can try a ghetto reflow on the chip with a heat gun, you wouldn’t have much to lose at this point anyway. good luck


#16

I really doubt that my friend was intentionally trying to rip me off, but what i’m seeing is most certainly artifacts. my friend offered to refund me if the card doesn’t work… but i took for fucking ever getting around to ordering the psu to put it in and fire it up so idk if i want to even ask for my money back at this point. I’ll give the reflow a shot this weekend.


#17

did your friend ever mention anything about bitcoin or crypto currency mining. amd gpus were very popular when the mining craze was at its peak. buying used amd gpus a year ago was very risky since crypto currency mining over stresses the card way more than gaming does and shortens its life by a significant amount.


#18

nah, he never got into bitcoin