Halp with 290x problem

ASCI Blue

ASCI Blue

New member
So my computer just seriously hiccupped on me, including cutting off fans for the GPU. I'm running a Corsair CX 750 power supply which I've read aren't the best when pushed.

Machine follows:

A10-7850
ASRock fatality FM2+
16 gig AMD R9 RAM
4 SSDs of various makes
300 gig Velociraptor
dual monitor setup

I've been having issues with odd screen artifacting and consistent but random driver crashes which recover sometimes I can continue using my system, sometimes not.

Is this a power supply issue? Should I check out the eggspert power supply list and pick a not shitty one nowish and get it here ASAP?
 
If you only have one 290X you shouldn't be pulling anywhere near 750W, even if overclocked. You shouldn't even be pulling over 500W. That isn't to say the PSU couldn't possibly be responsible, but it shouldn't be anywhere near its power limits. Corsair CX750, whether it's the best or not, is hardly a no-name. You shouldn't really be having problems if you have a PSU from any reliable company.

To me it sounds more like the graphics card is at fault. Are you overclocked? If so, try dropping back to stock and see if it helps. There's an outside chance of it being a memory problem, but that shouldn't generally cause screen artifacting.
 
It is a factory overclocked card, I dropped the GPU back to 1 Ghz from 1050. I also had the fans up in the event it was a heat issue. The card is an Asus DirectCU II OC.

My prior card was a 270 'nilla, and it ran like a champ with zero issues, it also draws about half (?) the power of a 290x.

The CX line from Corsair is their step or maybe two above their cheapest units.

The other oddities I've noticed since install is a seemingly slow start time for Edge from icon to usability and the top bar hangs for a second and is white rather than dark as my theme for it is.

While in the process of typing this it happened again. The previous owner had used the card without issue.


From some other things I've read this card alone pulls around 290 watts of power. 63 watts while idle for the CPU and closer to 170 under load which puts me at close to 450 under load before drives, fans, etc. If this has a less than stellar 12v rail it's gonna **** itself before it can do much else, which would in theory explain the fans on the GPU going to the off position and I dare to say the CPU fan also turning off for a second or so earlier too.
 
Last edited:
I'm sure you're going to be buying a nice new Ryzen/Vega system in the future so buying a good PSU right now is either a cure or a start on the new system.

Did you use DDU to clean out the old and install the new drivers? My 290x is happy with 16.12.1 but I don't use wattman or ReLive. I do use Afterburner 4.3.0 Beta 14 and all is good.
 
Intuition tells me you may need a repaste of your card.
It sure helps, though I have never experienced artifacts in any of my cards...

The other thing I would suggest and a good vaccuming and reseating everything short of the CPU.

All PCI cards and memory should be reseated.

Good luck and keep us posted!
 
Run GPU-Z in sensor mode while putting load on the card. If the voltage fluctuates or the GPU clock micro-stutters, you have a power delivery problem.
 
Hardwood said:
I'm sure you're going to be buying a nice new Ryzen/Vega system in the future so buying a good PSU right now is either a cure or a start on the new system.

Did you use DDU to clean out the old and install the new drivers? My 290x is happy with 16.12.1 but I don't use wattman or ReLive. I do use Afterburner 4.3.0 Beta 14 and all is good.
Click to expand...


I didn't use DDU, I tried to do the best I could by hand but I'm a bit out of my Regedit days. I'll grab DDU and try again.



MasterGoa said:
Intuition tells me you may need a repaste of your card. It sure helps, though I have never experienced artifacts in any of my cards...

The other thing I would suggest and a good vaccuming and reseating everything short of the CPU.

All PCI cards and memory should be reseated.

Good luck and keep us posted!
Click to expand...


Funny enough I ended up doing the reseating already, not much to do. :D. I still need some canned air however to be rid of the last of the dust bunnies. I'll keep repasting in mind for after GPU-Z stuff.


GTwannabe said:
Run GPU-Z in sensor mode while putting load on the card. If the voltage fluctuates or the GPU clock micro-stutters, you have a power delivery problem.
Click to expand...

Thanks for the heads up on this, GPU-Z has come a long way from when I last used it.
 
So just fired up The Division, played for all of 2 minutes once in game before it gave me a BSOD. The error is Stop Code Video TDR Failure. What failed: atikmpag.sys

Before crashing it did give me a few minutes of log, what is it specifically I'm looking for? I found 6 lines that are odd in regards to VDDC in and out. It's above 100 drops well below 100 for 2 lines and is back up again.

In is: 164, 72, 56, 104, 129, 89, 47.
Out is 145.5 42.3, 90.3, 129.0, 127.4, 107.1, 42.2.

I figure the last 2 is when the BSOD hit.
 
Last edited:
I've gone ahead and reseated the few things in the machine, video card and 2 sticks of RAM, still running into issues under load. Load and idle temps are within reasonable boundaries so a repaste is likely unnecessary, plus it's cold here at this time of year.

The one other time I have had a bad card, it was a Radeon 64 original OEM with a bad TCL unit which only caused artifacts while gaming. This happens on the desktop.
 
After looking at better rated power supplies I went back to my current and dug through some crap. I found that it has a single 8 pin input for PCI-express stuff. As I'm no power supply expert, can a single 8 pin reasonably break off to an 8 and a 6 or did I just find my problem and it really is a power issue?
 
Normally, the 8pin will have 2 black leads on a seperate sliding part of the plug.
You have to push those that part forward, as it is held in place should you plug in the whole plug.

Also, I am a bit weary of canned air as it will push dust in crannys it normally would not go.

Good luck!
 
You are of course correct. My statement is one of I have a single 8 pin on the power supply which has to go to an 8 pin plus a 6 pin on the video card. I explained it poorly.

I've also been reading that this power supply doesn't do too hot once it hits about 30c, which it likely is in that case.
 
So just put in an order for an EVGA G2 850. It'll be here Friday, gonna start deconstructing the machine soon to have it ready for arrival.

Keep reading pretty mediocre things about this corsair, seems it's fine for midrange gear but taps out pretty quick once high end stuff is thrown at it.
 
Hopefully that was the issue and you didn't just change PSUs for nothing, but at least the EVGA G2 series is top notch. I don't use anything else for my builds or others anymore -- top of the line in quality.

There is a chance it's the card itself but we'll see. Did you buy this card used? It could have been one of those cards that were tortured mining, back when that was popular.
 
According to the previous owner it was gently used. It looks to be in good shape physically. Based off the fact that I think all of the fans in my machine would hiccup on occasion while the GPU was under load it really leaned toward a power supply issue. I wish I had done my homework before and maybe this wouldn't be a thing.
 
ASCI Blue said:
So just put in an order for an EVGA G2 850. It'll be here Friday, gonna start deconstructing the machine soon to have it ready for arrival.

Keep reading pretty mediocre things about this corsair, seems it's fine for midrange gear but taps out pretty quick once high end stuff is thrown at it.
Click to expand...

Good choice. Just put in an EVGA P2 1000 to handle my Radeon Pro Duo and could not be more pleased so far.
 
That's what I gather about EVGA. I saw a few posts about them here then started doing more homework using the likes of Johnny Guru and other places that really test ****. It's super annoying to find that my purchase is of fairly low quality but that can always be fixed.
 
Initial testing shows that the power supply was the issue. Got it installed, running each plug on the video card off a separate cable, adjusted crap in the drivers (no one wants a fake high resolution AMD :mad: ) and fired up The Division. I didn't play long but it loaded and allowed me to play with no screen greebles. So far so good.
 
Back to square one as of a couple hours ago. I tried to install 14.X drivers and all that did was seem to corrupt things, I had to use system restore to go back to 1:30 pm today.

At this point my last venue is to pop this card into a 2.0 slot.
 
Update: Disabled R7 at the bios, peroxided the contacts of the board, and put it in a different slot. Got a solid 15 mins out of Doom and an hour prior of various animated things. I also discovered the maintenance thing which is scary over today.

Have GPU-Z running and logging and other stuff up monitoring things.
 
You must log in or register to reply here.
Back
Top