Problem: artifacts, dots and weird behavior

[GhostiNv]

Hello, i've got an old pc: Q9550, 4gb corsair ram, asus p5kr mobo, gts 250 video card, a sandisk ssd and a Sirtek highpower bronze II 600w psu running Windows 10.
A few days ago all of the sudden without any change Diablo 3 started crashing or just freezing. This happened before with other video card so i thought it was nothing special, but yesterday my pc started acting weird. Started wow and bam artifacts, weird textures, even the windows start bar was flashing. I restarted the pc and all seemed ok till after ~1h in diablo weird dots appeared and started moving around the screen. Again i restarted the pc and installed newer drivers (though the ones i had were 1 month old only) and since then it works without any problems. I did an artifact scan, full 3dmark test, furkmark for ~30 min to stress out the video card but nothing. This was yesterday.
Today when i started my pc i didn't have video signal in windows. Restarted the pc got to pin screen entered pin, no signal. Restarted again artifacts appeared on the pin screen but after entering the pin got on to the desktop. Everything was flickering, dots and other artifacts everywhere, the driver crashed and recovered and then all i saw was my desktop made of green dots, i could distinguish the battle.net launcher. I restarted and now again it works perfectly. No artifacts, no problems in games/furmark.
Is it the video card dying out or can it be the psu ?

this is how diablo/desktop looked with dots:

 

[MongGrel]

Looks like your GPU is just old to me.

Me wife even just web browses facebook and streams soap operas with my old Q9650 gaming rig from years back.

Just is how it works.

Least you have a SSD, I'd upgrade the GPU.

 

[EndlessWargasm]

Could be the gpu dying, could be the PSU.

Do you have any way to access a spare PSU to test with?

Fire up GPU-Z and monitor temps.

Open up computer and clean it out if dirty/dusty, special attention on the video card.

Try underclocking or raising voltage on GPU core and or vram if possible.

 

[GhostiNv]

I do have another psu to test with but right now it works perfect and doesn't do any problems in any tests even on high temperatures. Cleaned the pc yesterday, the temperatures are not a problem.
Will try with the other psu though not sure of the results. Need to find out which is the culprit since a psu is cheaper, because if it's the gpu i'd go for a gtx 970 even if my other components would be a bottleneck, but will have it for the new pc when i'll have the money to get one.

 

[MongGrel]

Yeah, I'd say checking the PSU out would be step one.

 

[tweakboy]

Take memory to stock settings. This is a vram issue I believe. gl

 

[GhostiNv]

The psu is not to blame since it does artifacts and stuff on other psu. Somebody suggested that it could be from the pci-e slot and so i trying wiggling it a bit and it did produce artifacts and crashed furmark (this was done when i just started the pc, so it didn't have time to "warm up"). So i cleaned both the pci-e slot and the video card connector and tried again. Did manage to produce some artifacts but had to wiggle it good. Now by this development i would say it's the slot but it did have time to warm up so to speak and if it is the slot then why did it produce artifacts when i started the pc today and stopped after "warming up" without me touching it at all ?

 

[DaveSimmons]

I really hope you are not "wiggling" your card while the system is powered on.

That is a good way to fry the card, the motherboard, the CPU and possibly you

 

[GhostiNv]

Well yes with it on. I know it's not the brightest idea i ever had but since i had no results by just moving it a bit starting the pc, turning it off, moving in other way turning on again... etc. Now i still don't know if that's actually the problem since as i said i didn't even touch it this morning when it misbehaved and. Will see what happens tomorrow morning.

 

[DaveSimmons]

Maybe it's thirsty and wants to drink a refreshing glass of water while it's powered on? *



* don't do that, it's a very bad idea, even worse that shorting out the card by pushing it around with the power on.

 

[xorbe]

My money is on a toasted GPU. Or motherboard slot.

 

[RussianSensation]

Hello, i've got an old pc: Q9550, 4gb corsair ram, asus p5kr mobo, gts 250 video card, a sandisk ssd and a Sirtek highpower bronze II 600w psu running Windows 10.

It sounds like you have a GPU issue of some sorts. GTS 250 is a rebadged 9800 GTX+. A lot of GPUs of that era, specifically GeForce 8 and 9 were prone to solder fissures that would develop over time. If your card starts exhibiting these symptoms again, make sure to run Unigine Valley or 3DMark, set your GPU fan speed to 100% and make sure that the GPU temperatures are below 95*C max in GPU-Z. If thermal throttling/GPU core temperatures are not the cause of the problem, it's likely your GPU is failing.

As a last resort, you can attempt the oven baking technique. If you primarily play Diablo 3, any $100-110 replacement card such as GTX750Ti 2GB or R9 270/270X is more than good enough. Q9550 would be a major bottleneck for a GTX970.

 

[MongGrel]

It sounds like you have a GPU issue of some sorts. GTS 250 is a rebadged 9800 GTX+. A lot of GPUs of that era, specifically GeForce 8 and 9 were prone to solder fissures that would develop over time. If your card starts exhibiting these symptoms again, make sure to run Unigine Valley or 3DMark, set your GPU fan speed to 100% and make sure that the GPU temperatures are below 95*C max in GPU-Z. If thermal throttling/GPU core temperatures are not the cause of the problem, it's likely your GPU is failing.

As a last resort, you can attempt the oven baking technique. If you primarily play Diablo 3, any $100-110 replacement card such as GTX750Ti 2GB or R9 270/270X is more than good enough. Q9550 would be a major bottleneck for a GTX970.

I read the first part and baking it came to mind.

Then you went there

YMMV.

 

[cholley]

is the oven method effective on an asus GTX560?

 

[GhostiNv]

I writing this with wow, diablo 3, hots opened on full background performance and furmark running on max resolution + 8xMSAA, video card is at 98% usage, video memory at 95%, temp at 73 degrees but no artifacts or problems whatsoever. So now i'm guessing it was just the pci-e slot or video card connector that was dirty.
Tbh i'm a bit disappointed, since if it was the video card i would've had a good reason to get a new one.
Thank you very much for your time and help.

 

[soccerballtux]

wiggling it could be flexing the PCB on the GPU, or it could be PCI-e slot, hard to say.

this is really peculiar though I've never heard of htis.

 

[sm625]

wiggling it could be flexing the PCB on the GPU, or it could be PCI-e slot, hard to say.

this is really peculiar though I've never heard of htis.

It's funny that the very first graphics card I ever bought had artifacts. And 2 out of the next 3 had artifacts as well. I just assumed it was a normal thing. Only one ended up being so serious I had to replace it. The AGP slot was notorious for this sort of problem too. PCIe, not so much.

 

[soccerballtux]

lol. I never had that problem. were you overclocking the AGP bus?

 

[GhostiNv]

I'm gonna do a last test tomorrow since i got 2 pci-e slots. No idea how it will work since the 2nd pci-e slot is 4x. That pci-e slot was stressed a bit since first video card was a 8800GT, then when it died went for this gts 250, then GTX285 and then that one burned (not acutal flames but smell of burned plastic and the pc doesn't even start with that card in) and now back to this gts 250.

 

[xorbe]

is the oven method effective on an asus GTX560?

Many 560 cards were simply under-volted from the factory. You can use MSI AB to either slow the gpu down a bit, or bump the voltage up a notch or two, which resolves many 560 cards. Baking wouldn't fix this type of problem.

 

[Gikaseixas]

It's not the PSU cuz I've been through a similar issue and changed PSU's just to end up having the same issues. It turned out my card was the problem no matter what.

I'm afraid you will have to invest on a GF soon

 

[GhostiNv]

So i've tested today with the other pci-e slot, moving the card just a bit would trigger artifacts or crash furmark...so i've ordered the gtx 970.

 

[Blue_Max]

I just had those dots in question - turns out my DVI cable was a little loose.

Easiest (and free) solution, make sure it's plugged in there snugly. Wiggle a little too and see if it's a bad connection.

 

[GhostiNv]

Not the case here, if it were just the cable not being plugged tight it wouldn't crash furmark or the whole pc.

 

[tweakboy]

80 percent chance its the graphic cards VRAM. gl