Catalyst 15.7 and Wolfenstein:The Old Blood

[Carnage Rules!]

The first game i played(in fact i tried to play) after Catalyst 15.7 was Wolfenstein: The Old Blood. I'm using R9 270X and Windows 8.1. (All of the driver settings are default) I ran the game yesterday and i saw weird artifacts on the screen. Taking a screenshot or record the gameplay video is meaningless because the image that taken from the frame buffer is very clear so i'm gonna try to describe them(it's very hard for me): There are some blue or violet(or the same color of the area) boxes flickering on the screen and the number of boxes increase when i move, or turn left or right. When i stop the move the character, most of the artifacts disappear. After playing couple of minutes stopping the movement doesn't help. It seems like pure hardware error a lot. I quitted from the game and tried another game(Dirt:Rally) i didn't see any artifacts.

I started to think and i remembered that i encountered some problems with Wolfenstein: The New Order. Here's the thread:

http://forums.anandtech.com/showthre...450&highlight=

But they wasn't horrible as i saw yesterday.

I rolled back to Catalyst 14.12(I preferred to stay away from the betas) and i the problem solved. There's no graphical corruption(even common problems like z-fighting or texture flickering)

Here's my questions:

1/Was it a hardware error that some drivers can easily mask and some of them can't? I'm asking this because it seems like very hardware problem.

2/Has anyone seen the same or similar problem? Is my card the only one that troubles Wolf:Old Blood with Cat 15.7?

 

[Deleted member 13524]

There are some blue or violet(or the same color of the area) boxes flickering on the screen and the number of boxes increase when i move, or turn left or right.

It looks like the core is overheating or just not getting enough power.

Are you overclocking your card? Is the weather hotter than usual where you are now?

 

[Carnage Rules!]

I didn't overclocked anything. The temps are not too hot.

 

[Deleted member 13524]

It could be that your PSU is failing to provide enough voltage stability and current to the 12V rails.

 

[Carnage Rules!]

Maybe but the voltages seem OK. I have to measure with a voltmeter.

I ran Furmark and Video Memory Stress Tester. The results are fine. I tried another games. They don't have a major issue.

I thought that maybe old driver was masking a big hardware error and i assumed there's no problem. Or maybe new driver triggered another hardware issue(like core voltages) or changed card's working methods.

What are you suggesting? Why another graphic intensive applications don't have the same problem?

 

[Malo]

Do you have New Order to test it with as well? Might be a problem with the driver and the texture swapping it's doing on fast turns for megatexture.

 

[Carnage Rules!]

Do you have New Order to test it with as well? Might be a problem with the driver and the texture swapping it's doing on fast turns for megatexture.

I tested it months ago and it has minor artifacts with Catalyst 14.4 WHQL. My PC requires 14.6 Beta to run the game without artifacts. I only played first level, and i didn't test it enough time.

Disabling and enabling texture compression doesn't help. But enabling Adaptive V-SYNC and lowering all graphical settings together reduces the number of these glitches. If I only disable V-SYNC and enable adaptive v-sync(in game menu) or only lower the graphical quality doesn't work.

I made a small Google research and i found this:

 

[Carnage Rules!]

My artifacts are not the same as in this picture but this artifacts are very severe. Their problem is related with CUDA accelearation at the megatexture compression. Enabling or disabling virtual texture compression doesn't make a difference on me.

 

[Davros]

When I had a 6950 cats 14.12 I never had any problems

 

[homerdog]

Very interesting that screencaps and videos don't show it.

 

[Rurouni]

FWIW, I believe the newer driver push the hardware a bit more, thus potentially more problem when the hardware isn't running perfectly. For example, on 14 something driver, my Kaveri could overclock as high as 4.7GHz (stable @4.5) with the GPU @900MHz. I must steadily decrease it when moving to a newer driver. Right now my CPU is at stock (I think I can run it at least @4.3 stable) and GPU @800 (900 is impossible!). Of course I'm using an APU, but maybe because I'm using an APU the bad effect becomes more severe (some windows file corrupted and I need to re-install mediaportal).

 

[swaaye]

I've seen texture corruption on The Old Blood with my GTX 970 too. Seems likely to be a problem with the game itself. IIRC it started when I was trying to crank the detail settings. I ended up just using a master quality preset.

 

[Carnage Rules!]

Very interesting that screencaps and videos don't show it.

It's the reason why i concern. Screenshots from frame buffer are perfect. This is why the problem seems like a hardware issue. But on the other hand changing driver solves the problem. Maybe 14.12 was masking a big hardware fault on the card.

FWIW, I believe the newer driver push the hardware a bit more, thus potentially more problem when the hardware isn't running perfectly.

I suspect the same thing. But why more GPU intensive applications like FurMark and Video Memory Stress Tester don't have a problem. Maybe I have to push GPU and VRAM simultaneously. But I don't know which test do this. I don't even know "is stressing GPU and video memory at the same time possible?"

@swaaye: What kind of artifacts do you have?

 

[Rurouni]

If I based my OC on FurMark, I definitely can OC higher. Not only GPU, but CPU can overclock higher if it's only based on synthetic benchmark/stress test stability. Probably a dynamic workload can lead to more instability vs a static but maximum load.

 

[Albuquerque]

After adding my 980Ti, I've bumped into a stability problem with an occasional game CTD that has never happened prior to the video card upgrade. I've gone back to stock clocks on the CPU and memory, and so far, all appears to be back to normal. The kilowatt powersupply I'm using should be more than sufficient for the equipment I have, so I assume that the CPU load with the 980Ti is simply "different enough" that it has exposed a previously undiscovered-by-me stability challenge.

I'm guessing this is linked either to the different threading model in NV drivers, or else generally higher load on the CPU now that it isn't bottlenecked (as much) by the graphics card. Either way, perhaps not equivalent to your challenges, but worth considering that instability can crop up in weird places where it's been stable for eons past.

 

[Putas]

2/Has anyone seen the same or similar problem? Is my card the only one that troubles Wolf:Old Blood with Cat 15.7?

Same issue for me, Win7.1 and R9 270, it is still the same with 15.8 beta. New Order was fine.