Amd crossfire games list - games without profiles/negative scaling WTF AMD

The best decision I ever made in the graphics hardware department was moving away from my 6990 to sli 670's. Sure they did not work with every game, however the microstutter problems vanished, and sli dates back to much older games, not to mention sli profiles being extremely compatible with other games. Since then I have discovered what nvidia inspector can do with hbao+ profiles, I often replace a game's native ssao with them and gain 50% back that i can then spend on something like sparse grid supersampling.

AMD have since gotten their frametimes under control, however they still seem to lack "we care" driver timelines and backwards compatibility, not to mension the lack of driver level features.

I guess if i was the kind of person to not give a rat's arse about gsync, hbao+, driver level AA, and DSR, then AMD would be an affordable alternative. I've just can't imagine going back and losing all of this.

Not to mention Txaa and physx. For the games that support those features.

That's one hell of an ecosystem lost.

 

You'll be pleased to know the new drivers have a Shadow Of Mordor crossfire profile, before the competition too.

 

Thanks, the game looks great without the flickering!

I've noticed pretty good scaling at 1080p with 2x 280x cards (50% fps boost). However, there are still a few issues:

- Scaling boost drops to about 10% when there's rain on the screen.
- There's still some flickering when using in-game supersampling, possibly related to HDR exposure calculations or something. I've noticed it when panning the camera around quickly

 

Thanks i will report those issues. I expect the profile will be improved and updated as future drivers are released. What's important is we have a solid working profile out first in a game that is sponsored by a competitor.

 

pCARS

 

Thanks for letting us know. Great work.

 

Single card was giving me half the framerate in places where GPU load under CF was ~96% ~96%. Thought that would be pretty obvious. 47FPS vs high 90s.

 

Shadow of Mordor benchmark: 1 GPU: 52FPS 2 GPU: 80FPS (average)
14.9.2

 

Is this using the build in benchmark? I've found it overestimates the max fps (and thus the average) compared to fraps. What settings are you using?

 

Yes with the the benchmark.

Custom
High
Ultra
Off
Medium
Ultra
Ultra
High
Ultra
DOF disabled
On
On

Hope that makes sense

 

My scaling goes from 78 (1 GPU) to 89 (2 GPU) with those settings [With the TombRaider*.exe profile, I get 154 FPS (97% scaling!) ]

Could you try and run the same tests, using these settings:
Everything on / maxed except textures on high.

 

Am I reading that right: The improved pCars profile gives you an extra 5 fps in crossfire?

 

That's a good point! It's a pity about the scaling, but kudos for getting it up and running I only hope the reverse doesn't happen with Dragon Age... Is that expected to work on day 1?

 

1920x1200
Settings as you stated. 1GPU: 53FPS 2 GPU: 77FPS (scaling starts 90/90 and drops to 70/70 and then back to 90/90).
Stock clocks 960/1250

 

Just a follow up to those issues:

- Using the TR profile in 14.9.2, fps increases by 99% over single card, regardless of rain in game. GPU load is above 90%, except in situations where the 100 fps cap is hit.

- Using the default profile in 14.9.2, fps increases by about 60% when there's no rain, and just 10% when there is rain. When rain shows up, GPU load on both cards drops to about 65%.

So it seems (at least in my case, I'd love to know if anyone can reproduce this) that the default profile a bit slower when there's no rain, and considerably slower with rain.

* note: these measurements come from fraps, rather than the built in benchmark, which seems to be very unreliable *

 

Dragon Age will have Mantle support working on day 1 as Mantle has already been integrated into the Frostbite engine, so if any developer that uses that engine wishes to use Mantle, the work has already been done.

Will respond to other queries on Monday.

 

Thanks for the info, and thanks for responding so quickly on a weekend I guess the next challenge will be getting AC:Unity scaling nicely.

 

Umm, yes it is.

CPU bottleneck = when at least 1 CPU core is pegged at 100%
GPU bottleneck = when the GPU is pegged at 100%

Plus, when I raised my CPU speed by 75Mhz, it upped the max fps when not using vsync by about 15-20fps.

It also upped min fps by about 2-4.

So yeah, the CPU is definitely "holding it back".

Doesn't really matter much in Crysis though as I am always over 60fps anyway.. and I just run it with vsync when I am playing.

Running with vsync off is lame unless you have a game that has bad input lag with it on.

All it really does is make your system use more power and introduces screen tearing.

At least that is how I see it nowdays.

If you have a game that is so taxing that you can't maintain fps equal to the monitor refresh rate AND it doesn't work with triple buffering, then yeah, turn it off.

 

Yes ur reading it right..a bit sad isn't it?
Perhaps I'm cpu depended with 1920x1080 , will try eyefinity and see if that improves.

Another thing is that Thief isn't working with mantle crossfire, it doesn't even start.
U know anything about it AmdMatt?

 

Pirates of Black Cove Gold CF and Vsync broken

Pirates of Black Cove Gold is not working properly with Crossfire.

At the main menu, GPU 2 runs at 100%, while GPU 1 is at 0%.

Once in game, GPU 2 runs at 100%, and GPU 1 runs at 43%.

It also looks like Vsync is not working at all in this game, even when forced in CCC.

When Crossfire is turned off, GPU 1 runs at 100%, even with Vsync enabled.

So yeah, both Crossfire and Vsync are broken in this game.

This is with:
Windows 8.1 Pro
2x 7970 Ghz
Driver 14.9.2 Beta