• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

Batman: Arkham Origins PC Performance Thread

Yeah, with the beta 6 drivers I'm able to play the game with 8xMSAA (maxed everything without PhysX obviously) at over 60FPS usually. It still drops sometimes, but it looks incredible.
R9 280X Dual-X here btw.
 
Perfect 60 here:

i5-4670K
8GB RAM
GTX 670

Running perfect apart from the com tower bug hopefully they patch it soon.
 
Alright, I wanna get this straight once and for all.

I'm trying to eliminate the tearing in this game. I used to use d3doverrider to enable vsync and triple buffering in games, but I've heard it doesn't get updated and that it comes with MSI Afterburner? I've downloaded afterburner, and I am not seeing anything about enabling both. I've even gone into the RTSS and added the game, but not seeing it. Should I still just use d3doverrider by itself then? I've tried vsync in-game, in nvidia inspector, and the tearing is still there. My machine is:

3770k
8 GB RAM
Geforce 680 4GB

I saw someone mention eariler in the thread about taking off vsync and enabling triple buffering in d3doverrider, so that is why I am asking about whether the program by itself is still worth using, or has its functionality being put in something else.
 
AMD 13.11 beta 6

Batman: Arkham Origins - improves performance up to 35% with MSAA 8x enabled

I installed this and lost half of the features on my Catalyst Control Center and the game runs no better. What the fuck? This cutscene bullshit is the part I hate about PC gaming the most. When something doesn't fucking work when it should, there's never a sound or sensible solution. It's always just throw everything at the problem and ask for help in 100 different places until either the problem fixes itself or some trivial inconsequential thing finally works.
 
Try running the Benchmark for the game by doing the following...

Create a new shortcut, and replace the Target line with the following one (assuming you have the game installed in your G:\ HDD): If not then replace G with whatever drive it is installed on.

“G:\Batman Arkham Origins\SinglePlayer\Binaries\­Win32\BatmanOrigins.exe” benchmark

In the benchmark you can tweak your settings whilst it is running and monitor what kind of FPS you are getting to give you a better idea of how the game will perform with your setup.

That's not working for me...It says the target is not valid.
 
jlxb.png


Gtx 770
3570k @ 4.4ghz
latest drivers

Everything maxed TXAA on High and Physx on High
 
I've been getting pretty solid performance in most situations. Noticeable framerate dipping when I zoom in on stuff, but gameplay is generally fine. Then I met (minor spoiler)
Barbara
and for some reason the framerate went into single digits when the camera was on her. Weird stuff.

i7 2600
8 GB RAM
Radeon 6870
 
i5 4670k 4.3ghz
2 GTX 780 1.2ghz sli
2560x1440
All settings maxed out DX11/Physx/etc.


I get anywhere from 90 to 140 frames, usually in the 120 area

Games is really gorgeous, especially all the snow.
 
i5 2500K @ 4.3GHz
GTX 770 4GB
8GB DDR3
Win8.1

batbench5mqzu.jpg
 
One thing I have noticed, there is literally no blur with TXAA on this game, it is implemented very well, as a matter of fact, when I was taking screenshots with fraps to compare It with MSAA, TXAA actually looks better with less performance impact.

Could you post the screenshots? Last time I messed with TXAA it was too blurry for my liking.
 
One thing I have noticed, there is literally no blur with TXAA on this game, it is implemented very well, as a matter of fact, when I was taking screenshots with fraps to compare It with MSAA, TXAA actually looks better with less performance impact.


One thing I did notice is how it won't let me force transparency x8 Supersampling in NV control panel though.

Looks blurry to me: http://www.youtube.com/watch?v=E0GnVSyYDjQ

01:50
 
Getting constant 60 FPS without any dips, with everything max except FXAA which is set to High.

My rig:

HD 7970
i5 4670
8GB RAM
Win7

Using 13.9 drivers.


Edit: And of course PhysX is off.
 
Anyone seeing issues with the game stuttering a little during auto saves? finding it abit annoying especially since the game savea quite often. 7970 here
 
Is it just me or will PhysX always be garbage?

GTX680
3770K@4,2
16GB Ram

With normal Physx settings my fps dip down to 30-40 during the arena fight on the Final Offer. Even without enemies there are frame drops. Turning PhysX off gives me 60fps back.
Honestly not worth it for some laggy flags waving around...
 
That's what I thought when I first watched this video pre-release, I thought to myself "no chance I am using TXAA then"

But honestly I can not notice any difference whatsoever, going to post some images in a minute to show you what I mean.

theres a slight blurring, but honestly? At 2560x1440 it gives me near perfect IQ even with TXAA Low. Looks mindblowing, almost like downsampling from 4k without the performance hit.
 
Not kidding when I say this, I hated TXAA up until this game, compare TXAA high with 8x MSAA or 4x at least to me there is no difference whatsoever and you will get more frames.

Or is my eyesight deceiving me?

First thing I tried was TXAA set to high and thought it introduced blur. I'm super sensitive to post-AA though and in basically 100% of cases will turn it off, even if that means no AA and thus jaggies. I'm a stickler for the sharpest IQ possible (relative to performance) and post-AA always hinders it in a way that seems obvious to me.
 
First thing I tried was TXAA set to high and thought it introduced blur. I'm super sensitive to post-AA though and in basically 100% of cases will turn it off, even if that means no AA and thus jaggies. I'm a stickler for the sharpest IQ possible (relative to performance) and post-AA always hinders it in a way that seems obvious to me.

Bad Post AA. SMAA is pretty much blur-free.
 
I installed this and lost half of the features on my Catalyst Control Center and the game runs no better. What the fuck? This cutscene bullshit is the part I hate about PC gaming the most. When something doesn't fucking work when it should, there's never a sound or sensible solution. It's always just throw everything at the problem and ask for help in 100 different places until either the problem fixes itself or some trivial inconsequential thing finally works.

Did you have a beta installed before?

Make sure you fully uninstall the previous driver otherwise it won't actually update as the betas have the same driver numbers.
 
Yeah, anyway I can uninstall this and go back to how I had it before? I installed this and it made my performance worse and I lost half the features on my AMD control center.

i recommend unistalling the drivers with this app and reinstalling. remember these are beta drivers.

sorry for the late reply.
 
Intel Core-i5 2500K @ 4.5GHz
Sapphire Dual-X R9 280X 3GB (13.11 beta 6 drivers)
8GB DDR3 1333MHz

What do you NVIDIA users get without PhysX?

batmanarkhamoriginsbebzskr.jpg
 
the benchmark is useless, there's a part on the benchmark ive already played on the game and the ran like crap for me there, my frames dropped to 30 there, in the benchmark 60 all the way, or it might be that the game is not as optimized as we think it is
 
the benchmark is useless, there's a part on the benchmark ive already played on the game and the ran like crap for me there, my frames dropped to 30 there, in the benchmark 60 all the way, or it might be that the game is not as optimized as we think it is

This is the issue with all in-game benchmarks, looking at you Gta IV with your average 60fps.
 
ive been comparing FXAA low vs HIGH and i see no difference at all, anyone else seen this?

I see a difference. FXAA High has noticeably smoother edges. In fact, it has the least jaggies out of all available AA methods in the game.

Using a Radeon HD 7970
 
Minimum spec peasant here, except for my CPU which is clocked at 3.06Ghz

At 1280x800 in fake fullscreen mode (as I'm using a 1920x1200 monitor) with almost everything off the framerate varies between 30 and 60 fps depending on the situation. It's a little jarring but still playable. What's weird though is that the sound stutters from time to time, which might be the CPU's fault. Still runs better than bloody Witcher 3 though lol.
...Yeah I need a new machine.
 
So i had a new monitor bought from a friend sometime ago a Yamakasi 2560x1440 LED Screen and 60hz. But i noticed in games when i switched from my Acer GD245HQ 120hz monitor i had some dips in framerates. And now that i tested it with AO i saw that i went almost below 30fps..that to me is a no go.

So i switched back to my other monitor and this is what i like more.

batmanbenchmark46pat.png
batmanbenchmark2fhr92.png
 
i7 4700m, 780m 32gb of RAM
Everything on high or on except depth of field on Normal and FXAA high at 1080p. Everything run super smooth.
 
I've never been 100% clear on the differences between the AA methods. Which has the least affect on performance besides turning it off? I'm playing on a 60" Pioneer plasma and sitting ~8 from it so I probably can't tell the difference between them anyways.
 
I've never been 100% clear on the differences between the AA methods. Which has the least affect on performance besides turning it off? I'm playing on a 60" Pioneer plasma and sitting ~8 from it so I probably can't tell the difference between them anyways.

Post processing AA, which is an effect generally applied after a frame has been rendered. This is stuff like FXAA and TXAA. It looks at the rendered frame data and, using smoothing algorithms, attempts to clean up jagged edges on the image. Post processing AA has the smallest impact on performance.

MSAA/SSAA are AA implementations that are generally part of the rendering process itself. MSAA/SSAA are brute force methods of AA and, on average, can be very expensive. SSAA in particular is very expensive as it renders scenes at higher resolutions than what you're actually playing, and downsamples the data in an effort to clean up aliasing. MSAA is less expensive, but can still have a large impact on performance.

Quality is mixed across the board. Ideally the best form of AA simply in terms of image quality is SSAA. It's basically saying "let's render everything at fucking stupid resolutions so we have a lot more data, then shrink it down". Unsurprisingly, it's also expensive as fuck. MSAA is a "results are mixed" case. Often it works really well if at a cost, but the advent of modern rendering engines has caused a clash. MSAA doesn't always play nice with aliasing on shaders, and going into next gen we'll be seeing a lot more deferred rendering, which is very difficult to get working nicely with MSAA (see: Battlefield 3).

This is why a lot of engines are opting for post processing AA solutions, like FXAA, because it's generally inexpensive. Results are mixed, but a lot of people hate it, because the idea of applying AA as a post processing solution, after a scene has been rendered with jaggies intact, is a recipe for image blur as that's the laziest method. There's a lot of games out there where turning on FXAA is like smudging vaseline over the screen. The jaggies are gone, but the cost of sharp image quality is very high. But a lot of this comes down to the algorithms and methods used, and programmers are trying to find methods that implement both post-AA and render-AA together. TXAA High in Arkham Origins combines 8xMSAA with post processing for it's own thing. Metro: Last Light has a post-AA built into the engine that cannot be turned off, and looks great with SSAA sharpening the image. SMAA is also a pretty nice post AA that keeps the image sharp while trying to clean up jaggies.
 
i just noticed that my GPU usage on this game is only 59%, there's gotta be something wrong here because every other game gives me 100% usage

GTX 770 4GB
 
Post processing AA, which is an effect generally applied after a frame has been rendered. This is stuff like FXAA and TXAA. It looks at the rendered frame data and, using smoothing algorithms, attempts to clean up jagged edges on the image. Post processing AA has the smallest impact on performance.

MSAA/SSAA are AA implementations that are generally part of the rendering process itself. MSAA/SSAA are brute force methods of AA and, on average, can be very expensive. SSAA in particular is very expensive as it renders scenes at higher resolutions than what you're actually playing, and downsamples the data in an effort to clean up aliasing. MSAA is less expensive, but can still have a large impact on performance.

Quality is mixed across the board. Ideally the best form of AA simply in terms of image quality is SSAA. It's basically saying "let's render everything at fucking stupid resolutions so we have a lot more data, then shrink it down". Unsurprisingly, it's also expensive as fuck. MSAA is a "results are mixed" case. Often it works really well if at a cost, but the advent of modern rendering engines has caused a clash. MSAA doesn't always play nice with aliasing on shaders, and going into next gen we'll be seeing a lot more deferred rendering, which is very difficult to get working nicely with MSAA (see: Battlefield 3).

This is why a lot of engines are opting for post processing AA solutions, like FXAA, because it's generally inexpensive. Results are mixed, but a lot of people hate it, because the idea of applying AA as a post processing solution, after a scene has been rendered with jaggies intact, is a recipe for image blur as that's the laziest method. There's a lot of games out there where turning on FXAA is like smudging vaseline over the screen. The jaggies are gone, but the cost of sharp image quality is very high. But a lot of this comes down to the algorithms and methods used, and programmers are trying to find methods that implement both post-AA and render-AA together. TXAA High in Arkham Origins combines 8xMSAA with post processing for it's own thing. Metro: Last Light has a post-AA built into the engine that cannot be turned off, and looks great with SSAA sharpening the image. SMAA is also a pretty nice post AA that keeps the image sharp while trying to clean up jaggies.

Thanks for the informative post. ;)

So generally when a game has a binary AA option (ie on/off) it is most likely FXAA/TXAA correct?
 
Thanks for the informative post. ;)

So generally when a game has a binary AA option (ie on/off) it is most likely FXAA/TXAA correct?

In modern games, yeah. Default postAA choice seems to be FXAA as that's supported by both Nvidia/AMD. TXAA is an Nvidia exclusive method. SMAA can be injected using injectSMAA, and I find that using this with 2xMSAA gives decent results.
 
Top Bottom