Mafia 2 Demo Benchmark

[v8envy]

8800GTX is actually better than the gt240.

Not just that but the original PPU from Agea is no longer supported. And even if it was supported it has processing power somewhere between an 8600GT and GT240/9600GT, the 8800GTX is waaay better.

 

[Wag]

I didn't find any difference uninstalling PhysX and running the version in the Mafia folder. I think I need a better GPU to run at 2560x1440.

 

[SHAQ]

I did the same last night. When I first loaded it with high PhysX the game ran at 10-15 fps. I then uninstalled and reinstalled PhysX and it stayed above 60 fps most of the time. However the benchmark still ran at 30 fps.

 

[Homerboy]

Ugh
Im guessing I'm going to have to upgrade my 8800GT to play this game (running with Phenom X4 @3.5Ghz @16x10 resolution).

 

[toyota]

Ugh
Im guessing I'm going to have to upgrade my 8800GT to play this game (running with Phenom X4 @3.5Ghz @16x10 resolution).

why don't you actually play the demo???

 

[smackababy]

why don't you actually play the demo???

That would be far to hard?

 

[Homerboy]

why don't you actually play the demo???

Uhh did I say I wasnt going to? I'm taking an educated guess off what people here are posting for their benchmarks. I'm at work, so I can't exactly play the demo RIGHT NOW... but thanks for your feedback.

 

[toyota]

Uhh did I say I wasnt going to? I'm taking an educated guess off what people here are posting for their benchmarks. I'm at work, so I can't exactly play the demo RIGHT NOW... but thanks for your feedback.

did you say that your were? lol. anyway I think you will be able to enjoy it with that card at just 1680.

 

[Grooveriding]

So has anyone tested whether using a dedicated card is better?

480 SLI not dedicated

2560x1600

PhysX off

Min 4.8
Max 250
Avg 62

Physx Medium

Min 6.3
Max 166.7
Avg 43.6

PhysX High

Min 4.6
Max 200
Avg 41.4


480 and 480 dedicated to PhysX

Medium

Min 4
Max 142.9
Avg 34.6

High

Min 13.7
Max 142.9
Avg 34.4


For me, it's better not using a dedicated. I'm sure if I threw in a third card just for physx it would improve, but eff that. This should be enough. I have an 8800gts 512mb lying around or a 285 from my gf's pc I could use but I'm too lazy for that.

Performance seems poor to me, especially for a console port, the graphics are using really low textures and AA does not seem to be working even when set to on, hate to see the fps once it works.

 

[evolucion8]

1280x1024 with everything maxed, Vsync off, High Quality Audio, No APEX PhysX

Min: 29.8
Max: 166.7
Ave: 77.4

Crossfire seems to scale well with the game, the game launched and crashed a couple of times until I reinstalled PhysX, seems that's the culprit.

 

[toyota]

1280x1024 with everything maxed, Vsync off, High Quality Audio, No APEX PhysX

Min: 29.8
Max: 166.7
Ave: 77.4

Crossfire seems to scale well with the game, the game launched and crashed a couple of times until I reinstalled PhysX, seems that's the culprit.

so are we supposed to guess your hardware being used? lol. and surely you aren't stuck at 1280x1024 with a modern gaming pc?

 

[Wag]

480 SLI not dedicated

2560x1600

PhysX off

Min 4.8
Max 250
Avg 62

Physx Medium

Min 6.3
Max 166.7
Avg 43.6

PhysX High

Min 4.6
Max 200
Avg 41.4


480 and 480 dedicated to PhysX

Medium

Min 4
Max 142.9
Avg 34.6

High

Min 13.7
Max 142.9
Avg 34.4


For me, it's better not using a dedicated. I'm sure if I threw in a third card just for physx it would improve, but eff that. This should be enough. I have an 8800gts 512mb lying around or a 285 from my gf's pc I could use but I'm too lazy for that.

Performance seems poor to me, especially for a console port, the graphics are using really low textures and AA does not seem to be working even when set to on, hate to see the fps once it works.

I always thought SLI always trumped using a card as a dedicated GPU? The game looks pretty good to me, especially with AO on.

 

[evolucion8]

so are we supposed to guess your hardware being used? lol. and surely you aren't stuck at 1280x1024 with a modern gaming pc?

Well, I prefer a P-MVA monitor at 1280x1024 than a 1920x1080 TN monitor with sucky colors, terrible angle views and horrible blacks and contrasts, and why do you care about it? I have a 42" LG LCD TV in the living room but I'm too lazy to move my PC there to test a simple game with blurry textures and questionable jaggies. Just worry about your hardware, I will take care of mine

Pst: I doubt that you can get such frame rate even at 320x240, you will be too CPU limited.

 

[happy medium]

1280x1024 with everything maxed, Vsync off, High Quality Audio, No APEX PhysX

Min: 29.8
Max: 166.7
Ave: 77.4

Crossfire seems to scale well with the game, the game launched and crashed a couple of times until I reinstalled PhysX, seems that's the culprit.

1280x1024 all settings high no physx gave me a B rating

6.5
142
63 avg

Now we all know the minimums and maximums are off, but with a 77.4 avg vs my 63 avg. What kind of crossfire are you running? I'm only running 1 5750.

 

[toyota]

Well, I prefer a P-MVA monitor at 1280x1024 than a 1920x1080 TN monitor with sucky colors, terrible angle views and horrible blacks and contrasts, and why do you care about it? I have a 42" LG LCD TV in the living room but I'm too lazy to move my PC there to test a simple game with blurry textures and questionable jaggies. Just worry about your hardware, I will take care of mine

Pst: I doubt that you can get such frame rate even at 320x240, you will be too CPU limited.

whats the point in listing your results if you are not going to list your hardware???

 

[happy medium]

whats the point in listing your results if you are not going to list your hardware???

BY the looks of his benchmarks, he must have 2 4770's or 3870's or a slower dual core cpu.

 

[evolucion8]

BY the looks of his benchmarks, he must have 2 4770's or 3870's or a slower dual core cpu.

I have an HD 4870 1GB in CF, and a Q9650 at 3.82GHz.

This are my results using different Crossfire profiles

Single GPU:

Min: 19.2
Max: 142.9
Ave: 59.2

Crossfire default:

Min: 29.2
Max: 166.7
Ave: 77.2

Crossfire SFR:

Didn't work, horrible artifacts which gave the game an ASCII art look, quite funny...

Crossfire AFR:

Min: 18.5
Max: 166.7
Ave: 77.9

Crossfire Super Tiling:

Min: 18.2
Max: 125.0
Ave: 52.1

Shows that the game is CPU bound at my resolution, increasing the resolution might reveal more stuff, when I feel like want to move my 45 pound computer to the living room, will do the benchmarks at 1920x1080.

Happy Medium, seems that my single GPU score is far much higher than your HD 5750 score, specially in minimums, in average the difference is quite small in your favor.

 

[happy medium]

I have an HD 4870 1GB in CF, and a Q9650 at 3.82GHz.

This are my results using different Crossfire profiles

Single GPU:

Min: 19.2
Max: 142.9
Ave: 59.2

Crossfire default:

Min: 29.2
Max: 166.7
Ave: 77.2

Crossfire SFR:

Didn't work, horrible artifacts which made the game an ASCII art style, quite funny...

Crossfire AFR:

Min: 18.5
Max: 166.7
Ave: 77.9

Crossfire Super Tiling:

Min: 18.2
Max: 125.0
Ave: 52.1

Shows that the game is CPU bound at my resolution, increasing the resolution might reveal more stuff, when I feel like want to move my 45 pound computer to the living room, will do the benchmarks at 1920x1080.

Happy Medium, seems that my single GPU score is far much higher than your HD 5750 score, specially in minimums, in average the difference is quite small in your favor.

Are you using fraps? Because min and max with this benchmark seems screwed up but the avg's seem right.

 

[happy medium]

here is what fraps shows me when I record 60 sec @1600x1200. My minimums are not bad. I don't know where its getting 6 fps minimum? Mabe in the very, very beginning?
FPS
61 59 59 65 65 65 65 67 44 61 58 60 58 52 53 50 59 60 57 58 55 43 47 52 57 56 45 52 48 52 54 53 50 52 60 61 65 73 75 59 56 63 58 59 46 49 50 49 50 52 53 53 54 51 54 54
56 55 44 41

The fire scenes kill my fps.

evolucion8,
The minimums and maximums are messed up. The avg seems right in line with fraps.

 

[evolucion8]

Are you using fraps? Because min and max with this benchmark seems screwed up but the avg's seem right.

Nope, the game comes with a built in benchmark which gives you the results at the end of the run ala Far Cry 2 style. Now that I digged deeper in those tests, seems that Crossfire has some scaling issues, or the game is too CPU bound.

 

[happy medium]

Nope, the game comes with a built in benchmark which gives you the results at the end of the run ala Far Cry 2 style. Now that I digged deeper in those tests, seems that Crossfire has some scaling issues, or the game is too CPU bound.

Thats what i'm saying the built in benchmark sucks.
The avg is right but the min and max are off.

Yea I agree,
Your crossfire must be not working correctly in this game, your avg should not be less then mine.

 

[Grooveriding]

I always thought SLI always trumped using a card as a dedicated GPU? The game looks pretty good to me, especially with AO on.

Eh, not sure, I've never paid much attention to physx. I'm just curious about performance and the strangely high recommended hardware to use physx in this game.

That and the claims this game would have amazing physx implementations. From the demo I didn't see anything I didn't see in Batman AA though.

The game looks decent, and I'm excited to play it, I loved the first one, and as far as gameplay, it's a lot of fun and looks to be a great game. But in strict graphics terms, this title looks about as good as Bioshock, but currently performs like a pig on 480 SLI for such weak visuals :thumbsdown:

Maybe a driver release will improve it. I know it's on console as well, not sure if it is a straight console port, if so, that could explain the poor performance as well.

For context I saw someone post a bench of 75fps avg at 1920x1200 with 5870CF, no physx of course, my system got about 100fps at that resolution without physx. Just seems low for a game that does not make use of anything demanding. No huge view distances, very low res textures, aa is not working etc.

 

[kami]

Here are my results. See system specs in my signature. One of my GPUs is set to handle the PhysX.

This is with the game set to the "High" preset and 1920x1080. The only thing I changed was turning off vsync.

No Physx:

Min 4.6
Max 250
Average 114

Medium Physx:

Min 4.6
Max 200
Average 52.7

High Physx:

Min 4.6
Max 200
Average 50.6


I also noticed something weird. When Physx was enabled both of my GPUs would stay around only 50-60 percent usage, but when Physx was off they were 90-100 percent (looking at afterburner charts). Why are the GPUs being so under-utilized when Physx is on?


I also tried dedicating one of my GTX 460's to Physx and these are the results on HIGH:

High Physx (1 dedicated GPU):

Min 9.1
Max 125
Average 48.6

Performance is similar but slightly lower to not using a dedicated GPU. During this test GPU1 was at nearly 100 percent (the one doing the graphics) and GPU2 which was doing Physx stayed around 50-60 percent.

It seems like I could get a lot more physx performance if the game would only fully utilize the GPUs when PhysX is enabled. And yes I uninstalled PhysX and re-installed it with the installer found in the mafia 2 demo folder.

 

[happy medium]

Mabe crossfire and sli are broken in this game?

 

[thilanliyan]

I don't think SLI is broken there because I get 30fps avg. with a single GTX 460 at 1920x1200 (all maxed with ambient occlusion) so those averages with GTX 460 SLI are about right aren't they?