ArmAII-Mark

[Ez3kiel 10]

Finally, I updated the spreadsheet with everyone's results, can I upload it somewhere so that anyone that's interested can review it?

Protegimus

Yea, plz. I did it yesterday morning but there is a lot of new one now.

To upload files, i used filedropper.

Wow you did a big job thanx

I used your link below in "proper" way

http://www.filedropper.com/scorearma2mark_2

Edited June 7, 2009 by Ez3kiel
add link

[Protegimus 0]

Yea, plz. I did it yesterday morning but there is a lot of new one now.

To upload files, i used filedropper.

Thanks Ez3kiel, spreadsheet was really useful.

You can download revised version at:

http://www.filedropper.com/scorearma2mark_2

My ArmA II Mark at 1680x1050 resolution:

4627.87 49.5940 49.7088 40.4595 57.0885 34.5425

all tests as directed by binkster.

Edited June 7, 2009 by Protegimus
Corrected URL

[Supernova 0]

Who cares about this synthetic benchmark. It means nothing really and is more of an e-peen competition rather.

[binkster 0]

Who cares about this synthetic benchmark. It means nothing really and is more of an e-peen competition rather.

It actually helps out a lot. Because of this I have found tweaks plus not to mention arma doesn't like vista on my system. I would have never tried xp without getting an idea on other users.

[jumpinghubert 49]

new bench with normal setting

thanks alot binkster and protegimus for detailed information.

XP and Arma2 on different harddrives.

My shortcut: "-maxmem=3072 -mod=@VopSound_2.0 -nosplash -cpucount=4

Edited June 7, 2009 by JumpingHubert

[Rista 0]

Who cares about this synthetic benchmark. It means nothing really and is more of an e-peen competition rather.

Easy for you to say it means nothing with a machine like that ;) While it may not be an accurate representation of in game performance, it's a lot more useful than saying "the game runs fine" or "the game runs crap on my machine". If it wasn't for this thread, I probably would have already upgraded to a card which wouldn't give me the kind of performance that I was expecting. So I'd say it's quite useful actually.

[Ez3kiel 10]

+1

And when you watched all score, best rig seems to have the best score but only for intel cpu (seems too low for amd cpu)

Edited June 7, 2009 by Ez3kiel

[Supernova 0]

Have any fellow i7 users tried turning off HT and seeing if that improves your score or ingame fps ?

[von_paulus 0]

+1

And when you watched all score, best rig seems to have the best score but only for intel cpu (seems too low for amd cpu)

Maybe Intel processsors performs better than AMD cpu's in this game.

[Supernova 0]

Maybe Intel processsors performs better than AMD cpu's in this game.

Or perhaps this benchmark is biased towards Intel processors.

[Cionara 10]

Hey ;-)

Got 5704 Points now with performing a 2nd run after the 1st.

1920*1080

Fillrate: 100%

Textures:Normal

Vram: Very High

AF: Normal

TerraiN: Normal

Objectdetail:Normal

Shadows: Normal

Postprocessing: Low

Q9550 @ 3,71Ghz

GTX 280 OC

4GB DDR2-Ram

X-fi Xtreme Gamer

[giantsfan24 10]

I think this proves how biased this test it. This score is better than I7 scores.

Hey ;-)

Got 5704 Points now with performing a 2nd run after the 1st.

1920*1080

Fillrate: 100%

Textures:Normal

Vram: Very High

AF: Normal

TerraiN: Normal

Objectdetail:Normal

Shadows: Normal

Postprocessing: Low

Q9550 @ 3,71Ghz

GTX 280 OC

4GB DDR2-Ram

X-fi Xtreme Gamer

[maddogx 13]

Or perhaps this benchmark is biased towards Intel processors.

That was probably the funniest thing I've read in a long while. How would an in-game benchmark be biased toward Intel?

[Cionara 10]

After applying sound-hardware acceleration and setting my rendersize to the size of my resolution in the config I now get 5842 points :D that's awesome, maybe I'll break the 6000 ^.^ next

I think this proves how biased this test it. This score is better than I7 scores.

Don't forget that my cpu is overcloked to 3,71 Ghz (normally 2,83Ghz).

Also my GTX280 is.

[Desrat 0]

Score 3530 :( most of the settings on low though...

C2D E6600 stock clock of 2.4ghz

4GB 667mhz DDR2

MSI P45 Chipset mainboard

Nvidia 8800GTS 320mb

Windows7 x64 Beta (build 7000)

Think its time I pushed my cpu to 3Ghz at least and get that Nvidia 275 I have my eyes on...

Edited June 7, 2009 by Desrat

[Supernova 0]

That was probably the funniest thing I've read in a long while. How would an in-game benchmark be biased toward Intel?

Perhaps because in most of the reviews I have read the Phenom II X4 955 has been able to keep up with the i7 920 in most if not all gaming tests. Then we see contrary results posted by this biased synthetic benchmark.

Don't forget that my cpu is overcloked to 3,71 Ghz (normally 2,83Ghz).

Also my GTX280 is.

This benchmark doesn't accurately system overclocks.

[Rista 0]

How is it a biased synthetic benchmark though? While it's far from perfect, it's an actual in game benchmark that gives a fairly good measure of real-world performance on a given system. Overclocking my E6750 showed the exact same improvements in both the benchmark and actual in game performance -it gave me a fairly large FPS increase in certain situations and slight increase in FPS overall. I don't really see how could a benchmark like this be biased towards Intel unless the whole game is "biased" or in other words, works better on Intel CPUs.

[Supernova 0]

How is it a biased synthetic benchmark though? While it's far from perfect, it's an actual in game benchmark that gives a fairly good measure of real-world performance on a given system. Overclocking my E6750 showed the exact same improvements in both the benchmark and actual in game performance -it gave me a fairly large FPS increase in certain situations and slight increase in FPS overall. I don't really see how could a benchmark like this be biased towards Intel unless the whole game is "biased" or in other words, works better on Intel CPUs.

First off CPU benchmarks are always usually run at lower resolutions than the resolutions suggested by OP. In some cases they are run at 1024x768 or even lower. This on the other hand is suggested to be run at resolutions higher than that which means it is GPU limited. There are about five tests in this synthetic benchmark but from what I have seen all the tests run in this thread favor an Intel processor more than an AMD processor. How can I see this well if an stock clocked i7 920 user scores much higher than a stock clocked AMD Phenom II X4 955 then there is obviously something wrong. Think about it how can there be such a discreptancy between Intel and AMD processors in the benchmark. Until I'm proven otherwise wrong I call this benchmark flawed and not a real indicator of real word performance.

[Rista 0]

First off CPU benchmarks are always usually run at lower resolutions than the resolutions suggested by OP. In some cases they are run at 1024x768 or even lower. This on the other hand is suggested to be run at resolutions higher than that which means it is GPU limited. There are about five tests in this synthetic benchmark but from what I have seen all the tests run in this thread favor an Intel processor more than an AMD processor. How can I see this well if an stock clocked i7 920 user scores much higher than a stock clocked AMD Phenom II X4 955 then there is obviously something wrong. Think about it how can there be such a discreptancy between Intel and AMD processors in the benchmark. Until I'm proven otherwise wrong I call this benchmark flawed and not a real indicator of real word performance.

I see what you're saying about CPU benchmarks being run at lower resolutions but this is more like an overall system benchmark that isn't supposed to be 100% accurate. Why do Intel processors perform better I don't know but they do achieve an actual higher FPS not just the overall score. The benchmark does show that that particular scene runs better at certain CPU/GPU/OS. If we were to change the benchmark mission to different scenes perhaps the results would be different but if Intel CPUs, ATI cards or whatever give better results in most tests I'd say it's safe to say the game runs better on those CPUs/GPUs.

[binkster 0]

I can easliy put together a cpu benchmark mission. Pretty simple just put like 1000 units with wp and an enemy to fight against.. But the problem is that the ai would act different in every time u would run it if there were enemy. I would say 1000 units walking/running to a wp with choppers dropping off soldiers and planes flying around could stress the cpu. We all know that FPS will drop when the cpu is overloaded so we could use same scripts to determine an average fps.

After looking at Protegimus spreadsheet that he made with everyones scores the i7's and the high quads cores q9550/q9650's are at the top in scores. I also noticed all the high scores people are using xp and all the the latest single nvidia cards. I wouldnt say that this benchmark is useless. Gives a good idea on what runs good and what runs bad based on that spreadsheet.

Best way that I used to open the file was with google docs online.

Edit: Maybe Protegimus should only put peoples scores that hotlinked an image of the score pluss settings. Im sure some folks might not have been truthfull. Plus some of these people are not following the standard on settings.

Edited June 8, 2009 by binkster

[o22cool 10]

Upgraded my CPU, motherboard & RAM. Score went from 1500 to 4700! Old system was a AMD Opteron 165 @2.6ghz aka X2 3800+, 4GB DDR1. I ordered the 2nd best AMD config, but I canceled it at the last min because looking at the results here told me to go Intel, and it was only $200 more. I figure I will be overclocking it to a $1000 processor spec so it's worth it. When I got my AMD Opteron 165, that was the best bang for the buck 4 years ago, it was a $350 processor, but I overclocked it to a FX60 a $1000 processor, I hope this one last me another 4years. I was going to drop another $150 for another 4850 so I can crossfire it, but seeing that ARMA II doesn't play well with multi-GPU setups, I decided to wait for the DX11 cards coming out this fall.

Texture Detail - Normal

Anisotropic Filtering - Normal

Terrain Detail - Normal

Objects Detail - Normal

Shadow Detail - Normal

PostProcess Effects- Low

Cpu - i7 920 OC (3.6ghz)

Ram - OCZ DDR3-1600 6GB

GPU - ATI 4850

OS - Vista

Resolution - 1680 x 1050

Normal Score - 4700

[maddogx 13]

@Supernova

Dude, no offense but I think you don't have a clue what you're talking about. You obviously have some kind of pre-formed opinion about benchmarks, probably because of programs like 3DMark/PC-Mark, so you're judging ArmA2Mark the same way. Let me put this clearly:

ArmA2Mark isn't some external program that measures CPU performance to generate a contrived benchmark number. It's an actual in-game performance test that tests actual in-game performance in different situations. If you try it with a varying number of CPU cores you can actually see the framerate change (with Fraps), and the ArmAII Mark score reflects that.

Just to make this clear: This is not a synthetic benchmark!

And your allegations of bias toward Intel are laughable. If anything, maybe the game performs better on Intel CPUs for whatever reason, but the benchmark has absolutely no way of preferring any architecture over another. It's just a scripted mission.

[FraG_AU 10]

Texture Detail - Very High

Anisotropic Filtering - Normal

Terrain Detail - Normal

Objects Detail - High

Shadow Detail - High

PostProcess Effects- Low

Fill rate @ 100%

Cpu - i7 920 OC (4.2ghz)

Ram - Patriot DDR3-1600 6GB

GPU - GTX295 @ 650/1400/1100

OS - Vista x64

Resolution - 1920x 1200

Normal Score - 4115

PS> My GTX295 is really a GTX275 as SLi is not working at all. I have run tests with 2x gpu on and just one, and also ran tests with different render modes. Obviosuly there is a lot of untapped performance here, to give you an idea my GPU temp after 30 min of play never went over 55C, wheras playing other games hits 70+

Also CPU @ 4.2ghz was barley used, so PLENTY of room for optimisations.

If anyone has managed to get SLi working please tell me :), I should have a 50+ FPS once this gets fixed up, lol may just go back to ARMA/ACE and enjoy the FPS there while the bugs are sorted here :)

[Supernova 0]

Texture Detail - Very High

Anisotropic Filtering - Normal

Terrain Detail - Normal

Objects Detail - High

Shadow Detail - High

PostProcess Effects- Low

Fill rate @ 100%

Cpu - i7 920 OC (4.2ghz)

Ram - Patriot DDR3-1600 6GB

GPU - GTX295 @ 650/1400/1100

OS - Vista x64

Resolution - 1920x 1200

Normal Score - 4115

PS> My GTX295 is really a GTX275 as SLi is not working at all. I have run tests with 2x gpu on and just one, and also ran tests with different render modes. Obviosuly there is a lot of untapped performance here, to give you an idea my GPU temp after 30 min of play never went over 55C, wheras playing other games hits 70+

Also CPU @ 4.2ghz was barley used, so PLENTY of room for optimisations.

If anyone has managed to get SLi working please tell me :), I should have a 50+ FPS once this gets fixed up, lol may just go back to ARMA/ACE and enjoy the FPS there while the bugs are sorted here :)

You should know that a GTX 295 consists of two GTX 260's not the 275.

[scottw 0]

Try renaming arma2.exe to crysis.exe, this got SLI working for me.