Arma 2 Benchmarks

J

jonneymendoza

Supreme [H]ardness
Joined
Sep 11, 2004
Messages
6,395
i think its Time [H] did a benchmark with this game http://www.hardforum.com/showthread.php?p=1034253222#post1034253222

its the new crysis ie very demanding game :p

URL=http://s47.radikal.ru/i115/0906/44/6b7e247c7ba0.jpg]
b9749e0a66b0.jpg
[/URL]







http://i006.radikal.ru/0906/7d/f58b7eeaeb45.jpg









 
Wow ill have to pick this one up. It is beautiful.
signature_gtx285.jpg

Ive heard that you can acctually see your body. like if you look down you can see your arms and legs and such.

that sounds kinda interesting.

hope my 9800gx2 can run it no problem.
 
overclockers said:
Ive heard that you can acctually see your body. like if you look down you can see your arms and legs and such.

that sounds kinda interesting.
Click to expand...

Yeah, but that's nothing new. It was the same way in Operation Flashpoint and Armed Assault. In several other games, too, as a matter of fact (L4D, Chronicles of Riddick, etc.) :cool:
 
i used the ARMAII Mark mission (see my link above) to test a 4870 512MB and a GTX285 1GB on my system. I used ARMAII version 1.01

Settings:
Resolution - 1680x1050
Fillrate - 100%
Texture Detail - Normal
Anisotropic Filtering - Normal
Terrain Detail - Normal
Objects Detail - Normal
Shadow Detail - Normal
PostProcess Effects- Low

(view distance does not matter, mission automatically sets this)
(forced vsync off - did not have an affect on performance)

Specs:
Q6600 (3.0GHZ)
4870 512MB (775/1050)
GTX285 1GB (694/1584/1296)
4GB DDR2 800
Vista 64bit
Catalyst 9.5
186.18 WHQL

this mission runs slow on the first time through due to loading textures, etc. i ran it 3 times and took the best score.

4870 512MB score: 3049 (avg fps 30.49)


GTX285 1GB score: 3002 (avg fps 30.02)


it seems i'm CPU limited with my Q6600 at 3.0GHZ.
 
Last edited:
Trying to download this benchmark. The 1 MB file appears to just include a mission file. I guess I am missing something here, do we need Arma 2 to run the benchmark?
 
Porter_ said:
(view distance does not matter, mission automatically sets this)
....
it seems i'm CPU limited with my Q6600 at 3.0GHZ.
Click to expand...

Either that, or the bit you glossed over in the beginning. In some games, a different view distance can double or triple performance. If the game is automatically adjusting the view distance to keep a set....oh....30fps, then the fact that you get the same thing with both cards just indicates the game does a very good job of dynamically setting it.
 
hhara said:
Trying to download this benchmark. The 1 MB file appears to just include a mission file. I guess I am missing something here, do we need Arma 2 to run the benchmark?
Click to expand...

installation instructions from the download page:
Installation:
Place the included .pbo file in your missions folder and select it from the SP missions screen.
Click to expand...
 
dderidex said:
Either that, or the bit you glossed over in the beginning. In some games, a different view distance can double or triple performance. If the game is automatically adjusting the view distance to keep a set....oh....30fps, then the fact that you get the same thing with both cards just indicates the game does a very good job of dynamically setting it.
Click to expand...

well if it did that it wouldn't be much of a benchmark tool. the view distances are fixed (but vary from test to test). there's an entire thread over on BIS forums where people are posting their results. it would be a thread full of nothing but 30fps posts if the test dynamically changed view distance to hold a fixed....30fps. you are right though, this game is one where view distance has a dramatic impact on performance.
 
sorry i didn't see your question. yeah you need ARMA2 to run the benchmark. not sure if it will work with the demo or not.
 
Porter_ said:
it seems i'm CPU limited with my Q6600 at 3.0GHZ.
Click to expand...

Perhaps if you increased some detail settings, the 285 would stay at 30fps while the 4870 would not. i.e. Perhaps you're CPU limited because your detail settings are too low. Surely some of the settings are not related to CPU speed in any major way?
 
thanks for the suggestion GHarris. i can increase some settings and not take a hit on performance. i can also max out the settings and take a relatively marginal hit, but still poor performance overall. i can drop the settings to their absolute lowest and get the same performance as medium settings. the problem is the ceiling i'm hitting is *just* a bit too low for my tastes.
 
GHarris said:
Perhaps if you increased some detail settings, the 285 would stay at 30fps while the 4870 would not. i.e. Perhaps you're CPU limited because your detail settings are too low. Surely some of the settings are not related to CPU speed in any major way?
Click to expand...

How could a Q6600 @ 3.0 be limited so soon?

:rolleyes:

It's a bad benchmark. Look at the thread it game from. Similar systems varying greatly by processor more than video card.
 
I made some benchmarkings with the rig at the bottom
I was not being super precise or anything. Its not science etc.

All settings MAX with 200 % Fillrate optimizer and Visibility set at 10,000
2560x1600: 10 fps (frequently below)
1920x1200: 13 fps
1680x1050: 14 fps

All settings HIGH with 100 % Fillrate optimizer and Visibility set at 2,000
2560x1600: 24 fps
1920x1200: 26 fps
1680x1050: 27 fps

All settings NORMAL with 100 % Fillrate optimizer and Visibility set at 2,000
2560x1600: 30 fps
1920x1200: 34 fps
1680x1050: 35 fps

Edit: as I have discussed with people on other forums I am most likely CPU-bound. It is well established now that ArmA likes quad cores, i7 and hefty overclocks.

Gaming rig
-Intel Core2 Quad CPU @ 2.83GHz
-8192MB ram
-ATI 4870x2
-Dell 30'' 3008WFP (2560x1600)
-TrackIR 5 Pro
-Saitek X52 Pro
-Microsoft Sidewinder X8 mouse
-Xbox 360 Controller for Windows
-Vista Home Premium 64bit
 
The fillrate( odd form of AA ) and terrain detail options seem to be the fps killers. Unfortunately they're also the two options that make the biggest amount of difference in quality.
 
shansoft said:
There is no way it will have a CPU bottleneck problem for a Q6600 @ 3.2G - 3.6G...

i7 is not a option for gamers, there is no advantage at all.... unless the game itself is bad coded.....
Click to expand...

whether it's coded poorly or not is difficult to determine, but the game engine apparently responds well to more cores. the more cores you have the better off you are. the i7 (with 4 cores, 8 threads) is indeed the best cpu for this game. my Q6600 @ 3.2GHz is holding me back, but i haven't had problems in any other games.

who knows, maybe it's a bug that will eventually be patched out.
 
Atech said:
Multi-threading is not a bug, it's a feature.
Click to expand...

my Q6600 never breaks 50% cpu usage, which indicates the game only utilizes two threads. i guess what i'm trying to say is i hope they eventually find a way to better utilize multi-core processors and patch it into the game...if such a thing can be patched into the game.
 
overclockers said:
Wow ill have to pick this one up. It is beautiful.
signature_gtx285.jpg

Ive heard that you can acctually see your body. like if you look down you can see your arms and legs and such.

that sounds kinda interesting.

hope my 9800gx2 can run it no problem.
Click to expand...

crysis can do that.
 
Porter_ said:
my Q6600 never breaks 50% cpu usage, which indicates the game only utilizes two threads. i guess what i'm trying to say is i hope they eventually find a way to better utilize multi-core processors and patch it into the game...if such a thing can be patched into the game.
Click to expand...

Read the article again, more cores are better than MHz, no one can promise that all threads will cap the CPU at 100%...stop moving the goalposts.
 
Porter_ said:
i used the ARMAII Mark mission (see my link above) to test a 4870 512MB and a GTX285 1GB on my system. I used ARMAII version 1.01

Settings:
Resolution - 1680x1050
Fillrate - 100%
Texture Detail - Normal
Anisotropic Filtering - Normal
Terrain Detail - Normal
Objects Detail - Normal
Shadow Detail - Normal
PostProcess Effects- Low

(view distance does not matter, mission automatically sets this)
(forced vsync off - did not have an affect on performance)

Specs:
Q6600 (3.0GHZ)
4870 512MB (775/1050)
GTX285 1GB (694/1584/1296)
4GB DDR2 800
Vista 64bit
Catalyst 9.5
186.18 WHQL

this mission runs slow on the first time through due to loading textures, etc. i ran it 3 times and took the best score.

4870 512MB score: 3049 (avg fps 30.49)


GTX285 1GB score: 3002 (avg fps 30.02)


it seems i'm CPU limited with my Q6600 at 3.0GHZ.
Click to expand...

UPDATE: exact same settings, but switched from Vista 64 to Windows 7 64

GTX285 1GB score: 3761 (avg fps 37.61)


i've only tested the GTX285 so far, but it looks like Windwos 7 gives a huge performance increase over Vista 64. the bump is even more substantial during the SP campaign. on one single player mission during a pretty intense part of the mission, my framerate went from 27 to 37. framerate is up across the board.

to anyone getting poor performance in Vista 64, i recommend downloading the Windows 7 RC. at least after you've exhausted all posibilities in Vista.
 
Porter_ said:
to anyone getting poor performance in Vista 64, i recommend downloading the Windows 7 RC. at least after you've exhausted all posibilities in Vista.
Click to expand...

This should come as no suprise. Vista has always sucked.

Ghostbusters runs a ton faster for me in W7 too.
 
AMDXPCottonFire22 said:
How could a Q6600 @ 3.0 be limited so soon?

:rolleyes:

It's a bad benchmark. Look at the thread it game from. Similar systems varying greatly by processor more than video card.
Click to expand...

Soon? 3.0 Ghz per core has been a bottleneck for along time. 3.4+ is where you want to be to remove the CPU bottleneck. I prefer 4.0Ghz with my 9550 but thats just me ;)
 
Mmh, the graphics are decent, but they have not enticed me. The gameplay looks pretty yawn-tastic. Those spectacular night firefight videos they have demonstrate the gist of the concept: sneak forward on your belly for ten minutes, admiring the battlefield. Get shot from an unknown angle and location. Die. Wait to respawn. Repeat. Most games of this type have the same problem. :(
 
Soggy_Popcorn said:
Mmh, the graphics are decent, but they have not enticed me. The gameplay looks pretty yawn-tastic. Those spectacular night firefight videos they have demonstrate the gist of the concept: sneak forward on your belly for ten minutes, admiring the battlefield. Get shot from an unknown angle and location. Die. Wait to respawn. Repeat. Most games of this type have the same problem. :(
Click to expand...

These games are not trying to be COD or Gears of War. Thats their whole point....
 
whats fun about this game is not the action. its the fun of USING your brain cell to try and stay alive. i was in a coop game with two people and we had to attack this village yet we stood and crawled and shoot for about 30mins with hardly any action occurring. we thought it was fun and intense!
 
You must log in or register to reply here.
Back
Top