QuadCore VS Dual, Multiple OS's Test. Post Your Results Please

[-=seany=- 5]

I wanted to share some info regarding some testing I have done. I have luckily been able to test Arma2 on 2 different PCs; one with Vista64, Win7 64 and XP64 the other with just XP64. I have a very basic mission (link at bottom) that I use to gauge CPU performance, it's just the Player as a rifle man, standing in the middle of Chenogorsk. I also run FRAPS. Note that both the GTS250 and 9800GTX cards are pretty much identical performance wise. All tests used 190.38 nvidia drivers.

I use these settings on both PCs:

Arma2 GFX Settings

View Distance :1000

Resolution :1152*864

Render Res :1152*864

Textures :Normal

Texture Mem :Normal

Tex Filtering :Very High

AntiAliasing :Low

Terrain Detail :Normal

Object Detail :Normal

Shadows :High

Post Process :Very High

PC1 Specs (Win7-64, Vista-64SP1, XP-64SP2)

------------------------------------------

Corsair Dominator TWIN2X PC8500 4GB DDR2

2*XFX GeForce 9800GTX 740M 512MB sli

Core2Quad9300 @ 3.19ghz (OC)

XFX780i Mother Board

PC1 QuadCore Vista64/Win7 64: 30FPS

http://s72.photobucket.com/albums/i192/seanysherry/?action=view&current=QuadCoreVista.jpg

Performance was Identical between Vista and Windows7

PC1 QuadCore XP64: 38FPS

http://s72.photobucket.com/albums/i192/seanysherry/?action=view&current=QuadCoreXP64.jpg

PC2 Specs (XP-64SP2)

------------------------------------------

Crucial Ballistix PC8500 4GB DDR2

Gainward GTS250 512mb

Core2Duo E7300 @2.66 (Default)

Asus P5Q SE MotherBoard

PC2 DualCore XP64: 38FPS

http://s72.photobucket.com/albums/i192/seanysherry/?action=view&current=DualCoreXP64.jpg

As you can see, the DualCore at a lower clock speed and with only one GFX card has no problem beating the SLI, over clocked, Quad Core PC on Windows Vista and Windows 7. When the QuadCore is put to use with XP64 it matches the Dual core. Still, I would have expected it to beat the DualCore by quite a few FPS considering it has twice the Cores and a much higher clockspeed. I can't explain it.

In a less CPU intensive area, like a forest, the QuadCore SLI pc gets about 5 to 10 fps better than the single GPU, Dual core. Pretty much what you would expect and is not really of any use due to that fact that any way you play the game, online and offline, real game use is always CPU limited, unless the mission is tiny with few units etc.

Just to verify that this is most definitely a problem with Arma2 and QuadCores / CPU usage in general, here are the 3DMark06 scores for each Operating system on both PCs:

3DMark06 Score:

Core2Quad Vista-64 :16046

Core2Quad Win7-64 :16430

Core2Quad WinXP-64 :17395

Core2Duo WinXP-64 :11945

Exactly what you would expect, so what's the problem with Arma2, why can my 3.2ghz Quad Core only match the same FPS as a 2.66ghz Dual Core when using the Exact same GFX settings and OS in CPU limited areas? I hope it will be sorted soon, or at all.

Please add to these results:

I have included the mission "CityCPUTest" if anyone would like to share their results. Unzip and put it in your "Documents/Arma 2 Other Profiles/YOUR PROFILE NAME/missions/" folder. Load FRAPS (http://www.fraps.com/download.php), then go into the editor, load the mission and press "Preview". Don't move the mouse or anything, just note down your FPS when it steadies out after a second or two. For the test results to mean anything for comparison sakes, temporarily switch to the same Video settings I have above.

Post your FPS and your specs, it could be interesting to compare:

http://rapidshare.com/files/268814816/CityCPUTestChernarus.zip

Edited August 18, 2009 by -=seany=-

[joethe33 10]

I think you're a bit confused on how quad cores work. Refer to here. I don't feel like re-explaining it.

post #13

http://forums.bistudio.com/showthread.php?t=84028&page=2

[ICE-Raver 10]

Sorry man, I can't get your mission to even show up.

I'm curious though. Why not just use the Arma mark benchmark? I've already ran tons of vista 64 vs windows 7 tests on my system using arma mark. Results are on page 55.

http://forums.bistudio.com/showthread.php?t=73610

PC specs are in my sig.

I score just over 27,000 3dmark 06, and just over 26,000 on 3d mark vantage with my rig.

Edited August 18, 2009 by ICE-Raver

[randir14 10]

I think you're a bit confused on how quad cores work. Refer to here. I don't feel like re-explaining it.

post #13

http://forums.bistudio.com/showthread.php?t=84028&page=2

Thanks that was interesting. I was thinking of upgrading to a quad since everyone says this game is so CPU demanding (right now I have a 3.2 ghz dual core) but I'll hold off on it. I guess the only game that really needs a quad core to run perfectly is GTA IV.

[joethe33 10]

Thanks that was interesting. I was thinking of upgrading to a quad since everyone says this game is so CPU demanding (right now I have a 3.2 ghz dual core) but I'll hold off on it. I guess the only game that really needs a quad core to run perfectly is GTA IV.

My dual on 3.9 Ghz Air cooled runs it perfectly. =]

Now... 41 FPS while shits ''going down'' may not be technically perfect, but considering it's a shitty port and the coding is likely fucked to hell. It's playable. =p

[ModeZt 0]

Add some 100-200 bots and you will see the difference between C2D and C2Q. When nothing is going on in the world around you the game is limited by the gfx card.. But this game requires lots of CPU power to manage AI's.

[SWAT_BigBear 0]

Add some 100-200 bots and you will see the difference between C2D and C2Q. When nothing is going on in the world around you the game is limited by the gfx card.. But this game requires lots of CPU power to manage AI's.

Affirmative,

I disabled cpu #2 on my pc, ArmA 2 definitely runs better with 4 cores when you have a large # of AI's running around.

[-=seany=- 5]

I think you're a bit confused on how quad cores work. Refer to here. I don't feel like re-explaining it.

post #13

http://forums.bistudio.com/showthrea...t=84028&page=2

You say yourself a couple of times in that post about the main difference being price, that is not the issue here. I understand that a higher clocked, cheaper, dual core will be better than a lower clocked chip with 4 cores. In my test the dual core is at 2.6ghz while the quad is at 3.2ghz and there is no difference in performance between the two, which is strange. Especially when 3Dmark06 confirms that there is a very big performance difference between both pcs.

About the number of cores being utilised, the fact is that this game supposedly supports multicore when in fact it only supports Dual core, why not add Quad support too? Your point about gamers being better off with Dual Core chips is only really relevant when the application has no Multicore support at all. If an app supports multicore the Quad should shine.

Sorry man, I can't get your mission to even show up.

I'm curious though. Why not just use the Arma mark benchmark? I've already ran tons of vista 64 vs windows 7 tests on my system using arma mark. Results are on page 55.

http://forums.bistudio.com/showthread.php?t=73610

PC specs are in my sig.

I score just over 27,000 3dmark 06, and just over 26,000 on 3d mark vantage with my rig.

I don't use Arma mark because it is too inconsistent for proper benchmarks imo, the score always varies wildly for me. A simple mission like this is better because there are absolutely no variables other than your hardware, OS and gfx settings. Also the results are immediate and reproducible every time.

Don't know why you cant get the mission to show up :confused: . You unzipped it right? and your loading it through the editor? Its a shame you cant get it to work, that's a nice 3Dmark06 score :). I'd be interested to see what FPS it can get.

Add some 100-200 bots and you will see the difference between C2D and C2Q. When nothing is going on in the world around you the game is limited by the gfx card.. But this game requires lots of CPU power to manage AI's.

That's an interesting point and worth testing, although the area I'm testing in (Chenogorsk) is a CPU limiting area. I can tell this because decreasing the gfx settings like AA or shadows do not increase the FPS, that and my SLI indicators are very low. SLI Indicators basically show how hard the GFX cards are working, if you are unfamiliar with SLI.

Thanks the replies and suggestions :cheers:

[ICE-Raver 10]

Alright, I got it to work. I was trying to put the file only in the missions folder. Apparently you have to put the entire folder in it.

Here is my results using the specs you posted:

View Distance :1000

Resolution :1152*864

Render Res :1152*864

Textures :Normal

Texture Mem :Normal

Tex Filtering :Very High

AntiAliasing :Low

Terrain Detail :Normal

Object Detail :Normal

Shadows :High

Post Process :Very High

However that resolution is not native for my monitor so I took another screen @ 1680 x 1050 which is my native resolution if you care to see the results.

Edited August 19, 2009 by ICE-Raver

[joethe33 10]

You say yourself a couple of times in that post about the main difference being price, that is not the issue here. I understand that a higher clocked, cheaper, dual core will be better than a lower clocked chip with 4 cores. In my test the dual core is at 2.6ghz while the quad is at 3.2ghz and there is no difference in performance between the two, which is strange. Especially when 3Dmark06 confirms that there is a very big performance difference between both pcs.

About the number of cores being utilised, the fact is that this game supposedly supports multicore when in fact it only supports Dual core, why not add Quad support too? Your point about gamers being better off with Dual Core chips is only really relevant when the application has no Multicore support at all. If an app supports multicore the Quad should shine.

I don't use Arma mark because it is too inconsistent for proper benchmarks imo, the score always varies wildly for me. A simple mission like this is better because there are absolutely no variables other than your hardware, OS and gfx settings. Also the results are immediate and reproducible every time.

Don't know why you cant get the mission to show up :confused: . You unzipped it right? and your loading it through the editor? Its a shame you cant get it to work, that's a nice 3Dmark06 score :). I'd be interested to see what FPS it can get.

That's an interesting point and worth testing, although the area I'm testing in (Chenogorsk) is a CPU limiting area. I can tell this because decreasing the gfx settings like AA or shadows do not increase the FPS, that and my SLI indicators are very low. SLI Indicators basically show how hard the GFX cards are working, if you are unfamiliar with SLI.

Thanks the replies and suggestions :cheers:

I pointed out price as in that case because he needed new hardware. So it's mostly for a builder. ARMA 2 is coded poorly. It runs poorly regardless. You should always keep that in mind. My main point is that games are benefited the most from clock speed, NOT additional cores. ARMA 2 included.

When it comes to a builder, your budget matters. If you want the more expensive Quad. Go for it, though I'll say it again. Unless you do heavy video editing, you do NOT need it.

edit: One thing you might have missed in my point was that, if you had to choose between a Dual Core or a Quad Core at the same price range, the Dual would be more suitable for gaming. Generally they sell a Quad Core with the same arch as the Dual, but it's generally more expensive.

Edited August 19, 2009 by joethe33

[MJK-Ranger 0]

Hi.

First picture is from your settings:

View Distance :1000

Resolution :1152*864

Render Res :1152*864

Textures :Normal

Texture Mem :Normal

Tex Filtering :Very High

AntiAliasing :Low

Terrain Detail :Normal

Object Detail :Normal

Shadows :High

Post Process :Very High

34 PFS Click image

Next picture, same settings, just chanced the Resolution.

View Distance :1000

Resolution :1680*1050

Render Res :1680*1050

Textures :Normal

Texture Mem :Normal

Tex Filtering :Very High

AntiAliasing :Low

Terrain Detail :Normal

Object Detail :Normal

Shadows :High

Post Process :Very High

31 PFS Click image

Last picture, with my settings

Visibility: 1600

Screen Resolution: 1680x1050

3d resolution: 100%

Texture detail: Very High

Video Memory: Default

Anisotropic Filtering: Very High

Anti-aliasing: Normal

Terrain Detail: Normal

Object Detail: Normal

Shadow Detail: Disable

Post Process Effects: Disable

33 PFS Click image

[=wfl= sgt bilko 10]

A note on non-SLI vs SLI:

I ran a few tests before and after applying v1.03 (they're in the arma2bench thread somewhere).

My conclusion was that in v1.03 with my rig and my settings the non-SLI setup beat SLI in all subtests except the first one "UAZ Race at 2500m viewdist".

In v1.02 SLI beat non-SLI but with extreme fluctuation between test runs.

See post further below

Edited August 20, 2009 by =WFL= Sgt Bilko

[Fireball 16]

A note on non-SLI vs SLI:

I ran a few tests before and after applying v1.03 (they're in the arma2bench thread somewhere).

My conclusion was that in v1.03 with my rig and my settings the non-SLI setup beat SLI in all subtests except the first one "UAZ Race at 2500m viewdist".

In v1.02 SLI beat non-SLI but with extreme fluctuation between test runs.

Did you consider the latest SLI profile? http://www.nvidia.com/object/nvidia_sli_profile_patch.html

I always hear people bitching about lack of SLI support in ArmA II but never see anywhere that they mentioned installing the latest SLI profile - only the latest drivers.

[=wfl= sgt bilko 10]

Yes, but ...

That profile was released before 190.38 drivers.

The only stuff in that patch is an update to nvapps.xml file.

The changes for ArmA2 in the patch is identical to what you get with latest drivers.

[-=seany=- 5]

Alright, I got it to work. I was trying to put the file only in the missions folder. Apparently you have to put the entire folder in it.

Here is my results using the specs you posted:

View Distance :1000

Resolution :1152*864

Render Res :1152*864

Textures :Normal

Texture Mem :Normal

Tex Filtering :Very High

AntiAliasing :Low

Terrain Detail :Normal

Object Detail :Normal

Shadows :High

Post Process :Very High

However that resolution is not native for my monitor so I took another screen @ 1680 x 1050 which is my native resolution if you care to see the results.

Woah, I'm green with envy :D. That is impressive, I see now why you say you have no probs with Arma2 performance wise. If you play online do you get high FPS or do you find that your limited by how many FPS the server can get? Like, have you ever got 40 or 50 fps online? It would seem you need a rather large clock speed (like yours) to get a better frame rate than I can get. Still disappointing that my extra 500mhz overclock doesn't give me any performance boost :/ .

I pointed out price as in that case because he needed new hardware. So it's mostly for a builder. ARMA 2 is coded poorly. It runs poorly regardless. You should always keep that in mind. My main point is that games are benefited the most from clock speed, NOT additional cores. ARMA 2 included.

When it comes to a builder, your budget matters. If you want the more expensive Quad. Go for it, though I'll say it again. Unless you do heavy video editing, you do NOT need it.

edit: One thing you might have missed in my point was that, if you had to choose between a Dual Core or a Quad Core at the same price range, the Dual would be more suitable for gaming. Generally they sell a Quad Core with the same arch as the Dual, but it's generally more expensive.

Understood, I hear where your coming from.

Hi.

First picture is from your settings:

View Distance :1000

Resolution :1152*864

Render Res :1152*864

Textures :Normal

Texture Mem :Normal

Tex Filtering :Very High

AntiAliasing :Low

Terrain Detail :Normal

Object Detail :Normal

Shadows :High

Post Process :Very High

34 PFS Click image

Next picture, same settings, just chanced the Resolution.

View Distance :1000

Resolution :1680*1050

Render Res :1680*1050

Textures :Normal

Texture Mem :Normal

Tex Filtering :Very High

AntiAliasing :Low

Terrain Detail :Normal

Object Detail :Normal

Shadows :High

Post Process :Very High

31 PFS Click image

Last picture, with my settings

Visibility: 1600

Screen Resolution: 1680x1050

3d resolution: 100%

Texture detail: Very High

Video Memory: Default

Anisotropic Filtering: Very High

Anti-aliasing: Normal

Terrain Detail: Normal

Object Detail: Normal

Shadow Detail: Disable

Post Process Effects: Disable

33 PFS Click image

Pretty much identical to what I was getting on Vista and Windows7, perhaps marginally better. Consider dual booting with XP64 for an additional free 10fps. Thanks for adding to the thread.

A note on non-SLI vs SLI:

I ran a few tests before and after applying v1.03 (they're in the arma2bench thread somewhere).

My conclusion was that in v1.03 with my rig and my settings the non-SLI setup beat SLI in all subtests except the first one "UAZ Race at 2500m viewdist".

In v1.02 SLI beat non-SLI but with extreme fluctuation between test runs.

Interesting about SLI, I was considering turning it off to see if it's usage cost's cpu cycles. In my experience SLi does work very well in Arma2, but in reality it never gets a chance to do it's thing because the CPU is always the limiting factor. So perhaps disabling it for Arma might give another FPS or two, who knows? I'll have to test! :)

Not to get off topic, but about the SLI profiles. The fix by EVGA, Nvidia and the 190.38 drivers are the same. And all of these are basically the same as renaming Arma2.exe to Crysis.exe.

Anyways, If any one with a quad core Q6600 or Q9xxxx could do the test that would be cool.

Cheers all :cheers:

[Game__On 10]

I can't believe this "Is sli working?" is still an issue.

It's working as good as it possibly can in this game actually. Turning off SLI is downright stupid.

UNLESS you are cpu limited in a horrible way that is.

Here's some pictures i took to demonstrate.

First one, typical scenario, in a town with some buildings and some trees.

SLI is working nicely, altough still a bit CPU limited as the scaling is not near about 75% .

Second, a clear case of cpu limitation. Not much going on in this screen. One building, lots of blue sky, and some grass and crap.

Definately cpu limited as SLI scaling is not max by a long shot. Still + 30 frames mind you compared to no sli .

Third one, close to a tree , much of screen filled with foliage. Video cards definately get stressed here . Still decent frames though thanks to both cards scaling quite well. Without SLI you can almost cut those frames in HALF. Which is not very pleasant to play with.

Last one, standing 3 feet from a tree looking into it. Video cards are getting their asses handed to them :butbut: . Still almost decent frames because SLI is scaling a very near 100% . Both cards are being used at fullest.

So unless your cpu is from the stone age, turning off SLI is definately not an option.

As for the multiple cores vs 1 core mentioned earlier in this thread, that may be true if you go back in time 2 years , but todays programs and games can definately utilize more then one core.

I would not even try to play arma 2 on a single core, not even if it were running at 6 ghz

[=wfl= sgt bilko 10]

Well, you're opinion is as good as mine.

I clearly stated that my conclusions was with my rig/my settings and nothing else, saying that non-SLI is "downright stupid" is being ignorant to the context.

Furthermore this game clearly benefits more from higher CPU speed than anything else.

You're OC'ing about 170% compared to my rig (4.1 vs 2.4), no wonder SLI might actually be able to show result.

You're hardware and mine are way to apart to be compared. I'm sure SLI works fine on your rig but seeing that many in here have about the same hardware as I have, it would be "downright stupid" to not and out my test results.

Edited August 20, 2009 by =WFL= Sgt Bilko

[Game__On 10]

Saying 'downright stupid', actually was downright stupid of me. I did not mean it in a bad way but i worded it very silly.

sry for that.

I did mention 'unless you are cpu limited' , which i think is going on for a lot of people here indeed.

My point was (and maybe i misunderstood your post) is only that SLI does work , and very good as well.

But you are right though, your cpu can't be the bottleneck, or else sli might actually hurt performance.

But again, for that to happen you must be severy cpu bottlenecked.

edit* You haven't included your pc specs in the post where you said non SLI beat SLI . What specs are you running ?

Edited August 20, 2009 by Game__On

[-=seany=- 5]

Damn, Game__On. You did all that posting and testing and never even did my test! Come on man show us the goods. :) Take 15 minutes and set your Video Settings to the same as in the first post and grab my mission so we can take a look, cheers ;).

I'm not seeing your specs also? You clearly can afford to run SLI. Do you get those FPS online? Far from the stone age, but most us with socket 775 CPU's at about 3.0ghz do seem to be CPU limited enough under normal gameplay to make using SLI not really worth it. However I still have to test it, although it is pretty obvious that I'm CPU limited online (coop) and in bigger towns and cities. My Sli indicators hit the deck.

Anyway discussing SLI is ok, but main thing here is Clock speed and cpu comparison.

[=wfl= sgt bilko 10]

@Game_On

Re-reading my own post above gives that I edited it to pieces, making little sense. So the confusion really started there.

======

Here we go again:

Visibilty - 2400
Texture Detail - Normal
Video Mem - Very High
Anisotropic Filtering - Normal
Anti Aliasing - Normal
Terrain Detail - Low
Objects Detail - Normal
Shadow Detail - High
PostProcess Effects- Disabled
res: 1280x1024 (100%)

OS - WinXP 32
RAM - 4GB
processor - Intel Q6600 2.4ghz @ 2.4ghz
video card - 2x nVidia Geforce 9600 GT 1GB
Game on separate drive

Test 1.02 (non-SLI vs SLI) and 1.03 (non-SLI vs SLI):

http://forums.bistudio.com/showpost.php?p=1399447&postcount=531

Some comments on overclocking 25% (2.4 vs 3.0):

http://forums.bistudio.com/showpost.php?p=1400637&postcount=534

Sum up:

- In 1.02 (no OC) non-SLI marginally better than SLI in all tests except test1 "UAZ Race vis 2500", where SLI outperforms non-SLI

- In 1.03 (no OC) non-SLI is slightly better than SLI in all tests except test1 "UAZ Race vis 2500", where non-SLI outperforms SLI

- In 1.03 (25% OC) non-SLI outperforms SLI in all tests except test1 "UAZ Race vis 2500", where SLI outperforms non-SLI giving SLI a better boost in total.

Makes you wonder what's going on with that first test?

V-Sync "application controlled" since I was mainly interested to see what gain I'd get in scenes normally below 60 fps.

Edited August 20, 2009 by =WFL= Sgt Bilko

[Karhis 10]

@Game__On,

What is that software you are using to display SLI utilization/scaling? Is it some Nvidia specific?

[nyran125 10]

i did this test years ago,lol. When Crysis came out. And the intel core 2 version of the E6700 2.66Ghz ran better than the quad core version of the same clock speeds. All i can say is if people are going to upgrade FFS do yourself a favour, dont listen to your friends (that includes intel websites, nividia websites, dont listen to them) and do your research on Frames per second at all the sites first (like anandtech or Toms hardware or SOMEWHERE that does honest benchmarking, or MOSTLY honest benchmarking) and get an informed decision on your upgrade , not just the graphic cards, check the processor fps benchmarks also seperately. Just because youve listened to Intel sell a load of rubbish to you about the fastest quad core bla bla and Microsoft about thier new friggin Oporatin system. Doesnt exactly meen its going to run your games WAY faster than the intel core 2 duo version with same clock speeds JUSt because its the latest thing out, i still dont know why ANYBODY went to Vista if they are a gamer, it makes no sense, DX10 was a waste of performance, money and trouble.

Secondly when doing an FPS test , at least stess your system out by going to editor mode and running 200 guys on screen at once. At least that way people can actually get a realistic view on things as opposed to runnign around on the editor by yourself, that tells you nothing. Everyone is getting good fps on editor mode with just a few units on screen at a time. Its when the units increase to 20-50 or 200 more units is when everything gets more difficult.

Just do your research before upgrading and only upgrade by Frame per second benchmarks for gaming. My computer runs all my games mostly fast on my Intel core 2 duo 2.66 Ghz. (Arma 2 is an exception obviously, ill be upgrading in about 2 years time judging by this game), more than the quad version. I want my games to run fast NOT anything else. For example, If your about to buy a card DONT buy a card if its benchmarked at only 10 or 15 FPS higher than the card you already have running the same games, whats the point? your not getting anything that different out of it. JUst do your research THEN upgrade.

Otherwise yoru going to upgrade your computer thinking you have the best stuff, play that special game you have just upgraded for and be dissapointed time and time again.

Remember this when pc gaming, "the grass is not always greener on the other side". Companies just want to make it look greener but underneath the green pastures could be rotting wood and quicksand. Do your research first ON the game you want to play and whats running it well and whats not. IF system benchmarks for the specs you want, runs MOST games really good then go for it. If a system is benchmarked at only running one game well and the rest bad , dont touch it.

Lol thats what ive learnt over the years anyway. :):):). My system runs all my games reasonably well, but it is getting on to 3 years old now. I dont particular like spending LOADS of money on things that i THOUGHT was going to play all my games really well based on friends computersor based on the latest stuff. Ive already made thos mistakes, thinking i had the latest hardware then BANG lates game comes out in a month's time and i cant run it. I now wait for the game to come out FIRST then upgrade a year or 2 later. (processor's, video cards etc). Research has defintaley paid off for my gaming. That includes not switching from XP to vista. XP ran all my games better than vista so i stayed with XP and am glad i did.

Edited August 20, 2009 by nyran125

[ICE-Raver 10]

Woah, I'm green with envy :D. That is impressive, I see now why you say you have no probs with Arma2 performance wise. If you play online do you get high FPS or do you find that your limited by how many FPS the server can get? Like, have you ever got 40 or 50 fps online? It would seem you need a rather large clock speed (like yours) to get a better frame rate than I can get. Still disappointing that my extra 500mhz overclock doesn't give me any performance boost :/ .

I generally run between 50 & 120 fps depending on whether grass is on or not. Some servers force on grass and that kills my fps. Knocks me down to 40's sometimes high 30's. BTW-that's with everything on very high/pp normal/aa normal 1680x1050.

[Game__On 10]

@Game__On,

What is that software you are using to display SLI utilization/scaling? Is it some Nvidia specific?

No it's in the driver.

To the opening post. Testing quad vs dual core performance can't really be done with standing somewhere and comparing frames in one single shot unfortunately.

There's much more going on when you for instance start to move, run, shoot and what not. I am not sure if the enemy A.I is as taxing on the system as it't thought of, but it definately plays a role to.

Maybe it's just best to download the demo and test with the little benchmark inside to see if dual core of quad core is faster.

Also, if you're GPU limited you're gpu limited, and dual or quad core will not make much of a difference.

Let me search for this test i found online, specific about arma 2, dual cores, single cores, and i7 cpu's . brb

edit* well that was fast :D .

http://www.pcgameshardware.com/aid,687620/ArmA-2-tested-Benchmarks-with-18-CPUs/Practice/

ArmA 2 CPU benchmarks: Conclusion

ArmA 2 needs almost more CPU than GPU power and is scaling surprisingly well with higher frequencies and additional cores - nevertheless a single core is too slow. Intel's Core i7 family is dominating the competition with the Core 2 Quad models about 20 percent behind them. The latter ones are about 20 percent faster than equal Core 2 Duos running at the same clock speed. While AMD's Phenom and Athlon 64 processors are able to compete with the Core 2, the Phenom II line is too slow in comparison - although this is obviously caused by the game itself.

I went from an E8600 @4,2 ghz to this current i7, and according to the article it should be about 40% faster.

Well i think that is a bit stretched. It's more like 25-30 % in my case.

Also this is just one of many articles, i really have no idea if it can be trusted. But i'm pretty sure it's better to run arma 2 on a dual core then a single core, and preferably on a quad core.

Last edit* Here's the graphic card article by the same people. http://www.pcgameshardware.com/aid,685770/Armed-Assault-2-Graphics-card-benchmarks-and-visual-quality-compared/Practice/

Edited August 20, 2009 by Game__On

[dogz 10]

About the number of cores being utilised, the fact is that this game supposedly supports multicore when in fact it only supports Dual core, why not add Quad support too? Your point about gamers being better off with Dual Core chips is only really relevant when the application has no Multicore support at all. If an app supports multicore the Quad should shine.

I am not sure how you are monitoring CPU utilisation but my Q6600 runs at 80% - 90% on all 4 cores while in game in Arma2 XPSP3 32