ArmA2 / OA (low) performance issues

[mmaker 10]

People with stuttering problems might try install the game on different hd then their os hd. I reinstalled arma2 from my slow hd to my os disc which is faster(raptor). However even though the drive is faster i got stuttering problem every now and then, was extremely annoying. So reinstalled it back to my slower hd and they went away.

[Pranne 10]

I still have this problem with sound. As long as i stand in the field, everything ultrafast, but when iam moving, the sound is beginning to crash and the game is stuttering for unknown reasons.

It is OBVIOUS to me the sound problem is causing the stuttering, but WHY??

[VenomHowell 10]

Specs:

Asus P5N-E SLI

BFGTECH GTX 295

Intel Q6600

6 GB RAM

Windows 7 64-bit

Newest forceware release on the GPU, and ARMAII updated to 1.05.

Kinda just want to know how my performance is measuring up to other people. Just completed Operation Red, and I was having heavy stuttering on many sections even with anti-aliasing turned off and most all of my settings on normal, with a view distance of only 1000. I get about 24-30 on benchmark 1 with these settings, but actually playing the game seems to cause it to dip drastically.

I've tried many of the shortcut edits like -winxp among others with little to no benefit, and have tried turning off physX, forcing alternate frame rendering 2, max pre-rendered frames to 8. I checked in the NVidia card topic and read something about altering 'Renderframesahead' but I'm not quite sure what that refers to in the .cfg file. I have two entries with similar names that have a value of 1000 and 3 respectively, for what that's worth. Aside from that, I haven't played too heavily with the config file, so if I'm missing something I'd like to know.

Anyways, regardless. My rig has been completely murdering borderlands, MW2, Operation Flashpoint despite maxing settings but it it positively crawling and stuttering on ARMAII. I'm wondering if I might not be heavily bottlenecked by that Q6600 and require an upgrade too, and I realize that this game is extremely intense, but I don't feel like my rig is even performing to the standard it should in ARMAII.

Edited January 27, 2010 by VenomHowell

[=kct=blackmamba 44]

your fps is a little low i think, i get the same fps with view distance of 4000mtrs..

but also i'm having trouble in the chernogorsk mission.

maybe when i set the viewdistance to 1600mtrs on chernogorsk would solve that,

but its just playable this way, and i go trough that mission under 20 minutes.

what made my game a little bether was turning of the pagefile....

[Joseph Archer 10]

1600 seems enough view distance for infantry? Post process to off and antiliasing low makes the best framerates for me. Also shadows and anistropic at high.

(50fps on a 1600 by 1200 in benchmark 1 - v.high settings w/ i7@4, 4870x2, ssd)

Edited January 27, 2010 by o0Jedi0o

[Leopoldo 10]

system specs

Intel Core i7 965 EE 3.20GHz

EVGA X58 Classified motherboard (E759)

3x2GB Kingston HyperX DDR3 2000 8-8-8-24

2x EVGA GTX 295

Thermaltake Toughpower 1500W PSU

Intel 80GB Gen2 SSD (OS Windows 7 Professional 64bit)

2x VelociRaptor 300GB HD (RAID 0) (games)

2x WD Caviare Black 1TB HD (RAID 1)

1x Raptor 150GB

I have the latest BIOS, latest drivers and all my other games run amazing.

I have tried my hardware at stock, overclocked but the game in general just runs like crap in campaign mode.

I can get decent FPS but its interrupted by slow loading textures etc, stuttering, lag, freezing.

Its a shame to because i was really getting into this game but the extremely poor performance gets in the way, no matter how hard i try to ignore it.

I have been looking at videos made by the dev team and i would like to know what type of system they used to get such good quality image but smooth FPS/footage/gameplay.

EDIT OK i managed to get tid of most of the stuttering by putting ARMA 2 under my anti-virus exclusions list (Kaspersky Internet Security 2010). Also as someone else mentioned turning Video Memory tod efualt under the graphics options was a big help. The game is pretty smooth now, its not perfect but playable. I have the rez (both) set to 1920x1200 (my monitors max) and i have all the other options set to very high or high (shadow is about the only one on high along with anisotropic filtering) apart from anti-aliasing which is set to normal. Game looks great and i get between 30-60 FPS in game with an average around 35-45. When looking at EVGA Precision which lists GPU usage i am still only using a max of 75% of my 4 GPU's and most of the time that sits around 50% so there could be alot more SLI optimization one to this game.

Edited January 28, 2010 by Leopoldo

[idiosis 10]

Doubt it'll make a huge amount of difference but there's new Catalyst drivers out (10.1)

The following performance gains are noticed with this release of ATI Catalystâ„¢ 10.1:

• Left 4 Dead 2 – Overall performance improves as much as 3% on ATI Radeon™

HD 5700 series products and ATI Radeonâ„¢ HD 5800 series products

• Crysis – Performance improves as much as 3% on ATI Radeon™ HD 5700 series

products and as much as 4% on ATI Radeonâ„¢ HD 5800 series products

[Leopoldo 10]

Something i have noticed with ARMA 2 the higher you clock a card the lower the GPU is utilized.

for example, i have 2 GTX 295's in Quad SLI, when clocked at stock 576/1242/999 the GPU utilization (which i can monitor with EVGA Precision) will go up to the 90% mark on all 4 GPU's. If i then overclock the cards to something like 648/1397/1188 the GPU utilization will only get up to around 70% and FPS remain relatively the same.

Its asthough ARMA 2 has a performance cap in place which is odd because the FPS i am getting are by no means optimal yet not all the GPU horsepower is utilized.

I am running a i7 965 @ 3.70GHz so i cant see that being a bottleneck.

[Bulldogs 10]

Something i have noticed with ARMA 2 the higher you clock a card the lower the GPU is utilized.

for example, i have 2 GTX 295's in Quad SLI, when clocked at stock 576/1242/999 the GPU utilization (which i can monitor with EVGA Precision) will go up to the 90% mark on all 4 GPU's. If i then overclock the cards to something like 648/1397/1188 the GPU utilization will only get up to around 70% and FPS remain relatively the same.

Its asthough ARMA 2 has a performance cap in place which is odd because the FPS i am getting are by no means optimal yet not all the GPU horsepower is utilized.

I am running a i7 965 @ 3.70GHz so i cant see that being a bottleneck.

It's possible the bottleneck is at your hard drive. I'm running a single 5850 on max details at nice FPS, so I don't see 2 x 295's getting fully used any which way, but if there's bad performance then there's likely a Hard drive bottleneck.

[Leopoldo 10]

It's possible the bottleneck is at your hard drive. I'm running a single 5850 on max details at nice FPS, so I don't see 2 x 295's getting fully used any which way, but if there's bad performance then there's likely a Hard drive bottleneck.

what exactly do you mean "max detial"? what is "nice FPS" also? What OS are you using (32 or 64 bit version)?

whats your resolution, your view distance and what do you have each thing set to under advanced in the Graphics Options? (just because you set very high on the overall section doesnt mean everything gets set to very high.

Also what is your hard drive?

I am running 2 Western Digital VelociRaptors in a RAID 0 for the game data and my OS is on an intel 80GB SSD Gen 2. there would be enough room to put ARMA 2 on the SSD but i have read that it really doenst make a difference in most instances.

Also i did have poor performance no in FPS sense but in the sense that there was alot of lag waiting for textures to load etc, i fixed this by putting the game under "exclusions" list for my anti-virus, so now i am not getting poor performance but i am still noticing that increasing frequency seems to decrease GPU utilization so you end up with the same FPS. I am still interested to hear any tips on getting optimal performance though.

Edited January 29, 2010 by Leopoldo

[mmaker 10]

Specs:

Asus P5N-E SLI

BFGTECH GTX 295

Intel Q6600

6 GB RAM

Windows 7 64-bit

Newest forceware release on the GPU, and ARMAII updated to 1.05.

Kinda just want to know how my performance is measuring up to other people. Just completed Operation Red, and I was having heavy stuttering on many sections even with anti-aliasing turned off and most all of my settings on normal, with a view distance of only 1000. I get about 24-30 on benchmark 1 with these settings, but actually playing the game seems to cause it to dip drastically.

I've tried many of the shortcut edits like -winxp among others with little to no benefit, and have tried turning off physX, forcing alternate frame rendering 2, max pre-rendered frames to 8. I checked in the NVidia card topic and read something about altering 'Renderframesahead' but I'm not quite sure what that refers to in the .cfg file. I have two entries with similar names that have a value of 1000 and 3 respectively, for what that's worth. Aside from that, I haven't played too heavily with the config file, so if I'm missing something I'd like to know.

Anyways, regardless. My rig has been completely murdering borderlands, MW2, Operation Flashpoint despite maxing settings but it it positively crawling and stuttering on ARMAII. I'm wondering if I might not be heavily bottlenecked by that Q6600 and require an upgrade too, and I realize that this game is extremely intense, but I don't feel like my rig is even performing to the standard it should in ARMAII.

What resulotion you playing at? I got same CPU i get 37 fps average on first benckmark with these:

got 4870 512 mb GFX, 4 gig ram, vista 7 64 bit, game installed on secondary disc 7200 rpm mech drive

[Leopoldo 10]

What resulotion you playing at? I got same CPU i get 37 fps average on first benckmark with these:

got 4870 512 mb GFX, 4 gig ram, vista 7 64 bit, game installed on secondary disc 7200 rpm mech drive

that benchmark is pretty misleading, for a portion of it the camera pans up to the sky which means less or no vegetation so the FPS shoots up but for the parts where its looking at the ground the FPS are really low.

I got an average of 27 FPS on the first benchmark with everything maxed out, even view distance (well i guess that means that one of the quality settings gets turned to normal) and my monitor rez was 1920x1200 but i had the 3D rez on the max setting. Most of the time i had unplayable or what should be unplayable FPS but when it pans up to the sky i had a good period of time where the FPS was up around 60-70 FPS which made my overall FPS look better.

[Bulldogs 10]

Q9400 @ 3.2Ghz

4GB Ram

HIS HD5850

10 Hard drives, no SSD's, no RAID

Windows 7 32-bit (mainly use this one for Arma 2, but performance is same across all platforms after 1.05)

Windows 7 64-bit

Windows Vista 32-bit

Windows XP 32-bit

10.1 Catalyst drivers

V-Sync enabled (with the 5850 performance is much smoother with V-Sync on in heavy unit areas)

Settings :

All settings to very high, AA disabled (good fps with AA on high except when there's a lot of units on screen, usually over 150 units)

View Distance 3000

1920x1080 (Res and 3d Res/100% Fillrate)

Mods and startup parameters :

Mods = ACE, ACEX, CBA, LandTex, WarFX particles (often disable WarFX because it causes performance issues sometimes), sometimes vanilla.

-CPUCount=4 (gives a slight performance increase, -Maxmem=2048 gives a slight performance decrease so disabled that)

I usually get 26-28 fps in Benchmark 1, 60 during normal play, 30 in heavy areas, 22 in campaign (although feels like 60fps, could be a Fraps misread, especially with V-Sync on)

Once every few hours I'll notice texture loading problems, but they vanish very fast.

[mmaker 10]

that benchmark is pretty misleading, for a portion of it the camera pans up to the sky which means less or no vegetation so the FPS shoots up but for the parts where its looking at the ground the FPS are really low.

I got an average of 27 FPS on the first benchmark with everything maxed out, even view distance (well i guess that means that one of the quality settings gets turned to normal) and my monitor rez was 1920x1200 but i had the 3D rez on the max setting. Most of the time i had unplayable or what should be unplayable FPS but when it pans up to the sky i had a good period of time where the FPS was up around 60-70 FPS which made my overall FPS look better.

Yep your right, however that part is not that big part of the whole benchmark so is not that far away imo. Atleast not for me, im usually at 25-40 fps in multiplayer.

[=kct=blackmamba 44]

i think the benchmark is just fine

its only to let u customize for optimal performance or quality

as long as u keep using the same benchmark instead of different missions

u can actualy see if your performance is bether or worse with sertain settings, drivers or tweaks..

---------- Post added at 05:06 PM ---------- Previous post was at 05:03 PM ----------

Q9400 @ 3.2Ghz

4GB Ram

HIS HD5850

10 Hard drives, no SSD's, no RAID

Windows 7 32-bit (mainly use this one for Arma 2, but performance is same across all platforms after 1.05)

Windows 7 64-bit

Windows Vista 32-bit

Windows XP 32-bit

10.1 Catalyst drivers

V-Sync enabled (with the 5850 performance is much smoother with V-Sync on in heavy unit areas)

Settings :

All settings to very high, AA disabled (good fps with AA on high except when there's a lot of units on screen, usually over 150 units)

View Distance 3000

1920x1080 (Res and 3d Res/100% Fillrate)

Mods and startup parameters :

Mods = ACE, ACEX, CBA, LandTex, WarFX particles (often disable WarFX because it causes performance issues sometimes), sometimes vanilla.

-CPUCount=4 (gives a slight performance increase, -Maxmem=2048 gives a slight performance decrease so disabled that)

I usually get 26-28 fps in Benchmark 1, 60 during normal play, 30 in heavy areas, 22 in campaign (although feels like 60fps, could be a Fraps misread, especially with V-Sync on)

Once every few hours I'll notice texture loading problems, but they vanish very fast.

ever went to utes in the editor..

start up fraps and look there ill bet u go over 100fps, so i dont think 60 fps is a misread

[kklownboy 43]

...

I usually get 26-28 fps in Benchmark 1,...

...

your bench mark numbers are fine, think of your card as one 4980...with DX11 and faster RAM.

[Bulldogs 10]

That's what I figured. The misread I was talking about is the 22 fps in the campaign. Not sure if it's a misread or not but with my old setup it said 22 fps in the campaign when I maxed out the settings and it was really bad, now it says 22fps maxxed out and it's extremely smooth.

[otrebla_snake_ita 2]

I think they should make the game able to use 4gb of RAM on 64bit systems (obviously on 32bit systems the have to make improvements): it's possible that we have to make a RAMdisk to run this game smooth? I (and a lot of users) can't make a RAMdisk: I have "just" 4gb of ram and making a 2gb ramdisk makes no sense...not all of us have a 16gb of ram systems...

Here's my PC config: it's not the top of the world but it's still a good config for a lot of games. Game details are all on medium and I still have bad performance

Motherboard: ASUS M3A

CPU: Athlon 64 6000x2

RAM: 4gb DDR2 800mhz Kingston (1*4)

HD: MAXTOR STM3250310AS-318AS 250Gb. SATA-II 7200 RPM 8Mb

Video Card: nVidia Point of View 9600Gt 512mb

Operative System Windows Vista Home Premium 64bit SP2

Edited February 1, 2010 by otreblA_SNAKE_[ITA]

[Leopoldo 10]

Yep your right, however that part is not that big part of the whole benchmark so is not that far away imo. Atleast not for me, im usually at 25-40 fps in multiplayer.

i dont know it was long enough and let the FPS go high enough for my unplayable settings to appear to be a playable 27 FPS.

As for using it for tweaking purposes, sure if you reuse the benchmark you can compare and it would work that way but i meant its misleading sometimes due to the pan to the sky, if it did not do this you would ahve a better idea of actual gameplay FPS, how often do people spend looking up at the sky when actually playing ARMA?

[kklownboy 43]

i dont know it was long enough and let the FPS go high enough for my unplayable settings to appear to be a playable 27 FPS.

As for using it for tweaking purposes, sure if you reuse the benchmark you can compare and it would work that way but i meant its misleading sometimes due to the pan to the sky, if it did not do this you would ahve a better idea of actual gameplay FPS, how often do people spend looking up at the sky when actually playing ARMA?

The benchmarks are nothing like real game play...

---------- Post added at 10:56 AM ---------- Previous post was at 10:55 AM ----------

That's what I figured. The misread I was talking about is the 22 fps in the campaign. Not sure if it's a misread or not but with my old setup it said 22 fps in the campaign when I maxed out the settings and it was really bad, now it says 22fps maxxed out and it's extremely smooth.

yes the Campaign is smooth now with no real increase in FPS...

[=kct=blackmamba 44]

it still have to render the same distance's not to mention the AI

but i believe with the latest beta patch things will get bether,

right now im playing my way to chernogorsk to find out how the campaign performs there

yesterday i tested in the editor with the AH64 and A10 and everything looked good

now to find out how it performs with all the AI in the campaign....

[Leopoldo 10]

The benchmarks are nothing like real game play...

---------- Post added at 10:56 AM ---------- Previous post was at 10:55 AM ----------

yes the Campaign is smooth now with no real increase in FPS...

I wouldnt agree with it being nothing like the real game, it uses game assets and renders them exactly the same as if you were playing the game. I am sure there are times i the game when you could come across a similar situation with the game being a sandox.

I would say its misleading of real gameplay performance at those settings

this thread is about game performance, someone listed settings and then said they had a certain FPS with benchmark 1, i was saying that the benchmark is misleading and although the FPS may look playable that those setting would probably not be playable for the game itself.

[kklownboy 43]

I wouldnt agree with it being nothing like the real game, it uses game assets and renders them exactly the same as if you were playing the game. I am sure there are times i the game when you could come across a similar situation with the game being a sandox.

No I am never in camera mode playing like that

I would say its misleading of real gameplay performance at those settings

this thread is about game performance, someone listed settings and then said they had a certain FPS with benchmark 1, i was saying that the benchmark is misleading and although the FPS may look playable that those setting would probably not be playable for the game itself.

I get much better performance playing than i get in the benchmarks. The issue with the Benchmarks other than being a camera on a rail, is that it doesnt start the frame at the correct time. It starts to soon,(while loading) which makes a "average" irreverent.

ARMA mark on the second run is a better indicator, due to the cache being loaded already.

[Leopoldo 10]

No I am never in camera mode playing like that I get much better performance playing than i get in the benchmarks. The issue with the Benchmarks other than being a camera on a rail, is that it doesnt start the frame at the correct time. It starts to soon,(while loading) which makes a "average" irreverent.

ARMA mark on the second run is a better indicator, due to the cache being loaded already.

that would just mean there is less to render

i am not sure what your trying to say, i think your trying to say that the benchmark is not a good representation of the performance you will get in game, which is exactly what i said and cant understand why you seem determined to argue that point with me.

I disagree that the benchmark is not like playing the game, its exactly the same, there is nothing done differently in the benchmark thats not done when actually playing the game. Whether you can control the character (it being on a rail) or in first/third person view is irrelevant. Thats the whole point of the benchmark, the results are misleading however due to how they have chosen to move the camera view. Misleading doesnt mean its not using the same resources and rendering exactly the same as it would if you were playing the game but the average FPS it reports is misleading about how those quality settings might perform when actually playing the game.

[Bulldogs 10]

It's really two things.

1) the camera angles. The camera doesn't rapidly change angles so it's rendering what it sees slowly, when playing you're more likely to move your sight around quickly.

2) the complexities. The benchmark is designed to stress, especially benchmark 2, so in most situations while playing you won't have that much stress on the hardware/engine.