A2/A2F/A2OA Tweak Guide

[orcinus 121]

Another thing I have found useful: partition alignment.

If the sectors of a partition are not correctly aligned with the actual physical sectors of an advanced HD, there can be lots of extra reads/writes. This happens when the drive has been partitioned with older OSs such as XP/XP Pro. PC World mag ran tests in 2011 that showed up to 30% decrease in performance. You can find detailed explanations on Paragon's site.

Despite the PC World discussion being almost entirely about drives of 2 TB or larger, when I checked my 500 GB WD Caviar Black (which had been formatted & partitioned with XP Pro 32) all the partitions were misaligned, including "C" - they were all well off. It seems the issue isn't so much related to the size of the disk (e.g., the 2GB limit in XP) but the low-level "Advanced Format" structure of modern HDs. The sector size is different and XP has problems.

Rather than pay Paragon USD29.95, I used the align tool in the excellent free program "MiniTool Partition Wizard Home Edition" from http://www.PartitionWizard.com,

Result? I got 4-5 more FPS in the EO8 benchmark without any other changes. YMMV.

It doesn't hurt to check, and MiniTool is a really good program for manipulating disks and partitions, replacing the EASUS s/w I used for several years previously. Be sure to check for & fix any disk errors beforehand - as you should prior to any disk work. Running a surface scan is good practice as well.

Note that with any disk work, do not rely on Robinson's second law ("When all else fails, read the instructions") - read very carefully the instructions for whatever program you use first, back up your data second, and only then then align, resize, or whatever.

See my earlier post comments about HD Hacker. Whenever you make changes to your drive(s) re-run HDHacker to back up the MBR and LBRs being very careful to include in the backup folder a text file recording the time & date and brief details of what you did. If you split a logical partition into two for example, then later restore an earlier HDHacker backup you will lose data unless you restore the correct one before any read/writes to that partition.

BR

Orc

Edited January 26, 2013 by Orcinus

[RadicalAtHeart 11]

Worst guide ever. I wonder why everyone treats him like a god.

Graphics Aperture Size only applies to AGP video cards. Are you kidding me? That tweak will never work.

VideoRAM is stored on the GPU not in the actual RAM.

No wonder he is banned. The only thing I find useful was the memory allocators and the HDR thing.

nonLocalRAM makes me actually wonder if 2GB+ values are actually allowed...

[orcinus 121]

It has some mistakes, sure. So does your post :)

My ATI 5850 has 1GB VRAM & also uses 1GB of system RAM - by design. From an analysis of my system:

0301 - Card Name : ATI Radeon HD 5800 Series

0302 - Manufacturer : Advanced Micro Devices, Inc.

0303 - Chip Type : AMD Radeon Graphics Processor (0x6899)

0304 - DAC Type : Internal DAC(400MHz)

0305 - Device Key : Enum\PCI\VEN_1002&DEV_6899&SUBSYS_29701682&REV_00

0306 - Display Memory : 1024.0 MB

0307 - AdapterRAM : 1.00 GB

"Adapter RAM" is system RAM.

ArmA uses some 'trick' to get round the 2GB limit - there's a blog about it (not that I understood it, bah).

OP wasn't banned for this thread - in fact a mod praised it - but, obviously, for whatever silly thing he did to merit the byline under his name.

[chrisb 196]

Worst guide ever. I wonder why everyone treats him like a god.

Graphics Aperture Size only applies to AGP video cards. Are you kidding me? That tweak will never work.

VideoRAM is stored on the GPU not in the actual RAM.

No wonder he is banned. The only thing I find useful was the memory allocators and the HDR thing.

nonLocalRAM makes me actually wonder if 2GB+ values are actually allowed...

I don’t see any worship here for the OP, but if you found just two of the tips useful then that’s a plus, surely.

If you have tweaks and tips yourself, list them up in the guide here, anything helps, or is at least worth checking out.

Regards the memory allocators, I used the W7 system one for a while, which was good for on the ground infantry, but not really for flying. I'm back now using default allocator, which I think the game decides which, as there are a few, 'ned' probably..

HDR, I use 'normal' setting, works great for me..

[Leopardi 0]

Shadows... wasn't it so that High or Very High is better since it takes load off from CPU to GPU? While Normal is rendered with CPU?

[frequence 1]

Yes, forcing shadows to GPU thru settings them to High or Very High did help some ppl.

[chrisb 196]

Mice:

Many players, as we all know, have problems setting the game up and getting good results. We have to remember that there are many other factors that effect the performance of your game, many mentioned in the OP of this thread plus many other threads.

Mice, however, are not really talked about much, yet can be a source of problems to do with stutter, lag etc, which lead onto poor frame rates. Now I am using at present a Roccat Kone +, this is a really nice mouse. I have used many types over the years and all have some good points, even Logitech:p;).

This is what I have found over the years playing this series, Arma 2 in particular, mouse set-up is important, not in-game so much ‘x/y’ but in your mouse’s software/driver pane.

My setting for the Kone+, that differ to the default are as follows, ignoring ‘button assignments/macro’ set-up:

Advanced sensitivity = +3 on both x/y

Polling rate = 250Hz

DPI settings = 800 / 1000 / 1200 / 1400 / 1600 (I play mainly using 1000dpi)

TCU (tracking control) = ‘On’ (this can depend on your gaming surface, I use a Cyborg Glide7, scans pretty well, but any surface should, mat wise)

Distance control = ‘normal’.

Mouse acceleration = ‘Enabled’

Save whatever set-up config/profile you have.

Other setting for the performance of the mouse are at default, as said, I have ignored button assignment/macro’s.

I am not saying use these setting's, its down to personal taste, however its a guide, my gaming is very smooth with good fps for the type of gameplay I want (other than when recording video:(), so its just an idea.

I find that I have to re-load my mouse config/profile or re-confirm it, reasonably regularly, I do the same for the Rat5 I also use every now and then for A2.

I find it helpful having the mouse driver/settings open on the bottom line, I can Alt-Tab out of the game if I experience any stutter, tweak then re-load my mouse config/profile, even though the profile should already have been running, hey presto when I go back into the game the stutter goes, same for lag. Its not that it happens often, but its worth remembering

Look closely at your mouse settings, if lag or stutter are persistent, tweak the settings, keep the pane open so you can tab out and tweak further if need be, may just make a difference to your gaming experience..

Would be interesting to know what other players use for their mouse setting’s, may make a difference and help other players..:)

[Leopardi 0]

Is this guide accurate with video memory very high using 2GB? Other guides say 512MB.

[tonygrunt 10]

Is this guide accurate with video memory very high using 2GB? Other guides say 512MB.

Use default, rest of the options are mostly for circumventing problems.

[chrisb 196]

I have a HD 5850 2gb oc’ed to 7850 stock clocks, when I set the vid mem to v-high it gives me 5-8 fps boost against default. I have run vid mem at v-high for a few years now and its been great. However, we went through a period of terrible texture popping etc in beta’s a little time back, so I moved over to using default vid mem, it costs fps but gave better stability graphics wise.

My card on the in-game default setting, uses anything upto 1850 gpu mem usage, all depends on what’s going on onscreen.

Juries still out, for me anyway, on default versus v-high setting, vid mem wise. I understand that default is the favoured one, but as said, I used it for a few years on v-high with no problems and a small fps boost upto recently, but I would say default for now. At least until the next official patch is released.

Shadows I have set to high, looks much the same as v-high but there is less of a hit..

[seba1976 98]

I have a HD 5850 2gb oc’ed to 7850 stock clocks, when I set the vid mem to v-high it gives me 5-8 fps boost against default. I have run vid mem at v-high for a few years now and its been great. However, we went through a period of terrible texture popping etc in beta’s a little time back, so I moved over to using default vid mem, it costs fps but gave better stability graphics wise.

My card on the in-game default setting, uses anything upto 1850 gpu mem usage, all depends on what’s going on onscreen.

Juries still out, for me anyway, on default versus v-high setting, vid mem wise. I understand that default is the favoured one, but as said, I used it for a few years on v-high with no problems and a small fps boost upto recently, but I would say default for now. At least until the next official patch is released.

Video memory setting should not produce any difference in FPS readings, for any still rendering used for testing; and testing impact on FPS of any setting "while playing the game", is at least difficult. You should average FPS during a period of time, during which most of the action should play out consistently during all the iterations so, like I said, difficult at least. The ever-present shadow of the placebo effect, tends to be the law there.

---------- Post added at 11:51 ---------- Previous post was at 11:40 ----------

I find that I have to re-load my mouse config/profile or re-confirm it, reasonably regularly, I do the same for the Rat5 I also use every now and then for A2.

I find it helpful having the mouse driver/settings open on the bottom line, I can Alt-Tab out of the game if I experience any stutter, tweak then re-load my mouse config/profile, even though the profile should already have been running, hey presto when I go back into the game the stutter goes, same for lag. Its not that it happens often, but its worth remembering

Look closely at your mouse settings, if lag or stutter are persistent, tweak the settings, keep the pane open so you can tab out and tweak further if need be, may just make a difference to your gaming experience..

Would be interesting to know what other players use for their mouse setting’s, may make a difference and help other players..:)

Alt-tabbing while in-game, do some things that might be the real cause for the afterward smooth engine performance. "Reflushing" of the video memory is the most probable real cause for why Alt-tabbing back and forth makes a difference for you. Or maybe is the mouse :).

[chrisb 196]

I’m certainly not on my own for seeing an boost in fps, makes a real difference from ‘Default’ to ‘Very High’, still does for me, but as I said texture popping issues made me use ‘Default’ after so many years on VH. But certainly there is a fps boost.

Regards the mouse, so many players take little notice of their mouse controls, which is a shame because it can cause real problems in a game, but there again, if you don’t experience any of these things, then that’s a bonus for you..;)

[seba1976 98]

I’m certainly not on my own for seeing an boost in fps, makes a real difference from ‘Default’ to ‘Very High’, still does for me, but as I said texture popping issues made me use ‘Default’ after so many years on VH. But certainly there is a fps boost.

Are you certain? I'm not going on for the sake of proving you wrong or just argumenting, I'd really like to see the effect myself. If it's not much of a trouble, could you set up a player in any place in Chernarus, take note of the FPS reading once it's stable, then going to video options, change the video memory setting, and see if the FPS has changed? (Always looking at the same initial spot.) And if you do, could you please take some screenshots of the proccess? If it's too much asking, please disregard.

[kklownboy 43]

I believe when you change the vidmem, it changes the frame-buffer usage, and depending on your Display rez and graphic setting you can get some (5~) extra frames.But not with Vsync on.

[chrisb 196]

Are you certain? I'm not going on for the sake of proving you wrong or just argumenting, I'd really like to see the effect myself. If it's not much of a trouble, could you set up a player in any place in Chernarus, take note of the FPS reading once it's stable, then going to video options, change the video memory setting, and see if the FPS has changed? (Always looking at the same initial spot.) And if you do, could you please take some screenshots of the proccess? If it's too much asking, please disregard.

Have a look at the discussion here, between ‘doveman’ and myself, third from the bottom is the main one to look at..

I told @doveman about the same thing, I tell anyone looking for some fps increase to try out changing from vidmem ‘default’ to ‘v/high’, he saw a good increase. Its something over the years I stumbled across. Not sure why it happens, but it does.

There are screenshots of fps bits and bobs, concerning fps increases.

Problem is though some get texture issues using it that way, some don’t. I never got any texture problems for all the time I was on v/h which was years really. Its only recently over one of the last 3/4 beta’s I have noticed that I started to get the texture issue so I switched to ‘Default’ for vid mem, dropped a little fps, but I had the room for that. I think @doveman got the same, soon after he changed, so decided to change back.

As said works for some and not for others, just one of those things..

I usually am 'certain' of anything I say concerning the game or indeed most things. I only say it if I have experienced it myself or seen it working..;)

Same with the mouse thing, players should really take time to set a gaming mouse up, can impact game-play quite considerably, if they don't..:D

Oh yes, v-sync off, I never have it on...

[seba1976 98]

I usually am 'certain' of anything I say concerning the game or indeed most things. I only say it if I have experienced it myself or seen it working..;)

And, there's a thing called "error", with too much tradition behind it.

And I still don't see your before and after picture of the same scene, with the same settings except the different video mem. setting, to prove you're right. I mean don't get me wrong, but I can't take your word for it, nor would I expect you to take mine if I would saying something it's against your empirical data and theoretical analysis. I want to learn, but you have to show me something... or you can leave it at it, if you don't feel like it.

---------- Post added at 16:39 ---------- Previous post was at 16:37 ----------

I believe when you change the vidmem, it changes the frame-buffer usage, and depending on your Display rez and graphic setting you can get some (5~) extra frames.But not with Vsync on.

That would make more sense, I mean, one can consider that to be valid, but still, the proof...

[kklownboy 43]

bla bla...

So I did test the Vidmem settings. Low, Normal and High will change your Texture Detail to match, and so will increase the FPS accordingly.

Default and VeryHigh will not change your Texture Detail.

Standing still looking / and then zooming with a micro CCO;

Default Mem =55 / 66

VeryHigh =66 / 78

Sooo there you go. VH is +12fps in zoom where it counts most. Ill play with it for awhile, see if i get poly-spikes etc...

Hadn't noticed this type of gain before. Always was small. hmmmm

Well after more testing, I see why the fps is higher, the IQ is reduced. Its very subtle, and if you dont run Max/VH settings, you may not notice. But you loose finer detail on a lot of objects untill you get close to them. Not worth it.

Edited March 4, 2013 by kklownboy
i didnt know wtf i was talking about.

[seba1976 98]

Well after more testing, I see why the fps is higher, the IQ is reduced. Its very subtle, and if you dont run Max/VH settings, you may not notice. But you loose finer detail on a lot of objects untill you get close to them. Not worth it.

Great, thanks a lot for testing, really. That was the exact kind of objective test I was striving for. I got very a similar result when playing with the -maxVram parameter the OP mentioned. He said that lowering the value to 50% of your available VRAM would help FPS, and I did notice the gain but the was noticeable change of detail levels (having only 512 MB I couldn't really be an example of what the average user experience might have been, but still).

[kazenokizu 10]

not that my fps is bad but could be a bit better. but i noticed no increase playing with memory allocator or vram settings

and well yea general stuff but ether way no increase in fps for me at all. and i dont think the memory allocator is switching me over to the one i want and is keeping me on the bis one idk why tho. any help none the less would be appreciated it

[chrisb 196]

not that my fps is bad but could be a bit better. but i noticed no increase playing with memory allocator or vram settings

and well yea general stuff but ether way no increase in fps for me at all. and i dont think the memory allocator is switching me over to the one i want and is keeping me on the bis one idk why tho. any help none the less would be appreciated it

The allocators are in your Arma 2 ‘dll’ folder. The game however uses the best one for the job I think, so it will probably be ‘ned’.

If your using win7, you could try the default, system malloc, just put it in your memory line:

-malloc=system

Not sure if it would work however, but just worth a try if you haven’t already done it..

[timbee71 1]

A big thanks to all the posters of this thread's useful information.

My post is for Nvidia users. I have a Gigabyte Geforce GTX680 (rev A2 4096MB GDDR5) which struggles to deliver its potential, because it's hindered by old CPU (Q6600) and memory (DDR2-800). I use Windows' own memory management. I had been using the excellent OC tool from Orbmu2k to push my FPS - the tool has some registry hacks specific to ArmA2OA, but even on a 10% OC setting was only able to achieve 20-30 FPS in multiplayer (spiking to about 80 when things are quiet). I recently discovered Nvidia GeForce Experience software which makes the below game specific changes to the graphics preferences in the Nvidia control panel - these are quite different to those I had picked up from reading the BI Forums (admit I'm not the most thorough of readers) but I'm now enjoying a better image quality and an FPS that is seldom lower than 20 in MP and mostly around 30-35 (spikes over 90):

Anisotropic filtering - very high

Antialiasing - very high

ATOC - all trees and grass

Fullscreen - on

HDR quality - very high

Objects detail - very high

PPAA - disabled

Resolution - (monitor native)

Shadow detail - normal

Terrain detail - high

Texture detail - very high

Video memory - default

Visibility - 0 (!)

Once in game, I found my profiles reflected these changes - the NE control panel doesn't mention Vsync, which was and still is disabled in the games Video Options, but I also found FSAA/SMAA disabled (which surprised me as I thought these would be the preference over Aniso.filtering and traditional AA). In-game I found the visibility was actually set to 500.

I hope this info is useful to some members.

Edited May 19, 2013 by timbee71
take out inaccuracies

[domokun 515]

Interesting feedback though you might want to increase your Shadow Detail to High, even Very High as these higher settings offload the calculations from the CPU to the GPU and seeing as your machine has a weak CPU but strong GPU, it should offer you better framerates AND better visuals.

You'll find that tweak and others listed here:

http://forums.bistudio.com/showthread.php?85124-ArmA2-OA-%28low%29-performance-issues&p=2081466#post2081466

If you try any of them out, let us know how you get on.

[timbee71 1]

Good suggestion, domokun, thank you. I have been following the tweaks per the link you provide for some time now (principally using PPAA rather than traditional AA), which was why the Nvidia Experience settings surprised me. What you said about transferring CPU calcs to the GPU made sense, so I tried your suggestion (using the ArmA2 Benchmark 01) and got average 30 FPS on both settings, normal and high, but 31 FPS on very high shadow. I also noticed the extra stress kicks the card into boost mode earlier, so I'll be running with your suggestion from now. Sorry for slow reply and thanks again.

Edited May 19, 2013 by timbee71

[domokun 515]

It's a good idea to test using empirical approach as it's less subjective.

IIRC the E08 benchmark is generally considered to be a more accurate representation of gameplay experience than the A2 benchmark.

Also bear in mind that 30 fps might sound low but in game, provided you don't have a high View Distance, you should enjoy much higher framerates, i.e. the game should be much fluid.

Although high framerates (60+) are not as important in Arma as twitch shooters (CoD, BF, etc.), it is important to keep them above 25 fps as below this threshold the game will simplify AI logic in an effort to maintain fluid gameplay.

[timbee71 1]

Interesting factoid about FPS and AI logic, thanks, but guess I'll have to stop patting myself for beating the AI :-(