Jump to content
Sign in to follow this  
Alabatross

When will the nVidia Arma 3 optimized drivers come?

Recommended Posts

Where do you get these number from, the ones that say what FPS you should get? How can you possibly know what FPS you should get? Simply referring to another game cannot be the answer, because as we all know ArmA treats its game world rather differently than other games. It's non player-centric, so all things on the map need to be given CPU time, not just stuff surrounding the player. IMO the engine is already highly optimised because it often surprises me with what it can do. Last night I had nearly 300 units on foot, transport & armour, and all of them required pathfinding & engagement logic. And it was playable, with not much appreciable slowdown.

Where i get those numbers from? Oh i don't know, from 15 years of playing games from all genres, from being in dozens of beta and alpha tests, from seeing games go from unoptimized to optimized and sadly seeing many games never bothering to actually get optimized? It doesn't take a genious to work out that Arma 3 is unoptimized for the vast majority of people. Same with Rome 2. Same with Arma 2 at launch. Same with Vanguard: Saga of Heroes lol.

It doesn't matter which genre or which type of game it is. Anyone that has had any experience what so ever in playing PC games knows when a game is unoptimized or not.

It's not about expecting a AAA amazing game to run on your granddads laptop, it's about seeing what the game delivers in complexity and depth, and comparing that with the framerate you're getting.

Simple as that. And yes we can compare other games. If another developer would manage to pull off the same things that BIS does, with a higher framerate. Then obviously BIS has something to learn, right?

Or would you argue that everything is perfect already and no improvement can possibly be made without intervention from alien technology?

I'd argue that i'm more of an optimist than you, since i see the possibility for vast improvements whereas you see a brick wall that can't be crossed.

You're one of the lucky few with playable framerates with, what did you just say? 300 units?

Share this post


Link to post
Share on other sites
Where i get those numbers from? Oh i don't know, from 15 years of playing games from all genres, from being in dozens of beta and alpha tests, from seeing games go from unoptimized to optimized and sadly seeing many games never bothering to actually get optimized? It doesn't take a genious to work out that Arma 3 is unoptimized for the vast majority of people. Same with Rome 2. Same with Arma 2 at launch. Same with Vanguard: Saga of Heroes lol.

It doesn't matter which genre or which type of game it is. Anyone that has had any experience what so ever in playing PC games knows when a game is unoptimized or not.

Oh OK, you're so experienced with gaming that you can just "tell". I think I see what's going on here.

It doesn't take a genious to work out that Arma 3 is unoptimized

QFT.

It's not about expecting a AAA amazing game to run on your granddads laptop, it's about seeing what the game delivers in complexity and depth, and comparing that with the framerate you're getting.

Simple as that. And yes we can compare other games. If another developer would manage to pull off the same things that BIS does, with a higher framerate. Then obviously BIS has something to learn, right?

I agree. Now all we need is another game that does the same thing that ArmA does.

Or would you argue that everything is perfect already and no improvement can possibly be made without intervention from alien technology?

I'd argue that i'm more of an optimist than you, since i see the possibility for vast improvements whereas you see a brick wall that can't be crossed.

Assigning unreasonable characteristics to people, then arguing against that characteristic might win you a point in your mind, but other people will recognise the situation. If by "vast improvements" that only you can see, you mean "more FPS", then I think people in general are more flexible than that :)

You're one of the lucky few with playable framerates with, what did you just say? 300 units?

That was 300 military units, I didn't count ambient civilians & animals etc. More than that and I get sluggish gameplay. But I think, with my current setup, I doubt I'll need to go over that :) in fact my preferred level is probably around 100 or so.

Share this post


Link to post
Share on other sites

It depends what those AI units do. Just laying around don't take much CPU time, but put "few" of them face to face and the fps will plummet even on fast CPUs. Also, on higher draw distance you come into the limitation of 32bit arch. Some 64 bit exe + optimizations or dx11.2 Tilled Resources is what will do the trick, but will most definitely won't happen soon.

Share this post


Link to post
Share on other sites
Where i get those numbers from? Oh i don't know, from 15 years of playing games from all genres, from being in dozens of beta and alpha tests, from seeing games go from unoptimized to optimized and sadly seeing many games never bothering to actually get optimized? It doesn't take a genious to work out that Arma 3 is unoptimized for the vast majority of people. Same with Rome 2. Same with Arma 2 at launch. Same with Vanguard: Saga of Heroes lol.

It doesn't matter which genre or which type of game it is. Anyone that has had any experience what so ever in playing PC games knows when a game is unoptimized or not.

It's not about expecting a AAA amazing game to run on your granddads laptop, it's about seeing what the game delivers in complexity and depth, and comparing that with the framerate you're getting.

Simple as that. And yes we can compare other games. If another developer would manage to pull off the same things that BIS does, with a higher framerate. Then obviously BIS has something to learn, right?

Or would you argue that everything is perfect already and no improvement can possibly be made without intervention from alien technology?

I'd argue that i'm more of an optimist than you, since i see the possibility for vast improvements whereas you see a brick wall that can't be crossed.

You're one of the lucky few with playable framerates with, what did you just say? 300 units?

Several problems here, first off, there are no other games like ArmA 3 (so there is nothing to compare it to).

Secondly, just because you can put 600 units in a scenario and then add 50 players to that scenario doesn't mean you should, or that it is going to run well.

Thirdly, as much as performance could be better (and probably will be when all is said and done), many people expect their 'granddad's laptop' to run A3 at enthusiast levels without enthusiast level hardware.

It's very possible to have decent FPS as long as you stick to missions/scenarios that don't take the piss.

We'd all love to be able to run 600 AI (in built up areas), armor, planes and helicopters etc but as much as the software exists to design this, I fear it is only DARPA that has the hardware to run it properly ;)

Share this post


Link to post
Share on other sites

Its a sandbox game that YOU can optimize for yourself. By not using a large Display, turning down Graphic filters, and limiting AI etc.. till it has the frames you want.

In SP you can get fine FPS(near or above Refresh) with 300 entities or so, on good hardware, BUT the same self optimizing is required. Saying BIS doesn't optimize... well yeah, SO. They have a lot of objects...and they just did do some optimizing of some.

The FSM of AI(Finite state machine) takes alot of CPU, and still needs to be smarter thus taking more CPU if we want awesome AI smarts. Where does that leave us.. well we have to buy bigger better.. it goes with the territory. It is how its always been. Unless your a young gamer(25~and less) who grew-up on consoles. And even with all that YOU still have to optimize it your self.

A large missions MP 30fps, small MP 70+fps, SP way more. Oh they actually sell more ARMA for SP then MP by a factor of alot(lol) more than 5 to 1 if I pull # out my a$$.

As for nvidia Driver... well just make sure you get a REAL dx11 Nvida card, first, no 5xx or below are. Even some 6xx are not. Then when you have a true DX11 card, you can get DX11 opti drivers. Lets not talk about you XP guys, with 275, or 290s...yikes!

Share this post


Link to post
Share on other sites

Can anyone running an SLI graphics card rig help confirm this for me?

If I disable multi-gpu mode in my Nvidia Control Panel, it makes absolutely no difference to my Arma3 FPS, as measured using ArmA3Mark mod.

Please don't bring up GPU percentage usage, I'm only talking about using two GPU's vs only one. Seems to make no difference at all. Can anyone confirm?

(on a bright note, seems to totally solve the flickering PIP issue)

Share this post


Link to post
Share on other sites
Can anyone running an SLI graphics card rig help confirm this for me?

If I disable multi-gpu mode in my Nvidia Control Panel, it makes absolutely no difference to my Arma3 FPS, as measured using ArmA3Mark mod.

Please don't bring up GPU percentage usage, I'm only talking about using two GPU's vs only one. Seems to make no difference at all. Can anyone confirm?

(on a bright note, seems to totally solve the flickering PIP issue)

Depends where you are testing - in built up areas (or when there is a lot going on) it won't make much of a difference, in less intensive situations (where you are not so CPU limited), it does.

PS: It's 'SLI Rendering mode' you need to change in the NVCP, not Multi Display/Mixed GPU acceleration.

Edited by BangTail

Share this post


Link to post
Share on other sites
PS: It's 'SLI Rendering mode' you need to change in the NVCP, not Multi Display/Mixed GPU acceleration.

Sorry, change SLI rendering mode to what? not SLI_PREDEFINED_MODE_FORCE_SFR, that does not fix the flickering PIP problem.

what does fix it is to go 3D settings/manage3d settings/program settings, then find arma3, then switch to one GPU. I forget the exact line, I'm at work and don't have an SLI setup here.

Also, just to clarify, when I run ArmA3Mark, I see NO difference at all in my resulting score. None at all. The benchmark is pretty diversified in it's environments and events...at least diversified enough to show SOME difference between using one and two GPU's. Wondering if anyone else can confirm that there is zero difference between the two setups, one and two GPU's.

Share this post


Link to post
Share on other sites
Can anyone running an SLI graphics card rig help confirm this for me?

If I disable multi-gpu mode in my Nvidia Control Panel, it makes absolutely no difference to my Arma3 FPS, as measured using ArmA3Mark mod.

Please don't bring up GPU percentage usage, I'm only talking about using two GPU's vs only one. Seems to make no difference at all. Can anyone confirm?

(on a bright note, seems to totally solve the flickering PIP issue)

SLI works when you are not limited by other factors. I just run the benchmark with View Distance at 3000 and Object Distance at 1000 and I got 74 FPS with SLI and 51 without.

Share this post


Link to post
Share on other sites
Sorry, change SLI rendering mode to what? not SLI_PREDEFINED_MODE_FORCE_SFR, that does not fix the flickering PIP problem.

what does fix it is to go 3D settings/manage3d settings/program settings, then find arma3, then switch to one GPU. I forget the exact line, I'm at work and don't have an SLI setup here.

Also, just to clarify, when I run ArmA3Mark, I see NO difference at all in my resulting score. None at all. The benchmark is pretty diversified in it's environments and events...at least diversified enough to show SOME difference between using one and two GPU's. Wondering if anyone else can confirm that there is zero difference between the two setups, one and two GPU's.

I can confirm, as Tony Grunt did, that SLI works fine - anyone who 'confirms' that there is no difference does not have their system set up correctly.

PS: When you ask technical questions, it is helpful if you provide your System specs.

Share this post


Link to post
Share on other sites
I can confirm, as Tony Grunt did, that SLI works fine - anyone who 'confirms' that there is no difference does not have their system set up correctly.

PS: When you ask technical questions, it is helpful if you provide your System specs.

Well, I guess that's really what I wanted to know. If it's not happening with you guys then it's something on my end.

Nvidia GTX590

i7 2600 quad OC to 4Ghz

250Gig SSD main drive, at about 515M/sec read

16Gigs ram

Not sure what to do. I don't see any difference at all, at least using the Arma benchmark, between two and one GPU usage. Only difference I see is with one GPU, I get no PIP flicker.

I can only assume that I am being very very abruptly bottlenecked by my CPU, like hitting a FPS wall...cause again, there is no difference between one and two GPU, not even 1fps difference.

Share this post


Link to post
Share on other sites

Well... I want the physx to be handled by my GPU and not the CPU.

I also want further & better optimization for my gfx card, running a gtx760 SC and so far I don't see much of a difference between that and my old gtx570 classified. (the difference is quite laughable)

Also... I want my CPU to be used @ its maximum potential as well, I got an i5 quad 2500k SBridge OC'd @ 4.5ghz and it doesn't even feel like I've got that much power under the hood -_-

Anyway, this is still a work in progress and I hope major improvements will be made in a near future.

Share this post


Link to post
Share on other sites
Well, I guess that's really what I wanted to know. If it's not happening with you guys then it's something on my end.

Nvidia GTX590

i7 2600 quad OC to 4Ghz

250Gig SSD main drive, at about 515M/sec read

16Gigs ram

Not sure what to do. I don't see any difference at all, at least using the Arma benchmark, between two and one GPU usage. Only difference I see is with one GPU, I get no PIP flicker.

I can only assume that I am being very very abruptly bottlenecked by my CPU, like hitting a FPS wall...cause again, there is no difference between one and two GPU, not even 1fps difference.

Best to test in an open field or something like that, where there are no buildings.

Share this post


Link to post
Share on other sites

3930k@5ghz

32gb@2400

Gtx680 sli

Revodrive 3x2

Arma in ramdisk

No differences here with or without sli... If to see differences I've to lower details and graphics... No, that's not exactly sli was made for :-P

Tell me about my hw bottlenecks ( lol ) and not about that needs for the game to be optimized and also drivers and sli profile..

If you think these are the game limits.. You probably have forgot arma 2 1.00 and how is now with 1.62 beta...

There's more space for tweak fixes and optimizations than you think , let's only wait and don't always say "hw bottleneck" please!

Share this post


Link to post
Share on other sites
3930k@5ghz

32gb@2400

Gtx680 sli

Revodrive 3x2

Arma in ramdisk

No differences here with or without sli... If to see differences I've to lower details and graphics... No, that's not exactly sli was made for :-P

Tell me about my hw bottlenecks ( lol ) and not about that needs for the game to be optimized and also drivers and sli profile..

If you think these are the game limits.. You probably have forgot arma 2 1.00 and how is now with 1.62 beta...

There's more space for tweak fixes and optimizations than you think , let's only wait and don't always say "hw bottleneck" please!

You can deny it all you like but it is a problem attributed to the game being CPU 'heavy'.

Of course the there is always room for optimizations, but SLI works fine (as it has done in every OFP/ArmA title).

There has been an SLI profile since the Alpha was released, JFYI.

Share this post


Link to post
Share on other sites

Cpu usage is around... 20-25% ... And I'm talking about the whole cpu usage... I'm not here to flame or say that game sucks cuz I don't get 120fps, I'm just saying that say always the same things ( cpu bottleneck and hw bottleneck ) is like saying : game is perfectly optimized and use the 100% of hw power so please don't optimize it anymore ...

I repeat, you remeber arma 2 at his first release and you know how is now? How many years? So please, don't put into the minds of the "new" arma3' user that this game wont get improvements because it's a lie!

Give to developers the trust and the time they need!

Share this post


Link to post
Share on other sites
Cpu usage is around... 20-25% ... And I'm talking about the whole cpu usage... I'm not here to flame or say that game sucks cuz I don't get 120fps, I'm just saying that say always the same things ( cpu bottleneck and hw bottleneck ) is like saying : game is perfectly optimized and use the 100% of hw power so please don't optimize it anymore ...

I repeat, you remeber arma 2 at his first release and you know how is now? How many years? So please, don't put into the minds of the "new" arma3' user that this game wont get improvements because it's a lie!

The problem with some people is that they believe (wrongly) that if GPU/CPU usage is not ~100%, something is wrong.

This denotes a fundamental misunderstanding of how CPUs/GPUs work.

In reality, very few games push GPUs or CPUs to 100% of their available processing capabilities (if you want to do that, run Furmark and Prime simultaneously ;) )

Secondly, I never said there wouldn't be improvements, but I wouldn't count on the 'night and day' differences you seem to be expecting based (erroneously) on CPU/GPU usage.

Share this post


Link to post
Share on other sites

I'm talking about 5 things ( connected each others )

Cpu load

Gpu usage

Sli profile ( also drivers of course )

ArmA 3 not optimized to use hw power (example: you know something called headless client for make servers running "faster" cuz of arma cpu/ram usage? Why they had to split run arma into 2 difference instanceson the same machine if it's so well optimized?! )

Bis always made miracles between first release and patches

Maybe I give too much trust to developers but atm they got 13 years of experience with this series and I think game will run more and more smoother within some months

Share this post


Link to post
Share on other sites
The problem with some people is that they believe (wrongly) that if GPU/CPU usage is not ~100%, something is wrong.

This denotes a fundamental misunderstanding of how CPUs/GPUs work.

In reality, very few games push GPUs or CPUs to 100% of their available processing capabilities (if you want to do that, run Furmark and Prime simultaneously ;) )

Secondly, I never said there wouldn't be improvements, but I wouldn't count on the 'night and day' differences you seem to be expecting based (erroneously) on CPU/GPU usage.

No. People see their frame rate tanking along with tanking GPU usage and realize something is clearly wrong.

If the game is so "CPU heavy", it should use more than two cores worth of total usage. Period. It is not using what it needs.

The issue is not that the game is not using 100% CPU, or 100% GPU. The issue is that the game doesn't utilize what it needs to run well (or has some serious coding problems that are severely bottlenecking the whole stream). When you watch your fps drop to 15 and see that it directly correlates to significant drops in GPU usage, it is clear something is wrong. When, open water yields high fps and high gpu usage (I know some have the opposite problem), but an empty city causes GPU usage to bottleneck and drop significantly (and fps goes with it) it is clear something is wrong.

I hear this comment that people believe the game should use 100% of this or that... But, I just don't actually see that many people making such a claim. They are responding to the clear observation that can be made when their game takes an fps hit. FPS drops (at least in single player, or in editor on an empty map where we can keep other issues introduced by mp out of the equation) are pretty much always accompanied by drops in GPU usage.

When you play BIS made showcase missions at 15-25 fps with GPU usage getting down into the freaking teens, it is clear something is very wrong. When, empty areas of the map are enough to cause fps tanking gpu usage drops, it is clear there is something very wrong. When settings adjustments from full on maxed down to 720p cell phone gaming don't significantly impact fps performance under these scenarios, it is clear something is wrong with the game.

It's one thing if the game is running at 20fps and my gpu is screaming for a break. But, that is not what is happening here. The main bottleneck in this game is NOT an end-user hardware bottleneck. It is the game. Either their engine, or some buggy coding or what, only they know (they sure better after this many years)... Enough people, with good enough systems have observed this. Enough people have tried every possible tweak out there, and monitored their systems and performance. It is clear on which side the real root of this problem lies. Even if there is a lot of cover-providing noise in the form of system specs, mis-stating people's observations, claims of the game being too big and grande, etc... and this and that that all get thrown about in the vacuum of non-participation, and non-information on this issue from BIS.

Share this post


Link to post
Share on other sites
No. People see their frame rate tanking along with tanking GPU usage and realize something is clearly wrong.

If the game is so "CPU heavy", it should use more than two cores worth of total usage. Period. It is not using what it needs.

The issue is not that the game is not using 100% CPU, or 100% GPU. The issue is that the game doesn't utilize what it needs to run well (or has some serious coding problems that are severely bottlenecking the whole stream). When you watch your fps drop to 15 and see that it directly correlates to significant drops in GPU usage, it is clear something is wrong. When, open water yields high fps and high gpu usage (I know some have the opposite problem), but an empty city causes GPU usage to bottleneck and drop significantly (and fps goes with it) it is clear something is wrong.

I hear this comment that people believe the game should use 100% of this or that... But, I just don't actually see that many people making such a claim. They are responding to the clear observation that can be made when their game takes an fps hit. FPS drops (at least in single player, or in editor on an empty map where we can keep other issues introduced by mp out of the equation) are pretty much always accompanied by drops in GPU usage.

When you play BIS made showcase missions at 15-25 fps with GPU usage getting down into the freaking teens, it is clear something is very wrong. When, empty areas of the map are enough to cause fps tanking gpu usage drops, it is clear there is something very wrong. When settings adjustments from full on maxed down to 720p cell phone gaming don't significantly impact fps performance under these scenarios, it is clear something is wrong with the game.

It's one thing if the game is running at 20fps and my gpu is screaming for a break. But, that is not what is happening here. The main bottleneck in this game is NOT an end-user hardware bottleneck. It is the game. Either their engine, or some buggy coding or what, only they know (they sure better after this many years)... Enough people, with good enough systems have observed this. Enough people have tried every possible tweak out there, and monitored their systems and performance. It is clear on which side the real root of this problem lies. Even if there is a lot of cover-providing noise in the form of system specs, mis-stating people's observations, claims of the game being too big and grande, etc... and this and that that all get thrown about in the vacuum of non-participation, and non-information on this issue from BIS.

I was replying to someone who made that very assumption, so please don't accuse others of obfuscating when it is, in fact you, that is replying out of context.

Cheers :)

Share this post


Link to post
Share on other sites

Coreparking...OFF. You will get one core up in the 70%+(lots of 80, some 90s, this is hours of play, so 90% is 8/10min runs). That alone bumps me 8 fps.

As for the days when SLI or CF doubled your frame rate... well, I remember, And I can if i lower my Resolution.

Two 580s on a stick with only 4gz CPU At what Resolution? Need 4.6 on that 2600k and a new GTX780, or R9 290x?

680SLI ,are they 4GB? if not, then your Resolution is too high. But you can max out the Graphics less the VD/ObjD.

Main disk is a SSD and RamDisk for ARMA? whats your"other" disk, and do you alow MS to move your page around and cache your RAMdisk?

There will be no magic, just a few extra here and there, maybe 10 more FPS to be had, from Bis on Altis. A2 really didnt improve much more than 10fps, it was desert maps and new hardware, that is the reason A2 got better. And now we need better hardware and API routines but same as always. Lower your rez or buy better gear.Right now H/W is behind the curve for Altis, relative to 1080p/60fps minimum de jour...

Share this post


Link to post
Share on other sites
I was replying to someone who made that very assumption, so please don't accuse others of obfuscating when it is, in fact you, that is replying out of context.

Cheers :)

He said no such thing, as far as I can see. And, I also didn't say anywhere that "no one" has said such a thing. I said "I just don't actually see *that many* people making such a claim."

To further, add... even if he had said it... this is the only stuff some people seem to pick up on as a point of conversation. And, it is just mis-direction at best, is my point. Anyone can easily monitor their hardware usage and see whether or not their hardware is being overtaxed or not. I am still looking for evidence to back up the claim that this game is "CPU heavy". Because, if that is true, then it only serves to prove the overriding point. That there is a serious, serious problem, because their engine can't currently utilize enough available cpu resources to meet the demands the game is placing on it. But, you already know this...

Read his post again. He said no such thing. In fact, looks to me like he was pointing out the reason he takes issue with people bandying about the "hardware bottleneck" claim is b/c it insinuates (to him) that the game is already taking full advantage of our hardware, only our hardware isn't powerful enough. Which, he has expressed disagreement with.

You quoted him directly. Your first sentence said:

The problem with some people is that they believe (wrongly) that if GPU/CPU usage is not ~100%, something is wrong.

I fail to see where he even hinted at holding this "belief".

Edited by Mobile_Medic

Share this post


Link to post
Share on other sites
Cpu usage is around... 20-25% ... And I'm talking about the whole cpu usage...

Stop trying to foment an argument over a discussion (that you seem to have difficulty interpreting) that had nothing to do with you in the first place :rolleyes:

Hell, A3 is by no means perfect, but as I said to you previously, deal with it or move on (it will improve as the patches roll out).

Edited by BangTail
Clarity

Share this post


Link to post
Share on other sites

680SLI ,are they 4GB? if not, then your Resolution is too high. But you can max out the Graphics less the VD/ObjD.

Main disk is a SSD and RamDisk for ARMA? whats your"other" disk, and do you alow MS to move your page around and cache your RAMdisk?

There will be no magic, just a few extra here and there, maybe 10 more FPS to be had, from Bis on Altis. A2 really didnt improve much more than 10fps, it was desert maps and new hardware, that is the reason A2 got better. And now we need better hardware and API routines but same as always. Lower your rez or buy better gear.Right now H/W is behind the curve for Altis, relative to 1080p/60fps minimum de jour...

I've heard of 1 person (just 1 for now) who is using the http://dimmdrive.com gaming ramdisk software with ARMA III and apparently texture draws, such as moving really fast, don't lag/bog out and result in fps drops anymore. Hopefully that'll help, and I imagine it would because reading and read/write to cache and the general texture files should be instant.

Share this post


Link to post
Share on other sites
Firstly, what you have said I quoted from Ramius86 is an original statement by me, not by Ramius86.

Huh? I said "Your first sentence said:" right before quoting you. I know you said that. That was my point. You quoted him directly, and then insinuated that he held the belief that the game must run at 100% usage, when he made no such claim in the text you quoted.

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
Sign in to follow this  

×