Low CPU utilization & Low FPS

[das attorney 857]

Myke;2678410']Sure' date=' with an easy example: imagine a room full of people solving math problems on their desk. Every persons math problems include results from other people math problems. Without those results they can't continue and have to wait. Get it?[/quote']

A bit like NASA mission control in the '60s.

[windies 11]

Myke;2678410']Sure' date=' with an easy example: imagine a room full of people solving math problems on their desk. Every persons math problems include results from other people math problems. Without those results they can't continue and have to wait. Get it?

And again the BF4 comparison. Calculating eyecandy is pretty straightforward and doesn't have the above described problem since nothing is really gameplay relevant. So please, stop taking BF4 as comparison or keep playing BF4. And maybe read some articles about multithreaded programming and it's downsides.[/quote']

Except this analogy can be proven false simply by the way AI is handled online in that client AI is calculated Client side and server AI is calculated Server side and it's sync'd over the network, so theoretically why can't Core 1 AI be calculated on Core 1 and Core 2 AI be calculated on Core 2? Why the need for it to be single threaded I ask again? Also what about the other issue's that don't stem from AI but rather simply from the way the engine works and is programmed to work?

[F3r4_N3n3m 10]

Hi guys....

There is no Need to fight about this.

Bottom line is... the game does not deliver the Performance that Players want. You may justify that by saying that Arma is a different game, more complex... bla bla bla. The fact is... People want a game they can Play. I invested a huge amount of Money on my PC and I feel really stupid when I Play Arma. And even if I give out more 1.000, 2.000 euros I know the game is gonna run just as bad.

Yes, I imagine multithreading should be complex as hell... But the Software optimization seems to be very poor on Arma. The fact that a much "simpler" game as BF4 is able to use more of my RAM, more of my VRAM, more of my GPU and more of my CPU... that says everything to me. They have a simpler Software that makes better use of my Hardware. Therefore they deliver 175% more fps.

At the same time arma runs a ridiculus 20% GPU utilization. And the CPU utilization is jumping up and down on all cores, going as low as 20% a lot of times. If the game is so CPU oriented it should at least use all the power I have available. There is no bottleneck, because everything is under optimized.

And fighting with customers ain't gonna help at all. We are a community that loves games and we are very passionate about the idea of Arma. So we would like to hear what BI is doing to solve their Problems. How they are going to make our experience better in the game. We want to know that we are taken seriously and that they are working on optimizing this game. Not jus cashing in on the suckers that buy this crap.

Show us some respect... that is all we want.

The community is here making tests and wasting time in order to help you guys. I wouldn't have spent so much time searching for ways to improve my Performance and sharing my experience if I wasn't commited to make this a better game.

To summarize all my nonsense talking... please don't give us this bs talk about enjoyable gameplay arround 30fps. Don't give me the excuse that it is because of the amount of AI opponents that the game is crapy. Other games also have these Situations, but they drop from 150 to 100 fps in such cases. That is the same expectation I have for Arma as a gamer.

Otherwise you guys should be developing Software for consoles that are played by Kids and Soccer moms! They don't dive a shit about Performance and they dind't spend this amount of cash on their console.

[Peter5on 10]

Gamer: "Hey Arma u walking insted of running whats up?"

Arma: "i need more power, i need more CPU power, iem so tired, iem so old (Real Virtuality Engine)

Gamer: "ok ok, here it is i OC my cpu to 6ghz fine?"

Arma: "nahh, no i change my mind i need to rest lets stay as it is"

Its sad becouse every year cpu power rise about 10% so we need to wait about 5-10 years to have constant

60 fps? What kind of power we need?

And maybe they going to relase Arma 4 of course with the same archaic engine and yes on 32 bits...

Its 2014!!!

There is no hope...

[jumpinghubert 48]

I got 49fps average.

SYSTEM DESCRIPTION:

CPU: i5-4670k@4.5GHz

GPU: 2x HD7990@1100MHz/1575MHz/1.2V (TDP@300W, clock is stable and constant on all 4 cores)

PSU: Corsair AX1200i

MOBO: MSI Z87-GD65

RAM: Corsair Vengeance 2x 8GB DDR3 1600MHz CL9

HD: SSD Samsung 840 Evo 250GB (Steam) + 120GB (Windows)

COOLING: Water cooled with a great deal of cooling power - GPU@50°C max and CPU@60°C max

OS: Windows 8.1 64-bit

CATALYST: 14.3 Beta

SETTINGS IN GAME:

VISIBILITY: 3800/3200/100

QUALITY SETTING: VERY HIGH for everything else

FPS BENCHMARK: 49 average

---------- Post added at 17:12 ---------- Previous post was at 16:51 ----------

As comparison from Benchmark I am running 3DMark Vantage: P43709, Graphics 59171, CPU 24501.

thanks. This benchmark is very cpu-dependent (only with lots of vegetation nearby or zooming into trees or grass arma3 is gpu-heavy) so your 7990 is idling. I have similar results, see my specs below. Please reduce visibility to 3000/3000 maximum to get better results in sp. Shadow radius you can set to 200 without problems. 1600er ram slow down your haswell. Try to overclock it :)

I´ve noticed project cars have similar cpu-utilization-range, constantly 50-65% (summarized over all cores) and in only one point opposed: The more ai in project cars cpu-usage gains from 50%up to 67% max. The more ai in arma3 cpu-usage drop from 65% (in empty editor) to 45-50%. Same behavior I observed in arma2 years ago :p

The quintessence is arma3 uses most cpu-power in empty editor :D

Edited April 30, 2014 by JumpingHubert

[[frl]myke 14]

Bottom line is... the game does not deliver the Performance that Players want.

Question remains: either it's the game faults or the players fault.

But the Software optimization seems to be very poor on Arma. The fact that a much "simpler" game as BF4 is able to use more of my RAM, more of my VRAM, more of my GPU and more of my CPU... that says everything to me.

This quote says everything to me. A simpler game is easier to make use of multiple cores since close to none calculations rely on each other. Non-gameplay relevant bling-bling can easily be made to fill up your memory and use your CPU/GPU up to 100%. This wont work as soon everything relies on each other.

They have a simpler Software that makes better use of my Hardware. Therefore they deliver 175% more fps.

At the same time arma runs a ridiculus 20% GPU utilization. And the CPU utilization is jumping up and down on all cores, going as low as 20% a lot of times. If the game is so CPU oriented it should at least use all the power I have available. There is no bottleneck, because everything is under optimized.

Why the hell is my Ferrari outrunned by a Subaru Impreza on a WRC dirt track? I have more cylinders, more horsepower, the wider tires and still this Subaru goes up and away. I have to call Luca Di Montezemolo what they plan to improve to solve this.

Now, please, stop comparing things that aren't comparable.

Its sad becouse every year cpu power rise about 10% so we need to wait about 5-10 years to have constant

60 fps? What kind of power we need?

The power to understand programming. Sadly this is one the few things BI can't patch into the game.

[jiltedjock 10]

Except this analogy can be proven false simply by the way AI is handled online in that client AI is calculated Client side and server AI is calculated Server side and it's sync'd over the network, so theoretically why can't Core 1 AI be calculated on Core 1 and Core 2 AI be calculated on Core 2? Why the need for it to be single threaded I ask again? Also what about the other issue's that don't stem from AI but rather simply from the way the engine works and is programmed to work?

There is no reason at all why this couldn't have been done for Arma 3. It has been obvious since Arma 2 that one of the best ways to stabilise client performance on a multi core is to move the AI to another thread and core by running the mission on a dedicated server on the same PC. Arma 3 should have been spec'd for quad core minimum and the client changed appropriately.

Myke's posts on this are embarrassing, frankly.

[[frl]myke 14]

There is no reason at all why this couldn't have been done for Arma 3.

This sentence is so definitive, mind to explain where you got the source code from ArmA 3? And what is your CV related to programming? If you're so confident that you know what has to be done, did you applied for a programmers job at BI?

[F3r4_N3n3m 10]

Myke;2678764']Question remains: either it's the game faults or the players fault.

It is a bit arrogant to say that the players are responsible for the poor performance of the game.

I think BI needs to start developing hardware that copes with their software then. Because the best hardware available to consumers isn't close to being enough.

And one more time... Fighting with the community simply doens't help.

I'd like to see some appreciation for the engagement that the community shows here, even when that means bringing problems to evidence. Being mad does not help solve problems and it draws customers away. :j:

[windies 11]

Myke;2678805']This sentence is so definitive' date=' mind to explain where you got the source code from ArmA 3? And what is your CV related to programming? If you're so confident that you know what has to be done, did you applied for a programmers job at BI?[/quote']

Well since it's already done across a network, it's hard to imagine it can't be done in the same fashion along the crossbar switch. Same basic principle. It also shows that the load can be spread across multiple physical entities, for instance a client and a server, or the Headless client and the server and the client. Don't need a degree in programming to see when the solution is already implemented in front of you and functioning, just not in the necessary way. Seems more to me that the problem is born out of laziness or obtuseness rather than a lack of an apt workforce or ignorant players.

[[frl]myke 14]

It is a bit arrogant to say that the players are responsible for the poor performance of the game.

It might sound like arrogance but it is daily experience. Players with sub-par PC's expect it to run on Ultra because game XYZ also runs on ultra, not understanding that ArmA 3 isn't comparable with game XYZ.

Well since it's already done across a network, it's hard to imagine it can't be done in the same fashion along the crossbar switch. Same basic principle. It also shows that the load can be spread across multiple physical entities, for instance a client and a server, or the Headless client and the server and the client. Don't need a degree in programming to see when the solution is already implemented in front of you and functioning, just not in the necessary way. Seems more to me that the problem is born out of laziness or obtuseness rather than a lack of an apt workforce or ignorant players.

Uhm, weren't we're talking about inferior performance in MP compared to SP? This negates your argument by 100%. According to your argument, MP should have superior performance than SP. But i see you have your opinion and wont let confuse you with facts. Please keep arguing, i'm no longer wasting my time in here. Haters gonna hate.

[jiltedjock 10]

Myke;2678805']This sentence is so definitive' date=' mind to explain where you got the source code from ArmA 3? And what is your CV related to programming? If you're so confident that you know what has to be done, did you applied for a programmers job at BI?[/quote']

I've got 20 years working in a dev environment, but that is hardly relevant. I am fully aware of how hard it would be to make the main simulation thread multi threaded, and how little gain there might be, which is why I am not suggesting it.

The point is that the server/client (and by extension headless client) architecture that exists in Arma 2 and 3 shows that the AI can be split out to different threads and cores without causing the client simulation thread to wait for it. These threads are independent of each other. So it stands to reason that the most straightforward way to reduce the load on the client simulation is to split out the AI processing to extra client cores in the same manner. Compared to making the simulation thread multi-threaded, which I don't think will ever happen, this is a relatively straightforward win. But it would mean a quad core minimum spec.

I have played all my Arma 2 missions on a dedicated server on the same PC for the last few years. The client is much more stable this way, giving higher average FPS and very few spikes.

Edited April 30, 2014 by jiltedjock

[windies 11]

Myke;2678842']It might sound like arrogance but it is daily experience. Players with sub-par PC's expect it to run on Ultra because game XYZ also runs on ultra' date=' not understanding that ArmA 3 isn't comparable with game XYZ.

[b']Uhm, weren't we're talking about inferior performance in MP compared to SP? This negates your argument by 100%. According to your argument, MP should have superior performance than SP. But i see you have your opinion and wont let confuse you with facts. Please keep arguing, i'm no longer wasting my time in here. Haters gonna hate.[/b]

It's been done like that since ArmA and OFP has existed and those titles haven't had the issue's in multiplayer that ArmA 3 has, don't be ignorant of that fact to try to prove your strawman argument. It's clear that something added to ArmA 3 has more than likely caused the multiplayer issue's since they didn't exist prior to ArmA 3, and that isn't AI locality.

[mobile_medic 43]

Question remains: either it's the game faults or the players fault.

It has to only be one or the other? Can you seriously not imagine a situation where both of these things can be true at the same time?

It might sound like arrogance but it is daily experience. Players with sub-par PC's expect it to run on Ultra because game XYZ also runs on ultra, not understanding that ArmA 3 isn't comparable with game XYZ.

Sounds more like to me you have already reached a conclusion (it is the players' fault) and are cherry picking from an enormous amount of posts over years on this subject in order to support your pre-determined conclusion.

Sounds more like a case of confirmation bias, rather than an objective look at the wealth of posts and feedback from a wide range of members with a wide range of systems and a wide range of expectations. Denying at this point, that the game doesn't have fundamental issues seems quite laughable to me.

But, the two things are certainly not mutually exclusive.

This thread is crickets most of the time, until someone finds a user with unrealistic expectations that they can pounce on, and then act like that represents the full demographic of users with issues.

[[frl]myke 14]

Sounds more like a case of confirmation bias, rather than an objective look at the wealth of posts and feedback from a wide range of members with a wide range of systems and a wide range of expectations.

So you're ignoring those players which don't post on this forums because they don't have any problems (with the sold units in mind, those are the majority) and call my view on it biased? Go figure. Most of the time there are the very same players nagging about problems.

Denying at this point, that the game doesn't have fundamental issues seems quite laughable to me.

Does the game needs improvements? Yes.

Does it have fundamental flaws? No. Not in my opinion. Am i allowed to have an opinion?

It is simple math:

250'000 units sold (fictive number) and 250'000 customers report problems = problem with the product.

250'000 units sold and 2000 (fictive number again) report problems (which are mostly the same customers which already were unhappy with the previous products) = ???

You may figure it out yourself. I for myself, if it works at most customers but not on a few (and i know the product is exactly the same) i wouldn't suspect a fault in the product itself at first place.

People say they get 20FPS no matter what settings. First thing i do, launching the game and try it out. I get 40 FPS on high settings (VD ~3200), i lower the settings and FPS goes through the roof. What should i think there? I don't have a high end rig, my GPU is already a few years old (HD5870). Now tell me, what youd you conclude?

People don't quit saying that the game is poorly programmed and/or poorly optimized. How can they know without having the source code (and more than often no programming knowledge either).

I've seen users asking why they can't play on ultra settings and after some time i find out that this user has a 5 year old mediocre budget laptop and really tried to play on ultra...what do you say there?

Face it, ArmA 3 will never reach 150FPS because it is not Call of Battlefield 34. It is ArmA 3, there is more going on under the hood than just fancy graphics and small areas. But there will always be some customers which will be never satisfied and will always see the game as the guilty. Even if suddenly ArmA 3 would run with 150FPS on ultra, there will always be unsatisfied customers still.

[jumpinghubert 48]

It's been done like that since ArmA and OFP has existed and those titles haven't had the issue's in multiplayer that ArmA 3 has, don't be ignorant of that fact to try to prove your strawman argument. It's clear that something added to ArmA 3 has more than likely caused the multiplayer issue's since they didn't exist prior to ArmA 3, and that isn't AI locality.

in my experience arma2 mp had bigger performance problems than arma3 mp. But not easy to compare....different missions, different islands and so on. In the last months arma3-mp performance made big steps forward. But I can´t speak for all mp-modes.

[windies 11]

in my experience arma2 mp had bigger performance problems than arma3 mp. But not easy to compare....different missions, different islands and so on. In the last months arma3-mp performance made big steps forward. But I can´t speak for all mp-modes.

I can't really take you seriously with the above bolded statement. I mean ArmA 2 had bigger multiplayer performance issue's than ArmA 3? I remember playing with GOL back around that time and we had literally no issue's with multiplayer. There were inherent issue's with the performance of ArmA 2 however, but they were not multiplayer exclusive and they were not simply a cause because of the AI.

[[evo] dan 79]

I can't really take you seriously with the above bolded statement. I mean ArmA 2 had bigger multiplayer performance issue's than ArmA 3? I remember playing with GOL back around that time and we had literally no issue's with multiplayer. There were inherent issue's with the performance of ArmA 2 however, but they were not multiplayer exclusive and they were not simply a cause because of the AI.

I thinks he is probably on about the warping issue that was fixed in OA 1.58 or 1.60

[windies 11]

Myke;2679053']So you're ignoring those players which don't post on this forums because they don't have any problems (with the sold units in mind' date=' those are the majority) and call my view on it biased? Go figure. Most of the time there are the very same players nagging about problems.

Does the game needs improvements? [b']Yes[/b].

Does it have fundamental flaws? No. Not in my opinion. Am i allowed to have an opinion?

It is simple math:

250'000 units sold (fictive number) and 250'000 customers report problems = problem with the product.

250'000 units sold and 2000 (fictive number again) report problems (which are mostly the same customers which already were unhappy with the previous products) = ???

You may figure it out yourself. I for myself, if it works at most customers but not on a few (and i know the product is exactly the same) i wouldn't suspect a fault in the product itself at first place.

People say they get 20FPS no matter what settings. First thing i do, launching the game and try it out. I get 40 FPS on high settings (VD ~3200), i lower the settings and FPS goes through the roof. What should i think there? I don't have a high end rig, my GPU is already a few years old (HD5870). Now tell me, what youd you conclude?

People don't quit saying that the game is poorly programmed and/or poorly optimized. How can they know without having the source code (and more than often no programming knowledge either).

I've seen users asking why they can't play on ultra settings and after some time i find out that this user has a 5 year old mediocre budget laptop and really tried to play on ultra...what do you say there?

Face it, ArmA 3 will never reach 150FPS because it is not Call of Battlefield 34. It is ArmA 3, there is more going on under the hood than just fancy graphics and small areas. But there will always be some customers which will be never satisfied and will always see the game as the guilty. Even if suddenly ArmA 3 would run with 150FPS on ultra, there will always be unsatisfied customers still.

So in other words, there's not enough people bitching so it must not be a problem? Also to say I was unhappy with the previous product and group me into that category in your mind simply because I take issue with the fact ArmA 3 runs terribly is blatant ignorance on your part because I basically loved OFP and ArmA. ArmA 2 was a bit of a letdown for me in some aspects but overall it wasn't terrible and OA was an improvement upon A2 but again there were still core issue's at that point.

Basically what you are getting at is that people's expectations ARE the problem because we've been spoiled by games that have a certain acceptable level of performance and we're spoiled because we ask the same of ArmA. No one here want's it to run at 150fps, what we want is for an acceptable level of performance and we question why it is something like BF4 can run at 150fps yet ArmA can barely maintain 30-40 fps on same said hardware. Excuses like "Well it's the AI obviously!" and "I'm fine with 10-15 fps, I mean this is a tactical shooter not a run and gun shooter so performance doesn't matter" are basically borne of shit when if that was the case, usage would be through the roof in most cases. It's actually quite the opposite in that the software can't make use of the hardware because it's programmed for old tech and that creates a large bottleneck. It may very well be an issue with the AI, but that doesn't mean that because the AI is monolithic by nature it can never be fixed. So simply saying "Yo dawg it's the AI" is a completely ignorant and false justification as are just about any other generic justification that people throw out there like this game being a "tactical simulator" etc...

You're allowed to have opinion, but I'm also allowed to tell you your opinion is full of crap and faulty. For one, you base your standard on your own presumptions and not on industry standards or commonly accepted standards. Secondly you're so biased in your opinion, and believe me it shows beyond a shadow of a doubt that you are, that you assume that anyone with a problem is using 5 year old equipment and trying to run on ULTRA SUPER DUPER settings and then you don't even read what people post half the time. I'm thoroughly convinced at this point that you literally ad lib in these things into people posts when you read them. One guy say's he has a beast of a system and you use his performance as an argument about some guys secondary system being like 5 years old, it's like you don't even care about correctness or accuracy, simply about proving your point which is utterly wrong and faulty.

The problem is that you don't even understand the problem, you're opinion is beyond faulty and you're completely biased when it even comes to trying to understand the problem by automatically putting the cause of the problem on the end user based on the fact that you say you encounter everyone trying to play on Ultra settings on 5 year old equipment which is hardly true at all and your responses in the 300+ pages of this thread are a testament to your ignorance. You're unwilling to be objective about it by any means and even provide a modicum of fact to your opinion.

In essence, you don't have the slightest clue what you're talking about and at this point in time you need to keep quiet and stop inflaming the situation with people, because that's all you ever seem to do. I'm glad you have an opinion but it's pretty much been proven to be a biased and nonfactual opinion.

Edited April 30, 2014 by Windies

[Varanon 892]

in my experience arma2 mp had bigger performance problems than arma3 mp.

Which ones ?

[windies 11]

Dan;2679208']I thinks he is probably on about the warping issue that was fixed in OA 1.58 or 1.60

Could be but hell we have 10x+ the amount of problems in ArmA 3.

[jumpinghubert 48]

Which ones ?

sorry, I have no apple to apple comparison. But isn´t the whole malloc-thing an improvement? In the malloc thread I got this overall impression.

[Varanon 892]

sorry, I have no apple to apple comparison. But isn´t the whole malloc-thing an improvement? In the malloc thread I got this overall impression.

That was basically an all-out problem. The only MP related problem in A2 I could think of was the rubberbanding, and that wasn't really a performance related thing.

But in A3, there's a lot of performance issues that basically only pop up in MP, and not in SP.

Note, no one claims A2 was perfect and A3 is utter trash, far from it. However, I think it is totally unproductive to deny any problems when clearly a lot of people are experiencing them.

[Alwarren 2765]

in my experience arma2 mp had bigger performance problems than arma3 mp. But not easy to compare....different missions, different islands and so on. In the last months arma3-mp performance made big steps forward. But I can´t speak for all mp-modes.

This is simply not true. Granted, Arma 2 had its fair share of problems, especially in the beginning, but issues in MP were not performance-related, not any more than they were performance-related in SP.

[jumpinghubert 48]

However, I think it is totally unproductive to deny any problems when clearly a lot of people are experiencing them.

on the other hand no one deny any problems.

edit:

@Alwarren

I talked from my experience but its of course not representative.

Edited May 1, 2014 by JumpingHubert