Performance is bad again, BIS enable us to help you

[brightcandle 114]

A long long time ago my community and I produced profiler traces of multiplayer games and showed two types of serious CPU limitations in the game that contributed to performance problems. We had builds that were enabled with diag_captureFrame that allowed us to capture the profiles that the developers were asking us for and we captured a lot of data (http://forums.bistudio.com/showthread.php?169484-What-I-found-about-performance-and-I-what-I-want-to-help-fix-it). Having captured the data and provided it as request we haven't heard a thing since. Performance hasn't changed either, if anything its gotten worse over time.

Given that people are still concerned about performance and given what we determined before I propose once again we look into getting profiles from real multiplayer games with highly modded gameplay so that you can look into the major contributors to poor performance and do something about. I can't say I am entirely happy with the radio silence last time so a condition of us helping again would be to be told what issues you find and the list of bugs/features you have added so we can track as and when they get improved. But neitherless I know after we did this last time there was a notable jump in performance and I think its time to do it again now that performance has become unacceptable again.

We need you to provide a diag_captureFrame call in the development branch again unless you now have a better approach. Its a pain for us to do these, takes a lot of effort to get everyone to download and install development branch and test all the mods and such but I want my Arma 3 game to run at the very least above 30 and mostly 60 fps and I am not finding 45 at mission start and 25 in mission to be acceptable where not more than a few months ago that was 90 and 50. So please let me know when we can do this exercise again.

[havatan19 10]

yeah i agree, its since OFP, since over 10 years, something should happen now, we are in 2014 - near 2015.

[.kju 3245]

1) dev branch finally has some diag commands back - however dev branch has no MP play

2) BI decided to go for mainly have their own tester servers with logging and do a broad analysis on these, as well as to lesser degree logs provided by server admins using the performance binary builds

as far as i can tell they went for this approach mainly as they think it costs them less time to do it this way

I think the cooperation with the CIT during A2 showed otherwise; however most people no longer care to provide that dedicated reports and cooperation after BI dropped the CIT and its community

in addition see A3 FT 12658: Give modders the tools to identify and avoid/solve performance bottlenecks

to their credit the A3 team has provide some of these to some degree, however i believe more elaborate commands and approaches are needed like some outlined in that ticket

the planetside 2 videos about their performance hunt going on over couple of months shows some elements and level of dedicated for the topic what would be need

[nikiforos 450]

I just tested the latest dev build and performance is better than 1.34 stable build :)

[Tyl3r99 41]

I just tested the latest dev build and performance is better than 1.34 stable build :)

by how much?

before and after?

[nikiforos 450]

by how much?

before and after?

Hello this is based on Helo's Benchmark on Altis.

1.32 80 FPS

1.34 75 FPS

Latest dev build 80-81 FPS

So for me we talking about 5-6 FPS more than 1.34 stable version.

I would like to know what they changed though :)

[Sterlingarcherz101 15]

Hello this is based on Helo's Benchmark on Altis.

1.32 80 FPS

1.34 75 FPS

Latest dev build 80-81 FPS

So for me we talking about 5-6 FPS more than 1.34 stable version.

I would like to know what they changed though :)

Niki had you tested the dev branch before 1.34 actual release?

---------- Post added at 21:47 ---------- Previous post was at 21:41 ----------

My tests so far have seen.

Single player stable 1.34 an average 5 fps loss.

On multiplayer, fps takes a back seat to big increase in dysnc, across a range of mission types and reported by all players involved playing at the time.

Really hope the performance going back to 132 levels translates into the stable release.

[chrisb 196]

So many great people out there playing this series, with huge amounts of talent in certain areas, would help BI. But in the real world, it is the same with anything, be it a game or any other form of creative process or anything. There is always someone out there, that is better at doing the same thing.

Doesn't mean the two shall ever meet, unfortunately.

Or indeed want to meet..;)

[nikiforos 450]

Niki had you tested the dev branch before 1.34 actual release?

---------- Post added at 21:47 ---------- Previous post was at 21:41 ----------

My tests so far have seen.

Single player stable 1.34 an average 5 fps loss.

On multiplayer, fps takes a back seat to big increase in dysnc, across a range of mission types and reported by all players involved playing at the time.

Really hope the performance going back to 132 levels translates into the stable release.

No I played with 1.32 stable , then moved on to 1.34 stable (bad frames) and now I moved to dev with much better FPS.

Sorry guys , forgot to mention that all my tests are made in single player. Haven't tested multiplayer at all.

[brightcandle 114]

So officially then BIS has stopped trying to improve performance then. They aren't willing to do anything to find the problems and fix them as is clearly apparent. Offer is now withdrawn.

[strangere 2]

So officially then BIS has stopped trying to improve performance then. They aren't willing to do anything to find the problems and fix them as is clearly apparent. Offer is now withdrawn.

Are you serious? They are testing new stuff right now http://forums.bistudio.com/showthread.php?183653-Arma-3-%28release-candidate%29-1-36-128482-stress-test-MP-%28MultiPlayer%29&p=2829507#post2829507

[brightcandle 114]

Are you serious? They are testing new stuff right now http://forums.bistudio.com/showthread.php?183653-Arma-3-%28release-candidate%29-1-36-128482-stress-test-MP-%28MultiPlayer%29&p=2829507#post2829507

Its not just about their scenarios, its about real world games. The problem has always been that performance in some particular circumstance can be made fine but real games use mods, scripts and different combinations of AI. Simply put the results of the program to improve performance using BIS servers has netted zero benefit for most groups out there. They can improve a few public games running limited scenarios but not real missions. Its simply not sufficient what they are doing and having done it for over a year the benefits have been nonexistent. Its not like this testing just started, it was happening during the beta. Frame rates were actually higher then than they are now.

BIS do this to show they are doing something about performance when they actually have no intention to fix it. Its just a ruse.

[fn_Quiksilver 1636]

Its not just about their scenarios, its about real world games. The problem has always been that performance in some particular circumstance can be made fine but real games use mods, scripts and different combinations of AI. Simply put the results of the program to improve performance using BIS servers has netted zero benefit for most groups out there. They can improve a few public games running limited scenarios but not real missions. Its simply not sufficient what they are doing and having done it for over a year the benefits have been nonexistent. Its not like this testing just started, it was happening during the beta. Frame rates were actually higher then than they are now.

BIS do this to show they are doing something about performance when they actually have no intention to fix it. Its just a ruse.

Relax. :)

It's just a game.

Quips aside, performance will get noticeably better in the .36 and .38 updates. It was better in past updates, I believe 1.24 and 1.26 were quite good performance, relatively speaking.

Aside from that, I'm sure most of us would appreciate more specifics and less vague musings. For instance:

They can improve a few public games running limited scenarios but not real missions.

There is nothing in here with which any of us can choose to agree or disagree with. No one knows what these limited scenarios or 'real missions/'real world games' are, that you speak of. If we knew, we could perhaps have a fruitful discussion.

There are two main noticeable performance gains on the way:

1. Re-gaining prior client FPS standards (1.24 standard). This is wizard stuff, not sure how it will happen.

2. Headless Client (HC). Having tested extensively HC code in MP, it does make a noticeable improvement to both server and client FPS in AI heavy scenarios such as Invade & Annex, Domination, Insurgency, Patrol Ops type scenarios. In full 40/40 session I was able to raise server FPS by ~15 and client FPS by ~10 by using all the connected clients as pseudo-HCs, while adding 50% the number of AI (from 150 to 225 units). From 150 AI only on the server, down to 75 AI on the server (results in nice and really noticeable server and client FPS gain) + 150 AI spread evenly (roughly a 4-man fireteam per client) among the client CPUs (no noticeable downward effect on client or server FPS).

The only downsides:

- The ability for AI to easily coordinate and communicate with one another. For instance if client 6 and client 14 were each the owner of a squad in the same small area, sometimes 2 units (1 from each client AI) would hide behind the same tree as if they didn't realize each other were there.

- Increased network traffic (which does have some downward client FPS effect).

[windies 11]

I find it odd that their idea of parallelizing the AI is to seed it across multiple machines rather than the multiple cores already in the machine but sorry to say that's typical BI. I wouldn't exactly say that's an advancement, even if it might be an advancement from where we currently are in terms of multiplayer performance. It's another band-aid on a bullet wound more or less, like the file mapping API was a band-aid for the limitations of 32 bit addressing instead of coding 64 bit binaries. Either way it's work, but one way is work towards a better future while the other way is work towards a quick fix that will end up biting you in the ass later on.

Ultimately the HC should be transparent to the end user and even the server administrator, it shouldn't actually be a "thing" but simply the way the game calculates the routines for the AI natively. Yet again though it's a burden thrust on to the end user to do for themselves while BI can focus on something else and it's going to end up being a bad implementation for the reasons you mentioned, increased network traffic and the AI being isolated from one another because of the medium you are using for parallelizing the AI, a network connection rather than say the internal memory bus and crossbar of the 4-12 cores in the desktop/server.

It's getting pretty moot arguing or discussing it at this point though, it's obvious which way the wind blows. They might as well stop worrying about their engine and just keep tacking on extra additions until the game ceases to function properly and then they might try to fix it properly, but somehow I doubt it and it would just be more band-aid fixes. Things like Jay Crowe's comment at E3 about the PC being the platform for them because of the power of it come to mind and yet they don't even seem to try or care enough to harness that available power and use it properly. The irony of DayZ going to the PS4 and how they will probably actually spend time getting it to run properly on basically an 8 core PC more or less, yet when it comes to ArmA it's apparently not worth the time or effort. I'm sorry to say but it's pretty sad to see from a company I once actually really respected. It's really no different than the Battlefield or CoD franchises at this point that this community loves to shake their superiority stick at.