brightcandle
-
Content Count
236 -
Joined
-
Last visited
-
Medals
Posts posted by brightcandle
-
-
On 17/06/2018 at 7:22 AM, SnakeDocc said:This sounds very cool but I'm also interested about the dynamic changing of skill settings in relation to running alongside vcom which also adjust things like accuracy sway ect on the fly during the mission.
I will look into how VCOM does this, if it adjusts skills then the two mods will likely fight to set it depending on how VCOM wrote it. I may be able to support VCOM_AI's adjustments but no promises yet.
-
1
-
-
On 16/06/2018 at 10:08 AM, domokun said:Thanks SO much.
My squad loves mods like this which improve the game's realism without breaking missions.
They improve the immersiveness so much that it's surprising that BI don't include it by default.
Any chance you could release on Steam Workshop like TPWCAS and ASR AI3?
Here you are: https://steamcommunity.com/sharedfiles/filedetails/?id=1415343823
-
1
-
-
6 minutes ago, kremator said:Thanks @brightcandle. Is this mod desgined to replace ASR_AI or VCOM_AI or perhaps run alongside them ?
Running alongside them is the most likely scenario, this isn't a replacement for those mods, it does a completely different thing. ASR_AI or VCOM_AI add different behaviours so that the AI on a micro level takes better actions, but they don't solve the imbalance in the game with woodland where to the players the AI seem to gain supervision and accuracy compared to urban or more open area combat. CF_BAI focusses on that balance aspect, it doesn't update vanilla behaviours in any way, it just relies on the vanilla mechanics for the skills and makes them change dynamically for the circumstances.
There is a bit of overlap with regards to the skill setup as CF_BAI can set the unit subskills and that is something ASR_AI does as well. ASR_AI has a much more sophisticated system that allows changing it for factions and class of unit and CF_BAI simply sets all units the same. Using fixed subskills across all units and factions while you dial in your skill settings might be beneficial to get the right average values faster before then returning to using ASR_AI and the various range of skills.
My group runs ASR_AI with CF_BAI (and TPWCAS), they are fully compatible out of the box.
-
2
-
2
-
-
Mod withdrawn, no longer available to the community.
-
8
-
3
-
2
-
-
The current flares just simply don't do the job they are meant to do anymore. The lack of lighting distance makes them more like a flying candle than a flare. We used to love them in Arma 2 but in Arma 3 we barely notice them as they have zero impact.
It would be nice from a tactical point of view to have them working as they should because currently they aren't worth using.
-
1
-
-
One thing we noticed last night in testing ACE Nightvision was that all IceBreakr terrains seem to be too light at night time compared to the base game. This means that the NVGs are actually less effective than if you don't use them, regardless of date/time set the night simply can't get dark enough, even when overcast. Can we get the night light settings updated to match the base setup of the game?
-
15 hours ago, zooloo75 said:Please post the logs here.
Its literally 900KB of:
21:07:46 Obsolete
Its hard to miss as it does about 50 of those a second.
-
1
-
-
JSRS Version 5.17.0807
with the RHS addon as well
I am seeing quite a lot of RPT spam on the client during combat saying the following:
21:34:12 Sound: Error: File: jsrs_soundmod_p_weaponsounds\noises\mg_chain_clipper (5).ogg not found !!!
The number in the () at the end ranges from 1 to 10 as far as I can see.
-
Bloodlust: v2.484
Scenario - Lock server game with around 12 v 12 AI combat in a small area.
I was testing this today and found it spams the RPT full of the message "Obsolete". After a while I found that the client would hit the 5 fps bug very rapidly (64 bit) so something is going wrong. I had been testing it with a group of other mods but eventually narrowed it down to bloodlust.
-
The sinking in the ground to attempt to compensate for the lack of grass rendering is just as big a concern really. It makes marksmanship a lot harder since the feet disappear into the ground even when there is no grass and does little to aid visibility reduction. A better solution using shaders to blend them into the background would be very welcome at this point, but also extremely unlikely.
-
1
-
-
Probably the best cheap option is to see if someone is selling something like a Ivy Bridge or Haswell i5 processor that is overclockable and get an older second hand platform. This will allow you to save a bunch of cash and still get a decent i5/i7 that you can overclock and get decent performance.
All the tests I have seen shown 4 cores is the sweet spot. The game is mostly single threaded but the parallel parts really matter as well so you can't just get a dual core and get great performance. Look into those second hand options and you might find a deal you can live with price wise that gets you a better CPU. Arma isn't easy to run, its going to struggle on a 1050 ti as well, simulation games in general hammer a lot of things.
-
2
-
-
Bohemia has done what they can for the game at this point, we aren't going to see any improvements. Unlike its predecessor this really is a sub 30 player game if you want playable framerates. Playable in BI's own words is 30 fps, that is their target. You can absolutely sustain 30 fps with decent hardware and below 30 players and about 60 AI and not too many scripts running. Arma 2 however we regularly had 60-90 player games with a lot of AI, so its disappointing to see the sequel perform so much worse, but BI continued to rely on the CPU core getting faster which it really hasn't done.
They are niche developer, they certainly aren't AAA and performance is clearly a tertiary consideration for them. Its been the number 1 request feature/bug fix since the very first alpha of the game and it'll be the number 1 problem when its obsoleted by Arma 4. The fix for the performance problems of Arma 2 was actually promised for Arma 3 but none of it materialised. I don't know if they will learn and do it differently for Arma 4 considering just how many sales Arma 3 has had. It seems no one cares enough that the performance is awful, its not hurting the growth of their sales for the series anyway.
Like everything else the feedback goes in the same skip out back marked "community requests".
-
2
-
-
Now the spotting distance of troops has been reduced to like 400m maximum (with a subskill for spotting of 1.0). Beyond that they are blind like they have a fog past which they can't see. Its kind of scary that this sort of tweak made it in and so few have noticed that you can walk around in the open with no consequences at this sort of range.
-
I am not so sure about that...I can imagine that BI will keep Arma 3 as a platform for a very long time and will merge the Enfusion tech in the existing engine step by step (as far as possible). This is something they will have to do for future Arma projects anyway and would make more sense for me than starting a complete new Arma game. I mean there would be even no need for creating complete new maps, they could easily do a Chernarus + version for A 3 once DayZ has gone final and/or use some of the older and smaller maps like Zargabad, Utes, Proving Ground etc. for Project Argo. The other way round they could use especially Altis for a DayZ dlc.
just my thoughts :)
BI has already said quite a few times that they don't intend to change Arma 3 performance significantly over the life of the product. They have no plans to overhaul the games engine whatsoever. Performance will stay about the same for the rest of the games playable lifetime, that is BI's stated intentions.
-
JSRS is a classic example of how annoying the relationship for the community can be. LordJarhead had his first release of JSRS for Arma 3 ready and releases it and 1 week later he gets completely blind-sided by a massive breaking audio shake-up that destroys a lot of his work and makes the mod no longer work. Given how important JSRS has been over the longer term for the series you would have thought that BI would have given him a heads up about the change coming, but for "marketing" reasons they chose to hide its existence from everybody including the modders they knew would have their work destroyed. We very nearly didn't get a JSRS release at all and it took a long while for it to come out.
That isn't right and every single mod author is seeing the same thing. Every community reporting bugs is seeing no progress on the problems they face. The performance problem with the game has been the number 1 priority since the very first alpha, and honestly we have seen little to no progress on it in all these years. BI does what it wants when it wants to and breaking changes are usually secrets until its too late and it breaks a lot of stuff. Visibility is all mod authors really need, that and an ability to hold onto older releases for longer so they can choose when they deal with the updates (and so communities can fix the game on an old version). We have a dependency graph we need to ensure we can control the most unstable part the game itself.
IMO Bohemia Interactive really needs to think long and hard about why every release is getting a hotfix. Often these bugs would have been obvious with a very simple initial test and if they were in a RC build the community would have reported the problems. They just don't seem to appreciate the point of having the community test stuff, this isn't a marketing strategy and if you treat it as one then the community becomes jaded and ignore it. We should not be receiving a game patch that has a breaking change not present in an RC at this point.. Its an abusive relationship, they treat people who are helping their game sell like dirt and its all unnecessary.
-
2
-
-
Having played and modded and run a community since Arma 2 I have had a lot of cause to provide feedback to BI. But so far I have found it to be a mostly abusive relationship where they want to market (their sitereps etc) their progress or want me to test something, but bug fixes or improvements we need? Forget about it. So there isn't much to say or do at this point, they want their thing and I want mine and there just isn't any point banging on about it because I know they are never going to do anything about what I want or need and hence what my community is suffering through. I think of the amount of bugs people still run into that I reported years ago its just downright irritating how BI prioritises things.
So the lack of participation is because while on the surface it looks like they are great company doing good by the community in practice what actually gets done is work that already aligns with their current plans.
-
8
-
-
Multiple times in MOUT training we have seen the AI actively track the team through an alley way through the actual buildings. We have had issues with grass where they appear to not care it exists but see straight to you and turning if off makes it obvious that they can see you. But other than that it seems to work OK.
-
Hmm... I have a 5820K which scores higher than the 6700K in CPU Benchmark, but it seems the 6700K is quicker in single core tests. So I guess this means Arma runs better with a 6700K.
http://www.hardware.fr/getgraphimg.php?id=223&n=14
Significantly better. The enterprise chips actually get hurt rather than helped and I think that is because they have a different design for the memory controller designed to spread out memory bandwidth across cores whereas Arma is heavily dependent on its main core thread.
-
When we talk about a new game engine in which to build the series what we are really talking about is using Unity or Unreal Engine, because those are the two remaining commercial engines out there. Neither provides a good base for this type of game although neither is awful either as they are general purpose game engines. But the real issue with a wholesale replacement engine is you start from nothing, you can't carry across anything you have currently. Arma/A2/A3 are all built on top of each other and there is a lot to the games code and unique rendering qualities. A new engine would mean none of today's mods could be ported, none of the content would work nor would any of the scripting and BI has a lot of features in their game. The redevelopment cost into a whole new engine would be enormous, potentially a decade or more and what would that achieve? Mostly it would just mean loosing a bucket load of time rewriting everything when the problem is with just a couple of percent of what they have.
The underlying issue in the Arma problem revolves around two key areas. The simulation for world updates is mostly serial and based quite a lot in SQF and the rendering engine itself is mostly single threaded but also still contains a lot of effects algorithms running on the CPU (smoke for example). Neither is easy to fix but changing the guarantees around the SQF and world updates in general combined with a more parallel scripting engine would produce vast improvements in performance (its around 1/3 of a frame currently) and require a lot less rewriting but would still require a chunk of change. The rendering engine on the other hand mostly just needs more of the CPU activities moved onto the GPU and porting to something like DX12 along with quite a substantial change in how it parallelises.
So while so many people are advocating that BI just throw what they have away and start again that is a ridiculously bad thing to do if you want an Arma 4 game any time soon. They have a world class engine of their own with a list of features 99.9% of games don't come close to matching currently but with some performance issues. When you have performance issues you don't just throw everything away unless everything is the problem, and in this case I highly doubt that is the case. What happened is they predicted poorly where the CPU market was going and lost that bet, but now given the issues they have had I suspect they can put the appropriate resources into evolving the situation for Arma 4. Adding parallelism and performance isn't easy but its much easier than rewriting the game from scratch. If each patch doesn't have to worry about a working game that doesn't break most of the existing user content and mods then its possible to at least change the situation in a future big release. I believe it can be fixed having looked at how it works from the DirectX, ingame and other profilers.
However until BI actually commits to fixing performance its not going to happen. We have had the same song and dance about performance in Arma 3 since Alpha, "we are working on it". Back in Alpha we determined the multiplayer limit was about 30 and the AI limit was about 150 to sustain reasonable frame rate and that hasn't changed. On a fundamental level the main thing that changed was 3 generations of Intel CPUs and they do improve performance in this game (where they don't in most others). Just like with Arma 2 it runs better on modern hardware not available when the game was released. I don't think that is the way this should work but you have to also remember the developers are aiming for 30 fps and they are achieving that quite well on their minimum and recommended hardware. When I take a diag_captureFrame and compare it to the ones I did in December after release in the same scenarios they are basically identical. People keep saying performance has changed but I can't see that in the actual objective data on the same CPU. There has been releases that decreased performance substantially but none that improved it past the base line I took all those years ago. Until BI commits to fixing it and we hold them to it they will continue to not bother, after all they got your money already they don't need to do anything more at this point.
They said before Arma 3 came out that they had a solution for parallel behaviour, I remember seeing the design for which the headless client was the first iteration. They abandoned it in order to ship the game. They said they were investigating DX12 (well I kind of said that wouldn't help) and they abandoned that. Every release they tell us performance has improved, they release "performance" binaries that seem to only fix bugs. There is a pattern emerging and the last time people got irritated they rolled out Dwarden to tell us it was all going to be OK, but has it been? No it hasn't been its the same story as Arma 2. You want something to change its going to take the community to actually mobilise. Throwing out their engine isn't the solution we just need them to actually do what they say and work on genuinely fixing performance and we will deal with the breaking mods and such that happens as a result. But hand waving "new engine" just gives BI the pass to keep not doing it and it'll be the same in the next game.
-
3
-
-
While I really want to try Arma 3 on my Vive I suspect it will never be able to deliver 90 fps at HMD resolutions in even the simplest scenarios with dual eye rendering, so adding support is kind of pointless as it will never be playable on a HMD. Whether BI see the value in making it possible for VR gaming to occur at least in cockpit mode for future games remains to be seen, it will vastly improve the perception of reality and provide mechanism and controls we don't have but its also a fundamentally different game completely.
-
I used AMD cards back in the Arma 2 days and they had this one bug where you would get white snow where the ATOC should be, you had to turn that setting off. It was like that years just broken. I used dual cards and there were these places in the map where crossfire scaling would just disappear and the frame rate would stutter like crazy. Then Arma 3 came out and I had the same issues. I finally gave up on the 7970 and used 2x680 and all my graphical glitches disappeared, my performance improved and the scaling was consistent. Considering the ATOC bug was there for years I have to blame AMD for that because it was a bug that they finally fixed for Arma 2 and Arma 3 at the same time when they finally got around to do something for Arma 3. Its just not a game on AMD's radar, they aren't doing much to help it along.
If you go look on gamegpu.ru and go through a tonne of games you see this trend where the common games all have the cards performing where you might expect with the usual back and forth variances. But then you look at the rarer and less popular games or the old ones and Nvidia is beating AMD decisively. There are too many "did not work" results in those less common games. As far as I can tell on the grand meta AMD focussed on particular games to save resources, they don't generally produce good performant drivers they produce optimised drivers for popular games. Arma 3 isn't on their radar, its not benchmarked enough to impact their comparison on most major websites and hence I suspect their support is largely abandoned.
The other possibility is that BI is doing a tonne to hurt AMD performance and is refusing to cooperate with them, but considering how often this happens in other less popular games too it would seem like it remains an AMD problem. AMD has had this problem for a decade, its been a consistent issue which up until recently has always been blamed on AMD, now for some reason its been framed as everyones problem (the game devs are shilling, Nvidia is hurting their performance with gameworks etc etc), and yet all the same evidence of issues on the less played games is still there.
Regardless the answer is the same, you want an Intel CPU with the highest clockspeed and latest architecture (not an -E chip they get "hurt" by their memory controller we think) and you want an Nvidia GPU. That is the best combination to actually play the game. Whoever's fault it is doesn't matter because the game runs poorly enough that if you don't build a machine with the game in mind you will get worse performance as a result.
-
We tried to run this mod with MCC by disabling its waypoint behaviour and it really didn't work. The units didn't get any waypoints initially from either mod and when combat started all units got the exact same waypoint after being "called for help" and produce a big mud ball. The AI moved a lot but they really didn't shoot a lot. I would say the two can't be used in any way.
-
We reported a few bugs and were asked to produce a video so here is our look at the 3 problem areas with the new release. The new features don't really work very well, I say it a lot in the video but its just weird, strange and inconsistent. Hopefully now we are reporting this in this form you are more interested in the bug report as the rejection and closes before were pretty unhelpful.
-
1
-
-
My experience is that SLI works pretty well in the game actually. I could max out 2x 970's on both Altis and Tanoa, the game is GPU limited given decent AA settings inside of the forests, its still CPU limited outside of the forests of course but you can be limited to 75 fps with a lot of trees on screen with 2x 970's.
So I think personally the current ideal Arma machine to get mostly ~60 fps in moderate multiplayer (sub 30 players) is a Skylake 6700k and a 1080. 2x 1070 is probably for Arma's sake a lot faster and around the same price but its obviously less good in other games. One thing I did notice however going from 2x 970's to a single 1080 was a sizeable reduction in input latency, Arma 3 has quite high latency and on dual cards its very noticeable but a single card is quite a bit snappier. But the game scales well with SLI and that higher FPS in the places where SLI helps is nice also.
CF_BAI - Dynamic AI skill adjustment for better balancing
in ARMA 3 - ADDONS & MODS: COMPLETE
Posted
Maybe I am not finding the latest code base but according to https://github.com/genesis92x/VcomAI-3.0
it appears that VCOM AI uses a similar mechanism to CF_BAI for setting the units base skills (and indeed similar to ASR_AI although ASR_AI remains the most sophisticated) and doesn't do any dynamic updates after that initial setup. So VCOM_AI is compatible in the same way that ASR_AI is, either use VCOM settings to set the AI subskills, in which case you should leave CF_BAI maximums on 0.00 (the default) or just use CF_BAI's maximum skill settings which are a bit easier to tweak since they are in a nice menu and can be updated mid game if need be.