Jump to content
Sign in to follow this  
Dwarden

ARMA 2: OA beta build 86055

Recommended Posts

Please report issues with details and RPT/Crash Dump files where applicable; http://dev-heaven.net/projects/cis/wiki/CIT

BI is also looking for feedback about the various memory allocator options; http://forums.bistudio.com/showthread.php?t=126493

Also changed:

Edited by Sickboy

Share this post


Link to post
Share on other sites

C'mon give people a brief and understandable description of using certain memory allocators. What do those memory allocators do and what not? Try it in no code-nerd-programmers-5p3ak and with no links!!! KISS ;):)

Share this post


Link to post
Share on other sites

Maybe a stupid Comparsion, but...do you know Tetris?

While a performant Memory Allocator will always try to put the Bricks (Data) in a good Order in your Field (RAM),

a bad or not so performant Memory Allocator can´t, what leads to Stress (stucking Data Flow). :D

Share this post


Link to post
Share on other sites
C'mon give people a brief and understandable description of using certain memory allocators. What do those memory allocators do and what not?

This is what an allocator does from a user perspective:

Some allocators can give you higher fps in some situations. Some allocators may cause crashes in some situations. Some allocators may cause the game to use more memory from the system in some situations. Your mileage may vary.

Share this post


Link to post
Share on other sites

Is it not time to make this a full patch to have a new foundation to work on.

There will still be things to fix even after 1.60. The eternal beta state we are stuck in now does not really help, is especially problematic in the few mp servers that are left from the better times of arma OA

currently the A2 CTI servers are more alive than the OA CTI servers.

Share this post


Link to post
Share on other sites

A full patch which brings (serious) new problems is also not in anybody's best interest, it's a fine line ;)

In the meantime imo people could use the beta patches even in MP - there are also linux betas out but not as up2date as the windows ones.

Still, of course it would be nice to have the v1.60 full patch finally :)

Edited by Sickboy

Share this post


Link to post
Share on other sites

Thanks for the new beta. Will test now. I don't mind more and more betas as I want 1.60 to be bug free.

Share this post


Link to post
Share on other sites

86055 performs very well on my Sys. Please have a look at the Malloc Thread for more Detail. :)

Share this post


Link to post
Share on other sites

How about some specific memory allocator benchmark features/missions so people can just see what changes by using "-malloc=..." parameter? Perhaps even something like "stress testing" to see clearly the limits of different systems and game settings for everyone?

Share this post


Link to post
Share on other sites

kju benchmark would do for the above. besides, i would use the malloc thread about such topics

Share this post


Link to post
Share on other sites

I really like the AI results in 1.60. Playing Laser Show is always a great experience in this area.

I played it on expert with a teammate and I tried to attack a group of four from the side so he can silently attack from the front.

Anyways I thought killing four AI players with an MG laying on the ground should be no problem but they reacted nearly like humans so that after killing two I got killed.

My teammate was able to kill a bunch of them from the distance but he was attacked by a group of four ai from behind.

One of this group always secured the back and they spread out to attack him. He was able to kill one or two but still had no chance.

Maybe it was random but it was still really impressive. Felt more like a fight against humans and not AI.

I know that the AI has still sometimes issues but at the same time they can do amazing stuff. And there are enough humans in MP games which are not good either. :D

The better AI in combination with the more realistic soldier damage system does a great job to create an immersive atmosphere.

Long text short, great job! :yay:

Edited by Black Russian

Share this post


Link to post
Share on other sites

about allocators read my answers

http://forums.bistudio.com/showpost.php?p=2045005&postcount=3

http://forums.bistudio.com/showpost.php?p=2047381&postcount=51

such discrepancy in results is possible

performance of allocators often show noticeable difference between

type of OS used (XP, Vista, Windows 7, Windows 8) and x32 vs x64

plenty of memory (4-dozens GB) vs less memory (<4GB) (as some allocators are more greedy than others)

CPU used (generation, brand, instructions available, cache, with or w/o memory controller etc.)

number of physical cores, cpus, additional threading bonus

etc.

to get meaning why result with some allocators results differs ...

anyway there is 'default' which was stable enough and fast enough for quite long time

while most so far claim the Intel TBB 4 one is better than v3, some say it's not and v3 is better ...

i myself would like to see TBB V4 as default tho

Edited by Dwarden

Share this post


Link to post
Share on other sites
I myself would like to see TBB V4 as default tho

Me too and I have a rather old system that marks the lower end of the "still playable" range albeit I have friend running the game on even older and less powerfull hardware in a still playable manner. Playable can be used as a synonym to enjoyable here....back in the late 90's we used to call 15fps in Janes F-15 or Falcon 4.0 playable, but that was another era.

Today it means around 25 fps minimum.

Edited by Beagle

Share this post


Link to post
Share on other sites
This is what an allocator does from a user perspective:

Some allocators can give you higher fps in some situations. Some allocators may cause crashes in some situations. Some allocators may cause the game to use more memory from the system in some situations. Your mileage may vary.

I'd rather hear it from a developers point a view. Not a 'programmers' point of view, but an application point of view. A realistic, technically based explanation on the differences between each and why they're included and what each should possibly do.

Share this post


Link to post
Share on other sites

I'd much rather have you figure that out yourself (/other community members) and the devs free to focus on non trivial matters ;D

I suspect that is one of the reasons why the memory allocators can be changed :D

Share this post


Link to post
Share on other sites

There is some fairly generic info about memory allocators on the internet (example), but I'm not sure anyone has actually gone to the trouble of making a direct comparison of several mallocs while explaining exactly how they differ and why.

From what I can tell, an important goal when creating a memory allocator is to balance read/write speed against memory overhead, amongst other considerations.

Share this post


Link to post
Share on other sites

Your test is flawed:

1. Your two units spawn in direct line of sight.

2. endless calculations

3. AI shoots player

3. NVG removal script is executed

Place the two units like this:

o->
---------WALL------------
o->

And walk around the wall - the NVG-free unit won't see you unless you literally bump into it.

BIS dev please check this CIT, I have found a very interesting and terrible issue for arma2 vanilla soldier in CO.

http://dev-heaven.net/issues/26117

This issue will cause many stealth action failure only if you won't use arma2 vanilla units.

Here is a demo video

http://www.youtube.com/watch?v=I7V_Ave37Bs

Share this post


Link to post
Share on other sites
I'd rather hear it from a developers point a view. Not a 'programmers' point of view, but an application point of view. A realistic, technically based explanation on the differences between each and why they're included and what each should possibly do.

As a developer, this is even my point of view. I gave up trying to understand inner workings of memory allocators, because it is too much complex area for me. I prefer "evidence based" approach - the allocator is almost a black box for me, and the one giving best results in empirical testing is the one I will prefer. The allocator crashing often is one I will not use. If you want to understand the differences, be prepared to face some difficult science.

Related: I also no longer have an ambition to create my own allocator from the scratch. This might have sense in OFP/Arma times, but those days are gone.

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  

×