General Discussion (dev branch)

[Dwarden 1125]

sorry for the incorrect details on the enableHT I was misinformed {all blame onto dazh :)}

[oldy41 61]

@Dazhbog: I tried just using -enableHT in the commandline and not defining the CPUcount, but I got the same results (no activity in the logical cores).

So then I went and checked the resource monitor and noticed they were all parked. I made some edits to the registry to unpark them and now Arma uses all cores.

Performance seems overall better and scripts are running faster.

AFAIK parked virtual cores become enabled automatically by the OS as soon as processor usage exceeds a certain threshold. So it should not be necessary to disable the parking algorithm of Windows to benefit from hyperthreading. But that's the theory.

So could you provide some measurements with enabled and disabled virtual core parking?

[das attorney 858]

Could you share the registry edits, please ?

I added a post in here with some explanation and links:

http://forums.bistudio.com/showthread.php?167746-Are-You-re-CPU-Cores-Unparked-You-d-Be-Suprised-At-The-Answer&p=2694800&viewfull=1#post2694800

So we don't derail this thread :)

@Dwarden - No worries, I like the way it's going - keep it up! :)

EDIT: @Oldy - All I can offer in the way of evidence is the empirical testing I've done. Before I unparked them, the usage on logical cores was next to 0. See here:

http://forums.bistudio.com/showthread.php?152866-General-Discussion-(dev-branch)&p=2693909&viewfull=1#post2693909

But then since unparking them, I'm getting regular usage on them (admittedly low). See this picture:

http://forums.bistudio.com/showthread.php?152866-General-Discussion-(dev-branch)&p=2694744&viewfull=1#post2694744

I'm not saying it works for all, but it has definitely made a difference on my machine. YMMV of course! :)

Edited May 21, 2014 by Das Attorney

[samb 15]

"ENGINE

Added: Support for RotorLib in the options"

What does it mean? Could You please explain guys? Thanks!

[chortles 263]

"ENGINE

Added: Support for RotorLib in the options"

What does it mean? Could You please explain guys? Thanks!

From two pages ago:

Yes, we're preparing the UI and other parts for the arrival of RTD, but the data is not yet in place. It will appear well ahead of the DLC for the existing copters though.

[Anachoretes 10]

I played with CPU monitor on the second screen and found that 2600K use 8 threads(25-65%time overall depends of scenarios or mode).

(But with any flags)

Edited May 22, 2014 by Anachoretes

[DnA 5155]

So, a little suggestion, but would it be possible to do the same with Fixed wing? When making full G turns, have the player experience very slight shaking as it does in real life due to wind resistance on the aircraft in High G turns? Well, it's very slight shaking, but it looks more like buckling, if you know what i mean...

Yes, we will not just do this for helicopters, but other aircraft ... and then other vehicle categories as well. For ground vehicles the implementation is slightly different (other forces playing a bigger role like terrain roughness). We are currently on a data lock for an upcoming update, after which the configuration of vehicles begins.

[ericfr 0]

No improvement with -enableHT, same as without this startup parameter:

1 random logical core 100%

1 random................... 65%

1 random................... 25%

5 random................... 10%

A firefight between 60AI

I7-3770

Gtx660

[Tonci87 163]

Yes, we will not just do this for helicopters, but other aircraft ... and then other vehicle categories as well. For ground vehicles the implementation is slightly different (other forces playing a bigger role like terrain roughness). We are currently on a data lock for an upcoming update, after which the configuration of vehicles begins.

So driving on a dirt road would actually make the camera shake? Nice idea. For full immersion you need to add a few sounds though. You know, suspensers and gravel hitting the wheel arch and stuff like that.

[Fushko 59]

Goddamn autotests.

[NoPOW 59]

On current DEV: After finishing campaign episode 'Survive: Blackfoot down', I receive a script error and the game halts loading the next mission (Maxwell)... Can someone confirm this?

'...

File A3\Functions_f\Arrays\fn_sortBy.sqf, line 131

95

_sortedArray set [count _sortedArray,|#|_selectedItem];

_sortedValues set [count...'

Error undefined variable in expression: _selectedItem

EDIT: After a while the mission does start, but any custom load out is not being transferred to 'Death Valley'.

EDIT2: Happens every time when entering Maxwell...

Edited May 22, 2014 by NoPOW

[Brisse 78]

So driving on a dirt road would actually make the camera shake? Nice idea. For full immersion you need to add a few sounds though. You know, suspensers and gravel hitting the wheel arch and stuff like that.

Don't we already have that?

[Sniperwolf572 758]

22-05-2014

No update today - failed auto-tests

Sounds like "I broke the build" money jar is receiving a sudden influx of funding. :D

[k0rd 3]

Sounds like "I broke the build" money jar is receiving a sudden influx of funding. :D

You should add "automatic passing of auto-tests" to your sig img :)

[xendance 3]

Goddamn autotests.

Shame on BI for having automated testing!

[DancZer 65]

@Dazhbog: I tried just using -enableHT in the commandline and not defining the CPUcount, but I got the same results (no activity in the logical cores)

Here is mine. I didn't noticed an FPS increase, but i think it perform better(used better LOD or sth)

https://dl.dropboxusercontent.com/u/71026113/LinkedContent/Arma3/-enableHT.png (327 kB)

[CaptainAzimuth 714]

Shame on BI for having automated testing!

You would have thought with AI needing improvement, they would've ruled out leaving their work for Artificial Intelligence to do the testing for them. =P

[k0rd 3]

on a serious note - the branch feels good now :) hoping for a push to stable soon :)

[DnA 5155]

If all goes well, 1.20 will be on main branch next week.

[dna_uk 30]

Fixed: Admin commands "kick" and "exec ban" not working properly

I'm fairly sure "exec ban" wasn't working, but what was wrong with "kick" before?

[DnA 5155]

I'm fairly sure "exec ban" wasn't working, but what was wrong with "kick" before?

The info from our programmers: "Kick: The game was checking input only against player nick and player number (UID wasn't checked). Ban: The game only considered a player number as a valid input. In order to make bans work with player UIDs and player nicks, the support for string input needed to be implemented, e.g. exec ban 'nick' or exec ban '12345' does the trick now."

"DnA_UK", are you my British alter-ego?

[dna_uk 30]

The info from our programmers: "Kick: The game was checking input only against player nick and player number (UID wasn't checked). Ban: The game only considered a player number as a valid input. In order to make bans work with player UIDs and player nicks, the support for string input needed to be implemented, e.g. exec ban 'nick' or exec ban '12345' does the trick now."

"DnA_UK", are you my British alter-ego?

"Cor blimey guv".

Cheers for the clarification :D

[james2464 177]

Nice dev branch, thanks for the ongoing transparency BIS.

But it appears Endstar has used his physx tweeting skills to fool us all haha ;)

[insumsnoy 4]

On current DEV: After finishing campaign episode 'Survive: Blackfoot down', I receive a script error and the game halts loading the next mission (Maxwell)... Can someone confirm this?

'...

File A3\Functions_f\Arrays\fn_sortBy.sqf, line 131

95

_sortedArray set [count _sortedArray,|#|_selectedItem];

_sortedValues set [count...'

Error undefined variable in expression: _selectedItem

EDIT: After a while the mission does start, but any custom load out is not being transferred to 'Death Valley'.

EDIT2: Happens every time when entering Maxwell...

This happens when i try to laod Gori but Camp Maxwell (after BHD) loaded fine for me. I put it in feedback tracker (here) and it looks like the problem has been found.

[blackpixxel 53]

Parameters for weapon inertia added to config (more details once implementation is wrapped)

Maximum weapon inertia multiplied by the current inertia coefficient of the weapon

Will it finally make sense to not use a sniper rifle in Close Quarter Combat? Will we get a nice inertia-effect?

Would be totally awesome!