To BIS dev's

[Rexxenexx 0]

Intel Open-Sources Threading Building Blocks

Can be useful for a ArmA v2.0 release?

EDIT: Here is Intel's press release(a non-Ziff Davis POS site): Intel Press Release

And a Direct Link

[Balschoiw 0]

It just takes me to an ad....

Edit:

Popup killer kept me from getting the content.

[Raptor 10]

ArmA v.2.0 ? What the hell?

[Rexxenexx 0]

ArmA v.2.0 ? What the hell?

Well its 1.08 now and this would be a major update, maybe 1.18?

[Rexxenexx 0]

It just takes me to an ad....

Edit:

Popup killer kept me from getting the content.

It shouldn't, but you never know its a Ziff Davis POS page so they might try and spam. heres the skinny on what the page is about:

"Intel Threading Building Blocks 2.0 is a C++ runtime library that simplifies threading for performance and scalability. It is a higher-level abstraction that requires less code for threading and delivers multicore performance. It is ready today for the multicore architectures of tomorrow," said James Reinders, software evangelist and director of marketing for Intel's software development products.

[Dwarden 1125]

mmmm ... optimized server code ....

[Rexxenexx 0]

It's supposed to be easily implemented. Even if its just used for a stand-alone server it would be a huge performance improvement.

[Jack-UK 0]

mmmm ... optimized server code ....  

Exactly my thoughts... mmm 200 player battles please...

[bravo 6 0]

i would like to use this thread to ask BIS what are their ambicions at this stage, without starting a hole new thread.

[*] Is ArmA fully optimized?

[*] Can we hope more about ArmA Engine?

[*] Have BIS Moved on to the next project living ArmA as it is?

don't want to disrespect anyone, im just looking for some answers that are hard to find in here -->

[snkman 351]

Quote[/b] ]Have BIS Moved on to the next project living ArmA as it is?

Well if this should be fact then i'm sure BIS will lose a lot of players.

[smellyjelly 0]

If they are, then I hope to get some kind of discount on the next game. I mean, Armed Assault is a great game, but it reeks of budget and unfinished.

All well, if it's good I'll probably still buy it...

[Beagle 684]

Well if this should be fact then i'm sure BIS will lose a lot of players.

The correct term is "consumers"... that's what counts, not players. The fear of companies like BIS should be to loose the consumers confidence in their product quality...this recently broke the neck of a lot of formerly succesfull game studios since more and more of the "consumers" nowadays are precautios, and wait how a software develops after Release.

Everyone knows these days that a fresh release is always full of bugs and showstoppers...but the half priced budget or "gold" or "GotY" version 6-9 or 12 month later will be playable.

Just for example... I waited till V.1.05 and the announcment of a 1.07 update.

And I wont buy a expansion until the patch after it's release since there is no fun in playing half baked games...my leisure time just became to rare and precious in the last years for the role of a after sale Beta tester.

[ramboofp 0]

EDIT misread sry

[ShrubMiK 0]

Intel eh?

Does it work on AMD?

[JFK 0]

Quote[/b] ]Does it work on AMD?

Yes it does. Has nothing to do with the architecture. It's "just" a c++ library which can be used to "easily" implement multithreading support to your existing or new code. Any c++ compiler can compile the code.

You can use the library to parallelize indepent code segments, loops and so on. Further they added container classes that can be used by parallel threads. All that stuff is quite easy to implement, you only need to add view lines of code.

The difference to classic thread programming is, you dont need to care about creating threads, managing data-access of variables and so on... you just create so called "tasks", the library distributes the work to the available processors.

I think chances are not bad, BIS could use the library in future patches.

Cya

[MATRA 0]

i would like to use this thread to ask BIS what are their ambicions at this stage, without starting a hole new thread.

[*] Is ArmA fully optimized?

[*] Can we hope more about ArmA Engine?

[*] Have BIS Moved on to the next project living ArmA as it is?

don't want to disrespect anyone, im just looking for some answers that are hard to find in here -->

Good point bravo.

But since Placebo has left the building, the link between Community and Producers, seems to be broken.

[baddo 0]

So Intel Threading Building Blocks is The Magic Bullet for multi-threaded programming? It's suddenly easy?

[JFK 0]

So Intel Threading Building Blocks is The Magic Bulletâ„¢ for multi-threaded programming? It's suddenly easy?

Quote[/b] ]It's suddenly easy?

I wouldn't say it's easy, but it's easier. You still need to know how to use the classes and methods provided by the library. Furthermore you need to know which pieces of the code can be parallelized.

I would say its the first step in making multi-threaded programming easier and accepted by a broader audience of programmers. As far as I've read about TBB, you can achieve quite good performance boosts with just a little rewriting of the code. Hope BIS will use it  

[Beagle 684]

Honestly..if I were a game engine programmer in a niché market like this, I would not go for the newes engine stuff...this always smells like squashing bugs for another year to come.

[DaSquade 0]

Bit offtopic maybe, but though it was good enough to add:

Quote[/b] ]MATRA

Quote (bravo 6 @ Aug. 18 2007,02:45)

i would like to use this thread to ask BIS what are their ambicions at this stage, without starting a hole new thread.

# Is ArmA fully optimized?

# Can we hope more about ArmA Engine?

# Have BIS Moved on to the next project living ArmA as it is?

don't want to disrespect anyone, im just looking for some answers that are hard to find in here -->

Good point bravo.

But since Placebo has left the building, the link between Community and Producers, seems to be broken.

Wasn't it not Placebo himself who told in an interview something along the lines of:

Quote[/b] ]Honestly, i'm afraid i can't answer these detailed questions as this is not my field of experience

My point is, only a 'true' inhouse PR/DEV would be able to answer these questions in a direct way. On the other hand i think it is fair 'they' will never show us their hand of cards, although they sometimes here and there try to answer or explain stuff.

Example: the truth about the current optimization of vegetation in ArmA.

Quote[/b] ] Have BIS Moved on to the next project living ArmA as it is?

Sorry for extra derail, but somehow this hit me with a hammer and hope there will never be an message like:

Quote[/b] ]XXXnr will be fully supported by Bohemia Interactive for years to come.

..joking offcourse and like said before BIS has more then 1 team for each of whatever project and untill that new project is officially out i think we can safely say they will try to make the best out of it.

[ShrubMiK 0]

"As far as I've read about TBB, you can achieve quite good performance boosts with just a little rewriting of the code."

I seriously doubt it is that simple...the major problem with multi-threaded programming is not identifying what can be parallelised and spawning off new threads to do it...it's making sure your code is safe from all those pesky synchronisation problems and potential deadlocks that crop up later.

But I'll take a look at it, if they have provided classes that really do make some of this stuff easier it could be very useful...especially if it is portable enough to work on Unix/Linux not just Windows?

[JFK 0]

Quote[/b] ]I seriously doubt it is that simple...the major problem with multi-threaded programming is not identifying what can be parallelised and spawning off new threads to do it...it's making sure your code is safe from all those pesky synchronisation problems and potential deadlocks that crop up later.

Thats the point, you dont have to spawn threads, the library does that for you. Sure you still have to worry about deadlocks an sync probs, but the implementation of threads itself is much easier, as you don't manage the threads, but use tasks instead. The library creates the threads, optimized for the available hardware, does load-balancing between the available processors, has thread-safe container classes which are much more performant than "manually proteceted" containers by e.g. mutexes, and so on...

Quote[/b] ]But I'll take a look at it, if they have provided classes that really do make some of this stuff easier it could be very useful...especially if it is portable enough to work on Unix/Linux not just Windows?

Yes it is portable/runnable on Unix/Linux.

Let me qoute a listing of an Intel page about TBB(http://www.threadingbuildingblocks.org/):

"For developers, the clear benefits of Threading Building Blocks are:

1. TBB significantly reduces the number of lines of code required to develop multithreaded applications;

2. TBB significantly reduces the programming complexity for developing multithreaded applications (by abstracting many details of thread management);

3. TBB's task manager automatically analyzes the system the software is running on, chooses the optimal number of threads, and performs load balancing that spreads out the work evenly across all processor cores;

4. As a result, TBB threaded applications automatically scale to fully utilize all available processing cores on whatever computer they run on – including future systems that will have many more cores than are available (or affordable) today."

Don't get me wrong, I dont have any experience with TBB, I was just reading some tutorials from Intel, and what I've read I can only assume that BIS (or any other dev which has potential of performance gains with multi-threading in his application) should consider TBB. I can remeber a statement of Suma, where he mentioned that the only signinficant performance gain they could get, would be implementing multi-threading. But he can not promise that this will occur.

I think with the usage of TBB, a lot of hurdles of the common multi-threading programming (the reasons you have mentioned, deadlocks, sync-probs, overhead caused by too many threads, etc) could be eliminated/disregarded. That's just my little opinion.

Cya