<!--quoteo(post=1969595:date=Aug 31 2012, 10:08 PM:name=Dictator93)--><div class='quotetop'>QUOTE (Dictator93 @ Aug 31 2012, 10:08 PM) <a href="index.php?act=findpost&pid=1969595"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I was actually curious if there was any chance the client side rendering could be multi-threaded. That means not just on its own thread as it currently is, but also divided among threads if necessary. That would improve the GPU utilization of most comps by quite a bit actually.
It may require a Direct X level higher than 9 though...<!--QuoteEnd--></div><!--QuoteEEnd--> Multi threding does not really work like that, you have to break up "classes" to work a long side the main code. For example the Physx code can and does run on a 2nd thread, but it would be very hard to have only player movement on a 2nd thread due to it relying on a lot of other peaces of code.
matsoMaster of PatchesJoin Date: 2002-11-05Member: 7000Members, Forum Moderators, NS2 Developer, Constellation, NS2 Playtester, Squad Five Blue, Squad Five Silver, Squad Five Gold, Reinforced - Shadow, NS2 Community Developer
<!--quoteo(post=1969588:date=Sep 1 2012, 07:23 AM:name=shad3r)--><div class='quotetop'>QUOTE (shad3r @ Sep 1 2012, 07:23 AM) <a href="index.php?act=findpost&pid=1969588"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Could you elaborate on that matso? I'm not sure if you're saying that the engine has a greater than usual potential for client-side prediction, or just that the work done handling player movement on the server is expensive but stops obvious client hacks. (Is this anything to do with the server using the client time step to handle movement?)<!--QuoteEnd--></div><!--QuoteEEnd-->
Basically, the game reads the user input from the keyboard/mouse and converts it into a time-stamped Move. This Move is then passed to the clients Move-handling routine, which it uses to move you and shoot, all client side, no lag etc.
The Move (and only the move) is also sent to the server, which uses the SAME code to process the Move on the server (after moving the game state back to the correct time for the Move), thus replicating what you did.
So there is no way of changing game state for a client, except through the Move structure. You can still do a lot through that (aimbots), but at least health/armor and speed hacks don't work.
The main problem, performance wise, is that the server has to run the input handling for every Move generated by the clients. As clients generate one Move per display frame, 20 players at 60 fps generates 1200 Moves per second, so the server has a bit of a load problem - it has to handle on one computer what is being generated by 20 computers.
In 217, the server performance increased by quite a bit, but the clients also improved, and now generates a lot more Moves for the server to handle. End result for server ops is that the number of players that a server can handle hasn't changed much.
Fortunately, the problem of handling input moves that have been run on a remote client is obviously embarrassingly parallel, so its an obvious target for multi-threading. Unfortunately, no way that is going to happen before 1.0. Fortunately, performance looks to be on the path to allow 12v12 games on really good hardware; barring clients not getting too much better. Unfortunately, client performance looks like its going to improve even further ... darn :-)
<!--quoteo(post=1969595:date=Sep 1 2012, 04:08 PM:name=Dictator93)--><div class='quotetop'>QUOTE (Dictator93 @ Sep 1 2012, 04:08 PM) <a href="index.php?act=findpost&pid=1969595"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I was actually curious if there was any chance the client side rendering could be multi-threaded. That means not just on its own thread as it currently is, but also divided among threads if necessary. That would improve the GPU utilization of most comps by quite a bit actually.
It may require a Direct X level higher than 9 though...<!--QuoteEnd--></div><!--QuoteEEnd-->
There is already a multicore rendering option - check your options screen. Or did you mean more than one rendering thread? Most things in the render process need to happen in a certain order, so probably not much potential for speedup there. Rendering time is not the issue on most machines, anyway.
<!--quoteo(post=1969602:date=Sep 1 2012, 05:16 PM:name=matso)--><div class='quotetop'>QUOTE (matso @ Sep 1 2012, 05:16 PM) <a href="index.php?act=findpost&pid=1969602"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Basically, the game reads the user input from the keyboard/mouse and converts it into a time-stamped Move. This Move is then passed to the clients Move-handling routine, which it uses to move you and shoot, all client side, no lag etc.
The Move (and only the move) is also sent to the server, which uses the SAME code to process the Move on the server (after moving the game state back to the correct time for the Move), thus replicating what you did.
So there is no way of changing game state for a client, except through the Move structure. You can still do a lot through that (aimbots), but at least health/armor and speed hacks don't work.
The main problem, performance wise, is that the server has to run the input handling for every Move generated by the clients. As clients generate one Move per display frame, 20 players at 60 fps generates 1200 Moves per second, so the server has a bit of a load problem - it has to handle on one computer what is being generated by 20 computers.
In 217, the server performance increased by quite a bit, but the clients also improved, and now generates a lot more Moves for the server to handle. End result for server ops is that the number of players that a server can handle hasn't changed much.
Fortunately, the problem of handling input moves that have been run on a remote client is obviously embarrassingly parallel, so its an obvious target for multi-threading. Unfortunately, no way that is going to happen before 1.0. Fortunately, performance looks to be on the path to allow 12v12 games on really good hardware; barring clients not getting too much better. Unfortunately, client performance looks like its going to improve even further ... darn :-)<!--QuoteEnd--></div><!--QuoteEEnd-->
Thanks for that. Multithreading input handling should work well. I guess you would handle the Move commands on a copy of each player object as it is received in dedicated per-player threads, then briefly lock all the proxy players and copy the data to the main thread player objects at one point in the main game loop.
- the performance: a competitive multiplayer game has to perform well, if the goal is having a "big" community. NS2 already runs bad ( compared to current multiplayer games ...), what if people wanna stream, which is a HUGE point nowadays?
- the pricetag: shootmania, cs:go, dota2, LoL ..... all those games cost 20 € > nothing. I know...full SDK and stuff comes with NS2, but for whom? I bet 90% don´t even care about that. I don´t care.....if i had the chance to chose between a 15 € Game-Only Version ....i´d take that.
( - public-friendlyness: for me, as a guy who´s COMPLETELY new to NS .... actually getting in the game is quite hard, i guess i´d be nice to have somekinda "beginner-tutorials" to explain BASIC stuff about NS2 :D )
I really think NS2 has the potential to be the next-gen kinda competitive game. Really. Im a long time CS player, SC2, Dota/LoL .... and NS2 has something unique and awesome about it. The competitive aspect of this game weights much more than the public one, so pushing that aspect would be quite logical to get this game going.
I mean... NONE of my friends has ever heard of this game .... and they see how bad it runs on my system ....they don´t even bother trying it on theirs.
Take a look at CS:GO: game doesn´t look good ....but it doesn´t look bad either. But it runs well...and is streamable.... same for every other current multiplayer game.
MASSIVE performance boost + lowering the pricetag .... that would help this game A LOT. Otherwise...im afraid it won´t attract many customers nor build a serious competitive scene..... and that would be really sad.
<!--quoteo(post=1969652:date=Sep 1 2012, 01:13 PM:name=G1R)--><div class='quotetop'>QUOTE (G1R @ Sep 1 2012, 01:13 PM) <a href="index.php?act=findpost&pid=1969652"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I don´t care.....if i had the chance to chose between a 15 € Game-Only Version ....i´d take that.<!--QuoteEnd--></div><!--QuoteEEnd-->
I'm running on a mid range graphics card... nearly native res with no problems. The graphics are awesome, it just gets kind of intense when too many things are up. Performance issues for me have been around the same as battlefield 3, so I'm not complaining. Time to save up for a crossfire rig :D Yeah I would definitely have an option though to just click a button and convert the graphics to medium-level from the sexy work of art it is just for smoothness sakes. Something totally streamlined but functional. I imagine in serious competition, people care more about the practicality and smoothness than the aesthetic appreciation.
<!--quoteo(post=1969438:date=Aug 31 2012, 04:43 PM:name=Imbalanxd)--><div class='quotetop'>QUOTE (Imbalanxd @ Aug 31 2012, 04:43 PM) <a href="index.php?act=findpost&pid=1969438"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->What this means is the following; a game programmed by the average Joe would run slower than a game scripted by the average Joe. A game programmed by an educated or experienced professional would run faster than a game scripted by an educated or experienced professional.<!--QuoteEnd--></div><!--QuoteEEnd-->
Scripting is the writing of scripts; tiny little helper programs that automate a series of user actions. They are usually but not necessarily written in an interpreted language like LUA or VB.
NS2 is not scripted in LUA, it is <b>programmed</b> in LUA. The NS2 LUA code is just as complex, and just as messy as mod code in C++. Writing entire programs in LUA is a nightmare from a performance, debugging and development standpoint. More skill is required to write code in LUA; there's no compiler to catch most of your mistakes at compile time, and you need to write more efficient code to deal with the inefficiency of the language.
Just writing a little script that you're not going to extend, you can be really quick and dirty. You don't need to comment the code. You don't need to do type checking. You don't need to adhere to any programming style. You don't need to carefully encapsulate. You don't need to care how other pieces of code are going to interact with you. You don't need to know how you're going to deal with errors. You can just go ahead and incur a lot of technical debt and never repay it.
If you're going to write an entire game in LUA, you're going to have to worry about technical debt, even in LUA. You're going to have to make readable, modular, extendable code that adheres to a sane programming style/code convention.
<!--quoteo(post=1969653:date=Sep 1 2012, 04:18 AM:name=Imbalanxd)--><div class='quotetop'>QUOTE (Imbalanxd @ Sep 1 2012, 04:18 AM) <a href="index.php?act=findpost&pid=1969653"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->A 15 € version that doesn't let you play mods?<!--QuoteEnd--></div><!--QuoteEEnd-->
CS:GO costs 15 $, let´s you play mods... "SDK" is free. I really don´t get why NS2 is so expensive .... i mean: i preordered it for 35 $ ( 30 € ), and i don´t get why it´s so expensive compared to other multiplayer titles. CS:GO 12 € on preorder, Shootmania 15 € on preorder, got BF3 for 20 € ( yeah yeah...i know..all the DLCs n ######, but then again BF3 is no real competitive shooter), LoL free, Dota2 free, SC2 around 30 € ( and SC2 is on a whole other level...feature wise..)and then: Natural Selection 2 .... 30 €. For what?
Not to mention that with the games i mentioned you get: a serious playerbase, enjoyable performance and a competitive scene ( more or less...). I don´t get ANY of this with NS2.
I can guarantee you: if NS2 was f2p ( yeah...i know....), had a good performance for a "big" audience + somekinda matchmaking system .... this game would crush the esports scene.
Well, I don't know. Game prices here are weird. Off the shelf, basically every game here costs R299, which is 35$. Since NS2 basically looks and plays like triple A, I wouldn't have to think twice about buying it for that amount, and didn't.
<!--quoteo(post=1969663:date=Sep 1 2012, 06:52 AM:name=G1R)--><div class='quotetop'>QUOTE (G1R @ Sep 1 2012, 06:52 AM) <a href="index.php?act=findpost&pid=1969663"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I can guarantee you: if NS2 was f2p ( yeah...i know....), had a good performance for a "big" audience + somekinda matchmaking system .... this game would crush the esports scene.<!--QuoteEnd--></div><!--QuoteEEnd-->
Free to play is built around selling you performance enhancing knick-knacks and e-sports is based on a level playing field with stock-copies of the game. They are inherently antithetical concepts.
<!--quoteo(post=1969669:date=Sep 1 2012, 05:17 AM:name=Soylent_green)--><div class='quotetop'>QUOTE (Soylent_green @ Sep 1 2012, 05:17 AM) <a href="index.php?act=findpost&pid=1969669"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Free to play is built around selling you performance enhancing knick-knacks and e-sports is based on a level playing field with stock-copies of the game. They are inherently antithetical concepts.<!--QuoteEnd--></div><!--QuoteEEnd-->
Yeah? Where does LoL make it able to buy something that gives you an advantage? Or Dota? Nowhere.
I'd like to know is NS2 is using Luajit. I was making an RTS with the love2d engine and was getting 10fps with Lua 5.1 after I switched to Luajit with no other re-factoring I got 50fps.
<!--quoteo(post=1969699:date=Sep 1 2012, 04:29 PM:name=perky)--><div class='quotetop'>QUOTE (perky @ Sep 1 2012, 04:29 PM) <a href="index.php?act=findpost&pid=1969699"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I'd like to know is NS2 is using Luajit. I was making an RTS with the love2d engine and was getting 10fps with Lua 5.1 after I switched to Luajit with no other re-factoring I got 50fps.<!--QuoteEnd--></div><!--QuoteEEnd-->
It was at some point, but not anymore. I forgot whether Max replaced it with a custom solution or not.
Interesting, It would be nice to hear from a developer why they dropped it and have they looked into it again recently, seeing as LuaJit has active development.
<!--quoteo(post=1969705:date=Sep 1 2012, 11:11 AM:name=perky)--><div class='quotetop'>QUOTE (perky @ Sep 1 2012, 11:11 AM) <a href="index.php?act=findpost&pid=1969705"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Interesting, It would be nice to hear from a developer why they dropped it and have they looked into it again recently, seeing as LuaJit has active development.<!--QuoteEnd--></div><!--QuoteEEnd-->
iirc - lua jit was removed to improve the c++/lua i/f. actually turned out to be a net perf improvement by doing this.
also, per matso, most of the time is in engine code, not lua (or at least it was until this past patch). so luajit may not bring the gains you expect.
as for a luajit equivalent for ns2, nothing public i've heard, but there have been hints of some work in this area in the dev videos.
<!--quoteo(post=1969709:date=Sep 1 2012, 08:37 AM:name=saltybp53)--><div class='quotetop'>QUOTE (saltybp53 @ Sep 1 2012, 08:37 AM) <a href="index.php?act=findpost&pid=1969709"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I have a 2008 overclocked computer, ns2 runs fine on 30 fps with occassional stutters.<!--QuoteEnd--></div><!--QuoteEEnd-->
Well, i have a 2011 top-notch gaming laptop which runs BF3 maxed out @kinda stable 40 fps, CS:GO at 200fps, Shootmania @120 fps and NS2 everything set to low/off at about 40 FPS dropping down to 20 at times ( critical times....). That´s not suitable at all.
<!--quoteo(post=1969693:date=Sep 1 2012, 09:06 AM:name=G1R)--><div class='quotetop'>QUOTE (G1R @ Sep 1 2012, 09:06 AM) <a href="index.php?act=findpost&pid=1969693"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Yeah? Where does LoL make it able to buy something that gives you an advantage?<!--QuoteEnd--></div><!--QuoteEEnd-->
You purchase champions. If you play for free you play only the free champion of the week. Being only able to chose whatever champions are free this week is an extraordinary meta-gaming disadvantage and doesn't allow you to practice consistently with one champion.
DOTA was not free to play(it was a mod), DOTA 2 is planing on only cosmetic items. This is an unproven free-to-play-model and I think it will be a commercial failure.
DghelneshiAims to surpass Fana in post edits.Join Date: 2011-11-01Member: 130634Members, Squad Five Blue, Reinforced - Shadow
<!--quoteo(post=1969723:date=Sep 1 2012, 06:44 PM:name=Soylent_green)--><div class='quotetop'>QUOTE (Soylent_green @ Sep 1 2012, 06:44 PM) <a href="index.php?act=findpost&pid=1969723"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->DOTA 2 is planing on only cosmetic items. This is an unproven free-to-play-model and I think it will be a commercial failure.<!--QuoteEnd--></div><!--QuoteEEnd--> The success of this model depends on the size of the playerbase. If the playerbase is huge, you will make enough money off it to support continued development. Sure, you'd make more money by charging for other things, but it's not necessarily a failure.
<!--quoteo(post=1969723:date=Sep 1 2012, 09:44 AM:name=Soylent_green)--><div class='quotetop'>QUOTE (Soylent_green @ Sep 1 2012, 09:44 AM) <a href="index.php?act=findpost&pid=1969723"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->You purchase champions. If you play for free you play only the free champion of the week. Being only able to chose whatever champions are free this week is an extraordinary meta-gaming disadvantage and doesn't allow you to practice consistently with one champion.
DOTA was not free to play(it was a mod), DOTA 2 is planing on only cosmetic items. This is an unproven free-to-play-model and I think it will be a commercial failure.<!--QuoteEnd--></div><!--QuoteEEnd-->
Yeah, i meant Dota2.
And no....you don´t have to BUY champions in LoL. For every game you play you gain points which you can get champions/runes for. No need to invest a single cent to get champions. F2p ..... gameplay-wise you don´t get an advantage through BUYING something. You can play, earn points and get everything for free, except skins....and skins don´t alter the gameplay in any way.
So both LoL and Dota2 are F2P without giving people a disadvantage if they don´t spend money.
<!--quoteo(post=1969731:date=Sep 1 2012, 01:03 PM:name=G1R)--><div class='quotetop'>QUOTE (G1R @ Sep 1 2012, 01:03 PM) <a href="index.php?act=findpost&pid=1969731"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Well.... all that is irrelevant to the opinion that NS2 is overpriced.<!--QuoteEnd--></div><!--QuoteEEnd-->
that opinion is subjective, and can be altered. the #1 problem is i don't think uwe has the brand recognition that AAA devs do.
<!--quoteo(post=1969735:date=Sep 1 2012, 10:26 AM:name=Wheeee)--><div class='quotetop'>QUOTE (Wheeee @ Sep 1 2012, 10:26 AM) <a href="index.php?act=findpost&pid=1969735"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->that opinion is subjective, and can be altered. the #1 problem is i don't think uwe has the brand recognition that AAA devs do.<!--QuoteEnd--></div><!--QuoteEEnd-->
Yep ofc.... so they kinda need to find other ways to bring their game to the customer. ###### performance, high price compared to other similar titles ( multiplayer games ), small community .... not things that are helping this game.
<!--quoteo(post=1969790:date=Sep 1 2012, 08:29 PM:name=G1R)--><div class='quotetop'>QUOTE (G1R @ Sep 1 2012, 08:29 PM) <a href="index.php?act=findpost&pid=1969790"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Yep ofc.... so they kinda need to find other ways to bring their game to the customer. ###### performance, high price compared to other similar titles ( multiplayer games ), small community .... not things that are helping this game.<!--QuoteEnd--></div><!--QuoteEEnd-->
UWE already knows about the price "problem", but F2P is not really an option for NS2. <a href="https://twitter.com/NS2/status/235433683308400641" target="_blank">https://twitter.com/NS2/status/235433683308400641</a>
CS:GO was expected to cost also around 30 Euros in the beginning before they started the preorder on Steam. Imho they reduced it to the half, because they realised that they would never achieve the goal of bringing players from CS:S and 1.6 to a new game with a game that expensive (yet they failed, because the CS:GO playercount is decreasing more an more over the past week and the CS 1.6 playerbase remained almost stable).
<!--quoteo(post=1969801:date=Sep 1 2012, 02:17 PM:name=falc)--><div class='quotetop'>QUOTE (falc @ Sep 1 2012, 02:17 PM) <a href="index.php?act=findpost&pid=1969801"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->UWE already knows about the price "problem", but F2P is not really an option for NS2. <a href="https://twitter.com/NS2/status/235433683308400641" target="_blank">https://twitter.com/NS2/status/235433683308400641</a>
CS:GO was expected to cost also around 30 Euros in the beginning before they started the preorder on Steam. Imho they reduced it to the half, because they realised that they would never achieve the goal of bringing players from CS:S and 1.6 to a new game with a game that expensive (yet they failed, because the CS:GO playercount is decreasing more an more over the past week and the CS 1.6 playerbase remained almost stable).<!--QuoteEnd--></div><!--QuoteEEnd-->
Well, cuz a bad game won´t keep players ...even if it would be F2P ... :D. Good game >> attractive to a big audience >> win. NS2 is a good game....but it isn´t really interesting for quite a big audience cuz of known issues. Performance, price.
Fix the performance as much as possible ( i´d be happy to run it at STABLE 50 fps + being able to stream ....), lower the price = <img src="http://4.bp.blogspot.com/-9QjARWd5a84/T3OXHJJbe1I/AAAAAAAAA34/OVtLNvwpMvI/s190/nod-of-approval.gif" border="0" class="linked-image" />
Lua vs LuaJIT doesn't matter as much as for example "profile" UI drawing every piece of text with separate draw call (and maybe more shader pipeline flushing).
<!--QuoteBegin-matso+--><div class='quotetop'>QUOTE (matso)</div><div class='quotemain'><!--QuoteEBegin-->Fortunately, the problem of handling input moves that have been run on a remote client is obviously embarrassingly parallel,<!--QuoteEnd--></div><!--QuoteEEnd--> You're totally wrong.
Comments
It may require a Direct X level higher than 9 though...<!--QuoteEnd--></div><!--QuoteEEnd-->
Multi threding does not really work like that, you have to break up "classes" to work a long side the main code. For example the Physx code can and does run on a 2nd thread, but it would be very hard to have only player movement on a 2nd thread due to it relying on a lot of other peaces of code.
Basically, the game reads the user input from the keyboard/mouse and converts it into a time-stamped Move. This Move is then passed to the clients Move-handling routine, which it uses to move you and shoot, all client side, no lag etc.
The Move (and only the move) is also sent to the server, which uses the SAME code to process the Move on the server (after moving the game state back to the correct time for the Move), thus replicating what you did.
So there is no way of changing game state for a client, except through the Move structure. You can still do a lot through that (aimbots), but at least health/armor and speed hacks don't work.
The main problem, performance wise, is that the server has to run the input handling for every Move generated by the clients. As clients generate one Move per display frame, 20 players at 60 fps generates 1200 Moves per second, so the server has a bit of a load problem - it has to handle on one computer what is being generated by 20 computers.
In 217, the server performance increased by quite a bit, but the clients also improved, and now generates a lot more Moves for the server to handle. End result for server ops is that the number of players that a server can handle hasn't changed much.
Fortunately, the problem of handling input moves that have been run on a remote client is obviously embarrassingly parallel, so its an obvious target for multi-threading.
Unfortunately, no way that is going to happen before 1.0.
Fortunately, performance looks to be on the path to allow 12v12 games on really good hardware; barring clients not getting too much better.
Unfortunately, client performance looks like its going to improve even further ... darn :-)
It may require a Direct X level higher than 9 though...<!--QuoteEnd--></div><!--QuoteEEnd-->
There is already a multicore rendering option - check your options screen. Or did you mean more than one rendering thread? Most things in the render process need to happen in a certain order, so probably not much potential for speedup there. Rendering time is not the issue on most machines, anyway.
<!--quoteo(post=1969602:date=Sep 1 2012, 05:16 PM:name=matso)--><div class='quotetop'>QUOTE (matso @ Sep 1 2012, 05:16 PM) <a href="index.php?act=findpost&pid=1969602"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Basically, the game reads the user input from the keyboard/mouse and converts it into a time-stamped Move. This Move is then passed to the clients Move-handling routine, which it uses to move you and shoot, all client side, no lag etc.
The Move (and only the move) is also sent to the server, which uses the SAME code to process the Move on the server (after moving the game state back to the correct time for the Move), thus replicating what you did.
So there is no way of changing game state for a client, except through the Move structure. You can still do a lot through that (aimbots), but at least health/armor and speed hacks don't work.
The main problem, performance wise, is that the server has to run the input handling for every Move generated by the clients. As clients generate one Move per display frame, 20 players at 60 fps generates 1200 Moves per second, so the server has a bit of a load problem - it has to handle on one computer what is being generated by 20 computers.
In 217, the server performance increased by quite a bit, but the clients also improved, and now generates a lot more Moves for the server to handle. End result for server ops is that the number of players that a server can handle hasn't changed much.
Fortunately, the problem of handling input moves that have been run on a remote client is obviously embarrassingly parallel, so its an obvious target for multi-threading.
Unfortunately, no way that is going to happen before 1.0.
Fortunately, performance looks to be on the path to allow 12v12 games on really good hardware; barring clients not getting too much better.
Unfortunately, client performance looks like its going to improve even further ... darn :-)<!--QuoteEnd--></div><!--QuoteEEnd-->
Thanks for that. Multithreading input handling should work well. I guess you would handle the Move commands on a copy of each player object as it is received in dedicated per-player threads, then briefly lock all the proxy players and copy the data to the main thread player objects at one point in the main game loop.
- the performance: a competitive multiplayer game has to perform well, if the goal is having a "big" community. NS2 already runs bad ( compared to current multiplayer games ...), what if people wanna stream, which is a HUGE point nowadays?
- the pricetag: shootmania, cs:go, dota2, LoL ..... all those games cost 20 € > nothing. I know...full SDK and stuff comes with NS2, but for whom? I bet 90% don´t even care about that. I don´t care.....if i had the chance to chose between a 15 € Game-Only Version ....i´d take that.
( - public-friendlyness: for me, as a guy who´s COMPLETELY new to NS .... actually getting in the game is quite hard, i guess i´d be nice to have somekinda "beginner-tutorials" to explain BASIC stuff about NS2 :D )
I really think NS2 has the potential to be the next-gen kinda competitive game. Really. Im a long time CS player, SC2, Dota/LoL .... and NS2 has something unique and awesome about it. The competitive aspect of this game weights much more than the public one, so pushing that aspect would be quite logical to get this game going.
I mean... NONE of my friends has ever heard of this game .... and they see how bad it runs on my system ....they don´t even bother trying it on theirs.
Take a look at CS:GO: game doesn´t look good ....but it doesn´t look bad either. But it runs well...and is streamable.... same for every other current multiplayer game.
MASSIVE performance boost + lowering the pricetag .... that would help this game A LOT. Otherwise...im afraid it won´t attract many customers nor build a serious competitive scene..... and that would be really sad.
A 15 € version that doesn't let you play mods?
Scripting is the writing of scripts; tiny little helper programs that automate a series of user actions. They are usually but not necessarily written in an interpreted language like LUA or VB.
NS2 is not scripted in LUA, it is <b>programmed</b> in LUA. The NS2 LUA code is just as complex, and just as messy as mod code in C++. Writing entire programs in LUA is a nightmare from a performance, debugging and development standpoint. More skill is required to write code in LUA; there's no compiler to catch most of your mistakes at compile time, and you need to write more efficient code to deal with the inefficiency of the language.
Just writing a little script that you're not going to extend, you can be really quick and dirty. You don't need to comment the code. You don't need to do type checking. You don't need to adhere to any programming style. You don't need to carefully encapsulate. You don't need to care how other pieces of code are going to interact with you. You don't need to know how you're going to deal with errors. You can just go ahead and incur a lot of technical debt and never repay it.
If you're going to write an entire game in LUA, you're going to have to worry about technical debt, even in LUA. You're going to have to make readable, modular, extendable code that adheres to a sane programming style/code convention.
CS:GO costs 15 $, let´s you play mods... "SDK" is free. I really don´t get why NS2 is so expensive .... i mean: i preordered it for 35 $ ( 30 € ), and i don´t get why it´s so expensive compared to other multiplayer titles. CS:GO 12 € on preorder, Shootmania 15 € on preorder, got BF3 for 20 € ( yeah yeah...i know..all the DLCs n ######, but then again BF3 is no real competitive shooter), LoL free, Dota2 free, SC2 around 30 € ( and SC2 is on a whole other level...feature wise..)and then: Natural Selection 2 .... 30 €. For what?
Not to mention that with the games i mentioned you get: a serious playerbase, enjoyable performance and a competitive scene ( more or less...). I don´t get ANY of this with NS2.
I can guarantee you: if NS2 was f2p ( yeah...i know....), had a good performance for a "big" audience + somekinda matchmaking system .... this game would crush the esports scene.
Free to play is built around selling you performance enhancing knick-knacks and e-sports is based on a level playing field with stock-copies of the game. They are inherently antithetical concepts.
Yeah? Where does LoL make it able to buy something that gives you an advantage? Or Dota? Nowhere.
It was at some point, but not anymore. I forgot whether Max replaced it with a custom solution or not.
iirc - lua jit was removed to improve the c++/lua i/f. actually turned out to be a net perf improvement by doing this.
also, per matso, most of the time is in engine code, not lua (or at least it was until this past patch). so luajit may not bring the gains you expect.
as for a luajit equivalent for ns2, nothing public i've heard, but there have been hints of some work in this area in the dev videos.
Well, i have a 2011 top-notch gaming laptop which runs BF3 maxed out @kinda stable 40 fps, CS:GO at 200fps, Shootmania @120 fps and NS2 everything set to low/off at about 40 FPS dropping down to 20 at times ( critical times....). That´s not suitable at all.
You purchase champions. If you play for free you play only the free champion of the week. Being only able to chose whatever champions are free this week is an extraordinary meta-gaming disadvantage and doesn't allow you to practice consistently with one champion.
<!--quoteo(post=1969693:date=Sep 1 2012, 09:06 AM:name=G1R)--><div class='quotetop'>QUOTE (G1R @ Sep 1 2012, 09:06 AM) <a href="index.php?act=findpost&pid=1969693"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Or Dota? Nowhere.<!--QuoteEnd--></div><!--QuoteEEnd-->
DOTA was not free to play(it was a mod), DOTA 2 is planing on only cosmetic items. This is an unproven free-to-play-model and I think it will be a commercial failure.
The success of this model depends on the size of the playerbase. If the playerbase is huge, you will make enough money off it to support continued development. Sure, you'd make more money by charging for other things, but it's not necessarily a failure.
DOTA was not free to play(it was a mod), DOTA 2 is planing on only cosmetic items. This is an unproven free-to-play-model and I think it will be a commercial failure.<!--QuoteEnd--></div><!--QuoteEEnd-->
Yeah, i meant Dota2.
And no....you don´t have to BUY champions in LoL. For every game you play you gain points which you can get champions/runes for. No need to invest a single cent to get champions. F2p ..... gameplay-wise you don´t get an advantage through BUYING something. You can play, earn points and get everything for free, except skins....and skins don´t alter the gameplay in any way.
So both LoL and Dota2 are F2P without giving people a disadvantage if they don´t spend money.
that opinion is subjective, and can be altered. the #1 problem is i don't think uwe has the brand recognition that AAA devs do.
Yep ofc.... so they kinda need to find other ways to bring their game to the customer. ###### performance, high price compared to other similar titles ( multiplayer games ), small community .... not things that are helping this game.
UWE already knows about the price "problem", but F2P is not really an option for NS2.
<a href="https://twitter.com/NS2/status/235433683308400641" target="_blank">https://twitter.com/NS2/status/235433683308400641</a>
CS:GO was expected to cost also around 30 Euros in the beginning before they started the preorder on Steam. Imho they reduced it to the half, because they realised that they would never achieve the goal of bringing players from CS:S and 1.6 to a new game with a game that expensive (yet they failed, because the CS:GO playercount is decreasing more an more over the past week and the CS 1.6 playerbase remained almost stable).
<a href="https://twitter.com/NS2/status/235433683308400641" target="_blank">https://twitter.com/NS2/status/235433683308400641</a>
CS:GO was expected to cost also around 30 Euros in the beginning before they started the preorder on Steam. Imho they reduced it to the half, because they realised that they would never achieve the goal of bringing players from CS:S and 1.6 to a new game with a game that expensive (yet they failed, because the CS:GO playercount is decreasing more an more over the past week and the CS 1.6 playerbase remained almost stable).<!--QuoteEnd--></div><!--QuoteEEnd-->
Well, cuz a bad game won´t keep players ...even if it would be F2P ... :D.
Good game >> attractive to a big audience >> win. NS2 is a good game....but it isn´t really interesting for quite a big audience cuz of known issues. Performance, price.
Fix the performance as much as possible ( i´d be happy to run it at STABLE 50 fps + being able to stream ....), lower the price = <img src="http://4.bp.blogspot.com/-9QjARWd5a84/T3OXHJJbe1I/AAAAAAAAA34/OVtLNvwpMvI/s190/nod-of-approval.gif" border="0" class="linked-image" />
Regards
You're totally wrong.
He is not, to a certain extent.