IronHorseDeveloper, QA Manager, Technical Support & contributorJoin Date: 2010-05-08Member: 71669Members, Super Administrators, Forum Admins, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Blue, Subnautica Playtester, Subnautica PT Lead, Pistachionauts
<!--quoteo(post=1829909:date=Feb 5 2011, 02:41 PM:name=luns)--><div class='quotetop'>QUOTE (luns @ Feb 5 2011, 02:41 PM) <a href="index.php?act=findpost&pid=1829909"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->^ well said.
i hope stability is honestly fixed, making beta playable is usually the main problem right now, and ever been. One of the main reasons I sometimes stop playing is due to my fps spiking like crazy, making game play really slow. The game itself is beta but the engine itself feels like alpha, imo.<!--QuoteEnd--></div><!--QuoteEEnd-->
anyone tried Wolf fire's Overgrowth alpha? <a href="http://www.wolfire.com/overgrowth" target="_blank">http://www.wolfire.com/overgrowth</a> most amazingly stable and optimized game engine i've ever seen and its barely in an alpha stage.. just a thought.
What's the deal with the horrible performance? Half a year since this went public and the engine still runs like a dog. Pardon my French, but it's true.
Is optimizing the engine not a high enough priority? It should be. I'm trying my hardest to be patient here, I really am, but I'm also starting to get quite worried.
Can I ever expect to get silky smooth performance with this game on my current Q6600, ATI 5770? How much more work needs to be or even CAN be done before the engine performs well for the majority of gamers with modest rigs?
<!--quoteo(post=1829997:date=Feb 6 2011, 10:55 AM:name=Arrowhead)--><div class='quotetop'>QUOTE (Arrowhead @ Feb 6 2011, 10:55 AM) <a href="index.php?act=findpost&pid=1829997"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->What's the deal with the horrible performance? Half a year since this went public and the engine still runs like a dog.<!--QuoteEnd--></div><!--QuoteEEnd--> That good? A dog can run like over 30-35 km/h or more! Seriously i got like 25-30 fps, most people too and this is ok, not good but for a earlie beta, its ok.
<!--quoteo(post=1830004:date=Feb 6 2011, 06:56 AM:name=Price)--><div class='quotetop'>QUOTE (Price @ Feb 6 2011, 06:56 AM) <a href="index.php?act=findpost&pid=1830004"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->That good? A dog can run like over 30-35 km/h or more! Seriously i got like 25-30 fps, most people too and this is ok, not good but for a earlie beta, its ok.<!--QuoteEnd--></div><!--QuoteEEnd-->
I assume that the reference to 'horrible performance' is regarding the 'glitchiness' mentioned above (as opposed to poor FPS); player movement rubber-banding, collision detection problems, poor hit detection, etc. which can be described generally as a feeling of 'unresponsiveness' or a marked disconnect between the player's actions and the avatar's in-game reactions.
<!--quoteo(post=0:date=:name=weezl)--><div class='quotetop'>QUOTE (weezl)</div><div class='quotemain'><!--quotec-->omg another such post...
just pretend this is a regular game that you don't have access to and simply DON'T PLAY<!--QuoteEnd--></div><!--QuoteEEnd-->
IMO, this kind of answer is <i>unacceptable</i>; it amounts, again, to rug-sweeping the issues away. As stated, we understand that the game is currently in Beta, and that what we have been given access to is a Beta version of the game which should be expected to have some issues. People following your advice is exactly what I'm hoping can be avoided, because simply doing that without raising any criticisms in the process amounts to <i>accepting not being able to play</i> and arguably in many cases--my friend, for example--this means basically giving up on NS2.
Another related point has to do with word of mouth publicity for the game. How do you expect there to be much at all in the way of positive word of mouth for this game--Indie titles such as this one <u>need</u> it--if the vast majority of people with access to the game are too frustrated by its performance to even play a full match?
The game looks great, the concept is exciting and innovative. All I ask is that this issue be officially addressed; something along the lines of, "The reason these issues haven't been resolved yet is __________." It's great that UWE is so involved with the community surrounding this game, but this issue has--for quite some time--been the seldom acknowledged 'elephant in the room' that most people are probably wondering about. As such, the flow of announcements about new feature ideas, prototyping etc, are starting to 'rub-it-in'.
Again: THANK YOU UWE for working on this project - we all appreciate it very much! I'm simply trying to contribute what I can to make sure this game doesn't become a forgotten gem.
P.S.
I like the idea of 'Community-Questions' video features. No offense intended, but if you're going to bother making videos like this; 1. Answer more than one question (r.e. one question about squads?), 2. Answer questions that you can provide more than a <i>vague</i> reply to (r.e. "We're going do to something to make squads better...")
I only mention this because it's something that contributes to the 'rubbing-it-in' I brought up above. In other words, how do you justify spending the time to make videos for question-&-answer if the chosen question pertains to something relatively minor (in comparison to the performance issues which have led to most people not playing) and the response is so vague as to be a non-response?
I apologize if any of that came across as offensive; not my intent at all! I'm just looking for logical explanations....
[Edited for Grammar; added P.S. re-'Community-Questions' video]
<!--quoteo(post=1830074:date=Feb 6 2011, 07:43 PM:name=Cpt.j2oss)--><div class='quotetop'>QUOTE (Cpt.j2oss @ Feb 6 2011, 07:43 PM) <a href="index.php?act=findpost&pid=1830074"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->All I ask is that this issue be officially addressed; something along the lines of, "The reason these issues haven't been resolved yet is __________." It's great that UWE is so involved with the community surrounding this game, but this issue has--for quite some time--been the seldom acknowledged 'elephant in the room' that most people are probably wondering about. As such, the flow of announcements about new feature ideas, prototyping etc, are starting to 'rub-it-in'.<!--QuoteEnd--></div><!--QuoteEEnd--> We HAVE acknowledged and addressed this on numerous occasions. Max has stated countless times that performance is his #1 priority at the moment, and that is what he has been focused on. Even tasks which may seem on the surface to not be directly related to performance usually are. Its not like there is one problem that is causing all the lag and low framerate, and Max can come on and say we are going to fix X bug and its going to solve all your problems. Max has a list of numerous optimizations that are planned, and hopefully each one of those will keep improving the game more and more.
Its not an exact science. Some optimizations that are expected to make a big impact sometimes result in smaller then expected gains, and vice versa. One optimization may have a large impact, but some other issues may be holding back the gains made from it. Add to that, everyone is running on a completely different system. Some people may see huge gains with one type of graphics card, while people with other cards, don't. Optimizing a game to run on a large spectrum of system specs is a very daunting process, compared to optimizing a game around 1 type of hardware, like an xbox or playstation.
These are all partly reason why we don't want to hype up certain optimizations that are being worked on, because everyone will expect a huge improvement and then be potentially disappointed when the patch comes out and they don't see any difference.
Trust me, Max is as frustrated as any of you. The poor performance eats away at him every day, and he is working hard to improve server lag and framerate. The new features and prototyping you see are coming mostly from Charlie's end (such as the DI at the moment) so its not taking time away from Max's optimizations.
@Cpt.j2oss as you see nothing is being swept under the rug, neither was it at all what i meant! point is game development must take it's course, and the developers know best what to do - afterall they want/need the game out most of all! community whine is pointless and simply would not exist for a regular game (at this stage).
the best we can do is discuss features, bring ideas, feedback, etc!
<!--quoteo(post=1830084:date=Feb 6 2011, 05:07 PM:name=Cpt.j2oss)--><div class='quotetop'>QUOTE (Cpt.j2oss @ Feb 6 2011, 05:07 PM) <a href="index.php?act=findpost&pid=1830084"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Thank you, Cory, for this reply! This is exactly the kind of response I was hoping for; I feel re-assured!<!--QuoteEnd--></div><!--QuoteEEnd-->
Ditto. Thanks for that response. My faith has been restored. Keep up the good work, I really do appreciate it. :)
If everyone wants performance so bad (I know I do), why don't we make a map with simple geometry and few props? A full map with a proper layout, but just simple textures.
I think this is the easiest, most straight-forward way to boost framerates for people who want smoothness. Surely it can't take more than a single day to make a simple map. Just make it an official map, so everyone can play it. That will tide us over until performance can be improved.
<!--quoteo(post=1830106:date=Feb 7 2011, 01:03 AM:name=FuzionMonkey)--><div class='quotetop'>QUOTE (FuzionMonkey @ Feb 7 2011, 01:03 AM) <a href="index.php?act=findpost&pid=1830106"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->If everyone wants performance so bad (I know I do), why don't we make a map with simple geometry and few props? A full map with a proper layout, but just simple textures.
I think this is the easiest, most straight-forward way to boost framerates for people who want smoothness. Surely it can't take more than a single day to make a simple map. Just make it an official map, so everyone can play it. That will tide us over until performance can be improved.<!--QuoteEnd--></div><!--QuoteEEnd-->
Evil_bOb1Join Date: 2002-07-13Member: 938Members, Squad Five Blue
<!--quoteo(post=1830106:date=Feb 6 2011, 06:03 PM:name=FuzionMonkey)--><div class='quotetop'>QUOTE (FuzionMonkey @ Feb 6 2011, 06:03 PM) <a href="index.php?act=findpost&pid=1830106"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Surely it can't take more than a single day to make a simple map.<!--QuoteEnd--></div><!--QuoteEEnd-->
Things take much longer and there is only so much you can do in a day. You could make a box room map in a day but it wouldn't be enjoyable. Basically, the more time and thought you put into it the richer and diverse the experience. A rushed map is a poor experience for players.
<!--quoteo(post=1830127:date=Feb 6 2011, 08:54 PM:name=Evil_bOb1)--><div class='quotetop'>QUOTE (Evil_bOb1 @ Feb 6 2011, 08:54 PM) <a href="index.php?act=findpost&pid=1830127"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Things take much longer and there is only so much you can do in a day. You could make a box room map in a day but it wouldn't be enjoyable. Basically, the more time and thought you put into it the richer and diverse the experience. A rushed map is a poor experience for players.<!--QuoteEnd--></div><!--QuoteEEnd-->
You can make a box room map in like 10 minutes. It shouldn't take long for someone experienced with the editor to make a basic, symmetrical map. You know whats a poor experience for players? Terrible framerates.
Actually, the map detail is really not that much of a problem. Its mostly other things running behind the scenes. We've even done tests with a version of the map wtih no lighting and shadows, and then even a version that had all geometry turned off...so nothing was being drawn. And the framerate savings was not as dramatic as you would think. So, the positive aspect of that is that this engine is clearly capable of pushing around a lot of visual detail. But it also means that spending much time to heavily optimize on the art side, or to make a very simple map for example, is not going to give us the biggest impact, compared to other code optimizations that can be done.
That said, we are in the process of testing out an new way of doing the occlusion, which we are hoping will make a difference in framerate.
<!--quoteo(post=1830168:date=Feb 6 2011, 10:28 PM:name=Squeal_Like_A_Pig)--><div class='quotetop'>QUOTE (Squeal_Like_A_Pig @ Feb 6 2011, 10:28 PM) <a href="index.php?act=findpost&pid=1830168"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->That said, we are in the process of testing out an new way of doing the occlusion, which we are hoping will make a difference in framerate.
--Cory<!--QuoteEnd--></div><!--QuoteEEnd-->
This is awesome, I love watching the problems encountered solved by trial and error. This Game has enlightened a lot of people to the inner workings of Small development teams devotion to there fan base!!
I can't wait to here more about what you all find out with the occlusion.
Nice! In my opinion, the engine itself is the work of a programming genius. Max and his buddies are really doing a great job! I'm really interested in how the engine will play with all the customization made in the future, by devs and modders. Keep it up!
<b>@Cory</b>: How about doing the same thing for network issues? Create a minimod stripped to the bare minimum, only marines and rifles (no checks for building, power, scores, other events) and see how that would run with up to 32 players. If that works fine, add skulks in etc.
<!--quoteo(post=1830084:date=Feb 6 2011, 10:07 PM:name=Cpt.j2oss)--><div class='quotetop'>QUOTE (Cpt.j2oss @ Feb 6 2011, 10:07 PM) <a href="index.php?act=findpost&pid=1830084"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Thank you, Cory, for this reply! This is exactly the kind of response I was hoping for; I feel re-assured!<!--QuoteEnd--></div><!--QuoteEEnd-->
I's nice that you're content now, and I realize you've registered only a few days ago, but there's a post like yours at least once per week, and usually UWE replies to them like they did here. So it's getting annoying when the next guy comes in and yells something like "ZOMG why don't you fix performance first?!?!?!!" as if UWE wasn't aware of and working on it, as we speak. And I'm sure next week we'll see another rerun of the same. That's why some people may respond to it in a somewhat rude manner.
<!--quoteo(post=1830168:date=Feb 7 2011, 07:28 AM:name=Squeal_Like_A_Pig)--><div class='quotetop'>QUOTE (Squeal_Like_A_Pig @ Feb 7 2011, 07:28 AM) <a href="index.php?act=findpost&pid=1830168"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Actually, the map detail is really not that much of a problem. --Cory<!--QuoteEnd--></div><!--QuoteEEnd--> Well i found a lot of things which make server lags and other people did, but its even not fixed. For example hydras, more then 3 in one room makes lags, lerk spikes hitting walls (even on local server with only ME!) inrease the ping from like 50 to over 200, i guess same with the gorge spitting, it get stuck on walls too. These things take down the peformence and make lags. Weapon spam for example did this too, and i guess there a lot of other things. Why have the weapon a timeout if they droped (and mostly can't picked up again) but the lerk spikes or gorge spit has none? Anyway good luck! Im excited for 163 with DI. I realy hope DI workds with Occlusion Culling, i am sure it increase peformence a lot if highpoly models get "eaten" by it ^^
<!--quoteo(post=1830194:date=Feb 7 2011, 12:38 PM:name=Vic)--><div class='quotetop'>QUOTE (Vic @ Feb 7 2011, 12:38 PM) <a href="index.php?act=findpost&pid=1830194"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec--><b>@Cory</b>: How about doing the same thing for network issues? Create a minimod stripped to the bare minimum, only marines and rifles (no checks for building, power, scores, other events) and see how that would run with up to 32 players. If that works fine, add skulks in etc.<!--QuoteEnd--></div><!--QuoteEEnd--> I would love to see a 32 player server but most time it start lagging with up to 6vs6 people.
<!--quoteo(post=1830168:date=Feb 7 2011, 06:28 AM:name=Squeal_Like_A_Pig)--><div class='quotetop'>QUOTE (Squeal_Like_A_Pig @ Feb 7 2011, 06:28 AM) <a href="index.php?act=findpost&pid=1830168"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Actually, the map detail is really not that much of a problem. Its mostly other things running behind the scenes. We've even done tests with a version of the map wtih no lighting and shadows, and then even a version that had all geometry turned off...so nothing was being drawn. And the framerate savings was not as dramatic as you would think. So, the positive aspect of that is that this engine is clearly capable of pushing around a lot of visual detail. But it also means that spending much time to heavily optimize on the art side, or to make a very simple map for example, is not going to give us the biggest impact, compared to other code optimizations that can be done.
That said, we are in the process of testing out an new way of doing the occlusion, which we are hoping will make a difference in framerate.
--Cory<!--QuoteEnd--></div><!--QuoteEEnd-->
*Bings: Occlusion*
Nice, I am once again reminded why I could never write a game engine. I would probably only go as far as implementing gravity. Haha.
I've investigated the occlusion performance problem in b162 and base framerate isn't the main problem; the issue is when dynamic occlusion culling calculations are performed and 1-5 frames are completely dropped. The disturbing hitching effect isn't really accounted for in the averaged out fps. If you have 40 fps but then you get an occlusion culling hitch, it's as if you have 7 fps, and that's how the game feels, because you don't notice when it's smooth, only when it hitches. This happens even in NS2_basic if you put your view in the right places.
But then once you rotate your view or move the camera so that the new problem geometry is in view, the fps is fine again. But if you turn so you can't see it, and turn again to include it in your view, it hitches all over again, doing the same occlusion culling calculation that totally HAMMERS the cpu. <b>Running my own game, the server tick rate takes a major dive whenever the occlusion calculation is performed, going from 30 to 8(!)</b>, because my entire computer is being dominated by trying to do the occlusion culling calc. Just from moving my mouse or pressing 'a'. It shouldn't be that way. Based on the angles in the map that maximize this effect, it may have something to do with the <b>interaction of the occlusion calc with spaces which house models/entities</b> like hive/egg/(player?). Can't you make most of those calcs at map load for full view angle at regular intervals through each axis through the playable part of the map geometry, and store them in memory as a rough guide?(vis?) "Dynamic" is a relative term. Or, shouldn't the gpu be doing this calculation? (so it wouldn't affect server tick)
I believe it even affects INPUT! Haven't tried this yet, but as a spectator, I think you can strafe back and forth in the right place, bringing new geometry into/out of view, causing hitching, and time-standardized keypresses will still give positional drifting. So part of what seems to be an input/movement physics problem really also has to do with the cpu load, from occlusion and/or resulting server performance...
<!--quoteo(post=1830168:date=Feb 7 2011, 06:28 AM:name=Squeal_Like_A_Pig)--><div class='quotetop'>QUOTE (Squeal_Like_A_Pig @ Feb 7 2011, 06:28 AM) <a href="index.php?act=findpost&pid=1830168"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Actually, the map detail is really not that much of a problem. Its mostly other things running behind the scenes. We've even done tests with a version of the map wtih no lighting and shadows, and then even a version that had all geometry turned off...so nothing was being drawn. And the framerate savings was not as dramatic as you would think. So, the positive aspect of that is that this engine is clearly capable of pushing around a lot of visual detail. But it also means that spending much time to heavily optimize on the art side, or to make a very simple map for example, is not going to give us the biggest impact, compared to other code optimizations that can be done.
That said, we are in the process of testing out an new way of doing the occlusion, which we are hoping will make a difference in framerate.
--Cory<!--QuoteEnd--></div><!--QuoteEEnd-->
That's actually good to know, it'll be weird to do mapping without having to spend ages thinking about limiting visibility as much as possible, and having a little freedom to splurge on some eye candy, but certainly not unpleasant.
GC has been causing hitches and it seems some progress has been made improving it. Juice makes some good mention of issues with the occlusion as well. I want to know the relation between the client (or server) networking code and all the other client (or server) code. Regardless if the client/server is still making certain calculations it should at least relay some of what it has already calculated (most importantly user input) with a priority queue that prioritizes physics/movement over sound and other effects. This way if we hit the rate limiter at least the most important information for a smooth game still goes through right away. From what I understand at least the GC has its own thread but it appears all other calculations are linear and if there are too many, forces the game to skip frames attempting to catch up giving us our stutters. So besides random code optimizations, it would seem either prioritizing the most important packet information or maybe adding thread(s) for some of the other intensive stuff could smooth out a few of the hitches?
McGlaspiewww.team156.comJoin Date: 2010-07-26Member: 73044Members, Super Administrators, Forum Admins, NS2 Developer, NS2 Playtester, Squad Five Blue, Squad Five Silver, Squad Five Gold, Reinforced - Onos, WC 2013 - Gold, Subnautica Playtester
Wow, the changelog for 163 is going to be a small novel. Imo, we might see a patch this Friday, but more likely next week (considering all of the stuff that's being jammed into this update).
Comments
i hope stability is honestly fixed, making beta playable is usually the main problem right now, and ever been. One of the main reasons I sometimes stop playing is due to my fps spiking like crazy, making game play really slow. The game itself is beta but the engine itself feels like alpha, imo.<!--QuoteEnd--></div><!--QuoteEEnd-->
anyone tried Wolf fire's Overgrowth alpha?
<a href="http://www.wolfire.com/overgrowth" target="_blank">http://www.wolfire.com/overgrowth</a>
most amazingly stable and optimized game engine i've ever seen and its barely in an alpha stage..
just a thought.
Is optimizing the engine not a high enough priority? It should be. I'm trying my hardest to be patient here, I really am, but I'm also starting to get quite worried.
Can I ever expect to get silky smooth performance with this game on my current Q6600, ATI 5770? How much more work needs to be or even CAN be done before the engine performs well for the majority of gamers with modest rigs?
That good?
A dog can run like over 30-35 km/h or more!
Seriously i got like 25-30 fps, most people too and this is ok, not good but for a earlie beta, its ok.
just pretend this is a regular game that you don't have access to and simply DON'T PLAY
A dog can run like over 30-35 km/h or more!
Seriously i got like 25-30 fps, most people too and this is ok, not good but for a earlie beta, its ok.<!--QuoteEnd--></div><!--QuoteEEnd-->
I assume that the reference to 'horrible performance' is regarding the 'glitchiness' mentioned above (as opposed to poor FPS); player movement rubber-banding, collision detection problems, poor hit detection, etc. which can be described generally as a feeling of 'unresponsiveness' or a marked disconnect between the player's actions and the avatar's in-game reactions.
<!--quoteo(post=0:date=:name=weezl)--><div class='quotetop'>QUOTE (weezl)</div><div class='quotemain'><!--quotec-->omg another such post...
just pretend this is a regular game that you don't have access to and simply DON'T PLAY<!--QuoteEnd--></div><!--QuoteEEnd-->
IMO, this kind of answer is <i>unacceptable</i>; it amounts, again, to rug-sweeping the issues away. As stated, we understand that the game is currently in Beta, and that what we have been given access to is a Beta version of the game which should be expected to have some issues. People following your advice is exactly what I'm hoping can be avoided, because simply doing that without raising any criticisms in the process amounts to <i>accepting not being able to play</i> and arguably in many cases--my friend, for example--this means basically giving up on NS2.
Another related point has to do with word of mouth publicity for the game. How do you expect there to be much at all in the way of positive word of mouth for this game--Indie titles such as this one <u>need</u> it--if the vast majority of people with access to the game are too frustrated by its performance to even play a full match?
The game looks great, the concept is exciting and innovative. All I ask is that this issue be officially addressed; something along the lines of, "The reason these issues haven't been resolved yet is __________." It's great that UWE is so involved with the community surrounding this game, but this issue has--for quite some time--been the seldom acknowledged 'elephant in the room' that most people are probably wondering about. As such, the flow of announcements about new feature ideas, prototyping etc, are starting to 'rub-it-in'.
Again: THANK YOU UWE for working on this project - we all appreciate it very much! I'm simply trying to contribute what I can to make sure this game doesn't become a forgotten gem.
P.S.
I like the idea of 'Community-Questions' video features. No offense intended, but if you're going to bother making videos like this; 1. Answer more than one question (r.e. one question about squads?), 2. Answer questions that you can provide more than a <i>vague</i> reply to (r.e. "We're going do to something to make squads better...")
I only mention this because it's something that contributes to the 'rubbing-it-in' I brought up above. In other words, how do you justify spending the time to make videos for question-&-answer if the chosen question pertains to something relatively minor (in comparison to the performance issues which have led to most people not playing) and the response is so vague as to be a non-response?
I apologize if any of that came across as offensive; not my intent at all! I'm just looking for logical explanations....
[Edited for Grammar; added P.S. re-'Community-Questions' video]
We HAVE acknowledged and addressed this on numerous occasions. Max has stated countless times that performance is his #1 priority at the moment, and that is what he has been focused on. Even tasks which may seem on the surface to not be directly related to performance usually are. Its not like there is one problem that is causing all the lag and low framerate, and Max can come on and say we are going to fix X bug and its going to solve all your problems. Max has a list of numerous optimizations that are planned, and hopefully each one of those will keep improving the game more and more.
Its not an exact science. Some optimizations that are expected to make a big impact sometimes result in smaller then expected gains, and vice versa. One optimization may have a large impact, but some other issues may be holding back the gains made from it. Add to that, everyone is running on a completely different system. Some people may see huge gains with one type of graphics card, while people with other cards, don't. Optimizing a game to run on a large spectrum of system specs is a very daunting process, compared to optimizing a game around 1 type of hardware, like an xbox or playstation.
These are all partly reason why we don't want to hype up certain optimizations that are being worked on, because everyone will expect a huge improvement and then be potentially disappointed when the patch comes out and they don't see any difference.
Trust me, Max is as frustrated as any of you. The poor performance eats away at him every day, and he is working hard to improve server lag and framerate. The new features and prototyping you see are coming mostly from Charlie's end (such as the DI at the moment) so its not taking time away from Max's optimizations.
--Cory
as you see nothing is being swept under the rug, neither was it at all what i meant!
point is game development must take it's course, and the developers know best what to do - afterall they want/need the game out most of all!
community whine is pointless and simply would not exist for a regular game (at this stage).
the best we can do is discuss features, bring ideas, feedback, etc!
Ditto. Thanks for that response. My faith has been restored. Keep up the good work, I really do appreciate it. :)
I think this is the easiest, most straight-forward way to boost framerates for people who want smoothness. Surely it can't take more than a single day to make a simple map. Just make it an official map, so everyone can play it. That will tide us over until performance can be improved.
I think this is the easiest, most straight-forward way to boost framerates for people who want smoothness. Surely it can't take more than a single day to make a simple map. Just make it an official map, so everyone can play it. That will tide us over until performance can be improved.<!--QuoteEnd--></div><!--QuoteEEnd-->
How is that different from playing NS1? :)
Things take much longer and there is only so much you can do in a day. You could make a box room map in a day but it wouldn't be enjoyable. Basically, the more time and thought you put into it the richer and diverse the experience. A rushed map is a poor experience for players.
You can make a box room map in like 10 minutes. It shouldn't take long for someone experienced with the editor to make a basic, symmetrical map. You know whats a poor experience for players? Terrible framerates.
That said, we are in the process of testing out an new way of doing the occlusion, which we are hoping will make a difference in framerate.
--Cory
--Cory<!--QuoteEnd--></div><!--QuoteEEnd-->
This is awesome, I love watching the problems encountered solved by trial and error. This Game has enlightened a lot of people to the inner workings of Small development teams devotion to there fan base!!
I can't wait to here more about what you all find out with the occlusion.
I'm really interested in how the engine will play with all the customization made in the future, by devs and modders. Keep it up!
I's nice that you're content now, and I realize you've registered only a few days ago, but there's a post like yours at least once per week, and usually UWE replies to them like they did here. So it's getting annoying when the next guy comes in and yells something like "ZOMG why don't you fix performance first?!?!?!!" as if UWE wasn't aware of and working on it, as we speak. And I'm sure next week we'll see another rerun of the same. That's why some people may respond to it in a somewhat rude manner.
--Cory<!--QuoteEnd--></div><!--QuoteEEnd-->
Well i found a lot of things which make server lags and other people did, but its even not fixed.
For example hydras, more then 3 in one room makes lags, lerk spikes hitting walls (even on local server with only ME!) inrease the ping from like 50 to over 200, i guess same with the gorge spitting, it get stuck on walls too.
These things take down the peformence and make lags.
Weapon spam for example did this too, and i guess there a lot of other things.
Why have the weapon a timeout if they droped (and mostly can't picked up again) but the lerk spikes or gorge spit has none?
Anyway good luck!
Im excited for 163 with DI.
I realy hope DI workds with Occlusion Culling, i am sure it increase peformence a lot if highpoly models get "eaten" by it ^^
<!--quoteo(post=1830194:date=Feb 7 2011, 12:38 PM:name=Vic)--><div class='quotetop'>QUOTE (Vic @ Feb 7 2011, 12:38 PM) <a href="index.php?act=findpost&pid=1830194"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec--><b>@Cory</b>: How about doing the same thing for network issues? Create a minimod stripped to the bare minimum, only marines and rifles (no checks for building, power, scores, other events) and see how that would run with up to 32 players. If that works fine, add skulks in etc.<!--QuoteEnd--></div><!--QuoteEEnd-->
I would love to see a 32 player server but most time it start lagging with up to 6vs6 people.
That said, we are in the process of testing out an new way of doing the occlusion, which we are hoping will make a difference in framerate.
--Cory<!--QuoteEnd--></div><!--QuoteEEnd-->
*Bings: Occlusion*
Nice, I am once again reminded why I could never write a game engine. I would probably only go as far as implementing gravity. Haha.
But then once you rotate your view or move the camera so that the new problem geometry is in view, the fps is fine again. But if you turn so you can't see it, and turn again to include it in your view, it hitches all over again, doing the same occlusion culling calculation that totally HAMMERS the cpu. <b>Running my own game, the server tick rate takes a major dive whenever the occlusion calculation is performed, going from 30 to 8(!)</b>, because my entire computer is being dominated by trying to do the occlusion culling calc. Just from moving my mouse or pressing 'a'. It shouldn't be that way. Based on the angles in the map that maximize this effect, it may have something to do with the <b>interaction of the occlusion calc with spaces which house models/entities</b> like hive/egg/(player?). Can't you make most of those calcs at map load for full view angle at regular intervals through each axis through the playable part of the map geometry, and store them in memory as a rough guide?(vis?) "Dynamic" is a relative term. Or, shouldn't the gpu be doing this calculation? (so it wouldn't affect server tick)
I believe it even affects INPUT! Haven't tried this yet, but as a spectator, I think you can strafe back and forth in the right place, bringing new geometry into/out of view, causing hitching, and time-standardized keypresses will still give positional drifting. So part of what seems to be an input/movement physics problem really also has to do with the cpu load, from occlusion and/or resulting server performance...
Just my uninformed speculations.
i was wondering if ns2_mineshaft will be part of build 163? cant wait to see it :)
That said, we are in the process of testing out an new way of doing the occlusion, which we are hoping will make a difference in framerate.
--Cory<!--QuoteEnd--></div><!--QuoteEEnd-->
That's actually good to know, it'll be weird to do mapping without having to spend ages thinking about limiting visibility as much as possible, and having a little freedom to splurge on some eye candy, but certainly not unpleasant.
* Decals
* Trailing flamethrower effects
* ARC playable (targeting, cooldown, etc.)
Though, nice to see the gameplay elements coming together, finally a way to deal with locked doors on Tram!