Is there really no way of lowering the performance floor?
Pyromaniac
Join Date: 2009-02-20 Member: 66498Members
After reading the <a href="http://www.ns2hd.com/2012/05/feature-creep.html" target="_blank">NS2HD blog entry</a> on the release performance goal I am disappointed. I have essentially the same machine as max and I consider 45 fps to be unenjoyable for a game like NS2. Perhaps with the <a href="http://web.archive.org/web/20100728234447/http://www.naturalselection2.com/faq/game-questions/what-are-the-system-requirements-and-when-will-ns2-be-releas.html" target="_blank">original requirements</a> the developers weren't being dishonest, but many consumers like myself have been effectively been bait and switched. Before people point me in the direction of refunds, I'm really not concerned over my $40, I'd just like a game I and many others without new computers can enjoy.
TF2 is being used as a standard, but TF2 has the ability to lower settings to the point where i can get it to rarely dip below 100fps on the machine that they claim gets 45fps in the game. I understand that lua is holding much of the performance back, but is there really no way to get significant gains from allowing further simplified effects and graphics? Even in cpu limited games I've been able to get large performance gains from lowering effects settings. Is this possible with NS2?
What niche multiplayer game that required a good computer to run well but didn't have a massive marketing budget was sucessful? Not many, and the current player numbers aren't hurting this case. The people that supported NS2 from the beginning don't care about how nice the game will look, they just want a smooth experience in a fun game. That's why they still like and support NS.
TF2 is being used as a standard, but TF2 has the ability to lower settings to the point where i can get it to rarely dip below 100fps on the machine that they claim gets 45fps in the game. I understand that lua is holding much of the performance back, but is there really no way to get significant gains from allowing further simplified effects and graphics? Even in cpu limited games I've been able to get large performance gains from lowering effects settings. Is this possible with NS2?
What niche multiplayer game that required a good computer to run well but didn't have a massive marketing budget was sucessful? Not many, and the current player numbers aren't hurting this case. The people that supported NS2 from the beginning don't care about how nice the game will look, they just want a smooth experience in a fun game. That's why they still like and support NS.
This discussion has been closed.
Comments
Go play the game and work on other parts of it, make the most of it, this ###### takes time, deal with it.
'45 FPS' is utterly unplayable. It'd be doable if it was consistent but it's going to turn into 10-20 FPS late game.
You either need a bigger performance margin or you need to prevent late-game battles with 20 structures, 10 players all visible at the same time, with struggling server performance.
The devs stated many times that the performance will get better. They want to get the game feature complete first. And with most patches there are actually performance increases, like the last build.
Though seriously. Does no one else remember playing Half-Life 1 on a Pentium 233 with a Monster3D card, with a whole 2MB texture memory, and when 10-15fps was the 'playable' cutoff? And unplayable was counted in seconds-per-frame? Spoiled <i>kids</i> nowadays. I remember the first time I saw Thresh playing Quake 2 running at a solid 30fps and thought it was the smoothest, silkiest, most unbelievable performance I'd ever laid eyes on, and spending days trying to figure out exactly how much it would cost to put together a rig THAT powerful.
Yes, it takes time, but I'm rather confident that it'll be worth it.
Your 45fps compared to the amount of content they want to impelement in the game, is insignificant. We are ALL experiencing this, it's not just you.
In summary, there is nothing you can do except post actual constructive threads which would actually help point out substantial issues to developers. This stage of game development is strictly designed for weeding out problems which are hard to reproduce in internal-only testing (Alpha stage). Your expectations that you deserve a finished game at this stage are completely unfounded and childish.
The game is already really great, enjoy what you have, and look forward to coming patches.
<!--quoteo(post=1937872:date=May 21 2012, 03:06 AM:name=Pyromaniac)--><div class='quotetop'>QUOTE (Pyromaniac @ May 21 2012, 03:06 AM) <a href="index.php?act=findpost&pid=1937872"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->So far most people have dismissed the topic while failing to address the main point: Is it possible to sacrifice looks for performance any further? If it is, please make this an option.<!--QuoteEnd--></div><!--QuoteEEnd-->
Your 45fps compared to the amount of content they want to impelement in the game, is insignificant. We are ALL experiencing this, it's not just you.
In summary, there is nothing you can do except post actual constructive threads which would actually help point out substantial issues to developers. This stage of game development is strictly designed for weeding out problems which are hard to reproduce in internal-only testing (Alpha stage). Your expectations that you deserve a finished game at this stage are completely unfounded and childish.
The game is already really great, enjoy what you have, and look forward to coming patches.<!--QuoteEnd--></div><!--QuoteEEnd-->
Sehr GUT
Your 45fps compared to the amount of content they want to impelement in the game, is insignificant. We are ALL experiencing this, it's not just you.
In summary, there is nothing you can do except post actual constructive threads which would actually help point out substantial issues to developers. This stage of game development is strictly designed for weeding out problems which are hard to reproduce in internal-only testing (Alpha stage). Your expectations that you deserve a finished game at this stage are completely unfounded and childish.
The game is already really great, enjoy what you have, and look forward to coming patches.<!--QuoteEnd--></div><!--QuoteEEnd-->
You trolling? I addressed the lua bottleneck in the first post. This game is in beta not alpha and I was speaking on release projections, which have historically been overly optimistic anyway. I never claimed to deserve a finished game at this stage, I implied it for release though.
The performance bottleneck of NS2 are not the graphics, but the insane CPU usage. Both client and serverside.
It's the result of LUA beeing used to an extreme extent.
Imho. it can only be resolved by massively porting tasks carried out by LUA to C++. While it would take a <b>long</b> time, it's the only way to get really noteable results.
There won't be any huge performance increases by just "optimizing" LUA, like in 208. Theres also not a lot of stuff left, to optimize.
Maybe simplified animations ?
Approximation of the prediction/interpolation maybe, although this would affect the gameplay quite a bit.
It's the result of LUA beeing used to an extreme extent.<!--QuoteEnd--></div><!--QuoteEEnd-->
I understand this, but I wonder if cutting the fat with any cpu intensive visual effects would free up some resources and cause an improvement. I can usually lower the effects setting in source games and see a huge performance increase at almost no visual cost.
Yeah, but you can't really compare NS2 to any other game right now.
The problems are almost purely based on CPU usage and lack of proper multithreading. (The first game I ever encountered with a bottleneck like that.)
Go ahead and try setting the details on "Ridiculously Awful", as well as trying this:
<!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->r_atmospherics false
r_bloom false
r_shadows false
r_instancing false
r_fog false
r_aa false<!--c2--></div><!--ec2-->
You will see that there won't be a lot of a difference.
I hope you do know more than me, to be so optimistic. =/
Yes, because it isn't 60. Do you have a 45hz monitor?
<!--quoteo(post=1937852:date=May 21 2012, 02:54 AM:name=Papayas)--><div class='quotetop'>QUOTE (Papayas @ May 21 2012, 02:54 AM) <a href="index.php?act=findpost&pid=1937852"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Actually, I have been quite impressed with the performance of build 208, and how is 45 fps unplayable? If you are getting that, consider yourself lucky!<!--QuoteEnd--></div><!--QuoteEEnd-->
It's unplayable because it isn't 60. In PC games, we're used to a higher standard than 625 line video at 30 fps, and our games are usually designed with this in mind (things move quickly and require actual aim to be shot).
Also, I don't see how this is a 'troll' or whatever idiots are claiming it to be. The guy wants to know about lowering the performance floor - he's not just saying "WAAH PERFORMANCE SUCKS MAKE BETAR". He's pointing directly at one (often overlooked) aspect and trying to discuss it. If you call that a troll, chances are you're just trying to avoid the issue and should instead not post in the thread!
Also here's a cute summary of page 1: kids think they're in on a big secret because they know the LUA interpreter layer is the performance bottleneck, so they keep repeating it until someone gives them a pat on the back and a gold star (even though it's mentioned in the first post and is not the topic of discussion)
I have tried this by manually lowering the resolution of alien and marine models to very (read: VERY) small sizes. The fact that the FPS didn't change much, only goes to show the rendering engine is not the cause of the performance hit. The devs can really start focusing on optimisations once the game is feature complete (which wont be too long now :P), but even now the game has gotten performance increases (build208 had a big improvement)
The only thing that changed for me was the hitching using this low res test, which was gone on the lower material/texture resolution. But the fact that the game does not use lower poly models yet for lower graphics settings, might be a next step once it is feature complete. But the render engine itself is very efficient so this might not even be needed once lua issues are addressed along with more optimisations.
It's unplayable because it isn't 60. In PC games, we're used to a higher standard than 625 line video at 30 fps, and our games are usually designed with this in mind (things move quickly and require actual aim to be shot).<!--QuoteEnd--></div><!--QuoteEEnd-->
Er. As a long time PC gamer, 60fps is not the minimum 'playable'. Even 30fps is much more than playable. I'd call it 'unplayable' around 15-20fps. *Requiring* more just makes you sound like an entitled prat.
Then again, I use vsync and triple buffering (when applicable) so I don't get tearing. And as most LCDs aren't 60hz either (they just report that to the OS as a communal update rate, their own update rate differs significantly due to the differences in how a CRT updates the screen versus an LCD), it's a bit moot as far as <i>that</i> old goose-egg goes. Unless you're playing on a rather low-spec CRT monitor still (as my remaining two CRTs are running at 100hz)... which you most likely aren't.
It's been stated repeatedly that the GPU is not generally what is causing the current framerates. You could turn Marines into single-polygon cutouts and turn down the visuals until you're running around in untextured boxes and still only get 45fps. Waiting for the CPU threading to be fixed and/or the single process to be optimized even more heavily is about the only thing to do at this point.
In a fast paced, competitive multiplayer game it is an extrarordinarily punishing disadvantage to play with only 30 FPS.
The primary factor in being a good player must not be to get a good CPU and overclock it to ~4-5 GHz.
<!--quoteo(post=1937951:date=May 21 2012, 11:40 AM:name=Talesin)--><div class='quotetop'>QUOTE (Talesin @ May 21 2012, 11:40 AM) <a href="index.php?act=findpost&pid=1937951"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I'd call it 'unplayable' around 15-20fps.<!--QuoteEnd--></div><!--QuoteEEnd-->
45 FPS average <b>IS</b> 15-20 FPS end game.
<!--quoteo(post=1937951:date=May 21 2012, 11:40 AM:name=Talesin)--><div class='quotetop'>QUOTE (Talesin @ May 21 2012, 11:40 AM) <a href="index.php?act=findpost&pid=1937951"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->*Requiring* more just makes you sound like an entitled prat.<!--QuoteEnd--></div><!--QuoteEEnd-->
This isn't the quake days. Players have many options. They can go and play a game that actually performs well instead of one that performs frustratingly bad. If NS2 can't wring decent performance out of mid-range hardware it won't sell very well, it won't be played very much and it will die an early death.
Well, you're assuming that problem is going to persist no matter what.
No. I'm assuming that it is a <i>problem</i> and that it <i>has to be solved</i>. And Talesin et al are defending the current state of affairs as being somewhere close to <i>good enough</i>.
Not compared to quake 3 or NS, but it is definitely fast paced compared to the standard fare of slow-as-treacle world war II shooters, COD, battlefield etc.
If you find 45fps on that machine to be 'unenjoyable,' you consider TF2, Left 4 Dead, and other games unenjoyable. The entire point of the target is that we are trying to get close to Source games on Max's rig.
<!--sizeo:3--><span style="font-size:12pt;line-height:100%"><!--/sizeo--><b>Asking us to beat <!--coloro:#FF8C00--><span style="color:#FF8C00"><!--/coloro-->Source<!--colorc--></span><!--/colorc-->, an extremely mature engine built by a very large development house, is more than a little bit unreasonable. </b><!--sizec--></span><!--/sizec-->
While we will be striving to get fps as high as we can, you do need to keep a little bit of perspective on your wishes for this game. Requesting the impossible will leave you nothing but disappointment.