<!--quoteo(post=1906483:date=Feb 24 2012, 05:02 AM:name=RockyMarc)--><div class='quotetop'>QUOTE (RockyMarc @ Feb 24 2012, 05:02 AM) <a href="index.php?act=findpost&pid=1906483"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I agree, the performance has been improving a heap considering a few patches back.<!--QuoteEnd--></div><!--QuoteEEnd-->
Has it really though or are this subjective feelings? I've been doing benchmarks on ns2_tram every few builds (as repeatable as it is currently possible). Between build 162 - which was released about a year ago - and build 195, the average frame rate has increased by only 15% on my machine. UWE needs to implement some massive improvements over the next months, so that players with average PCs can achieve acceptable frame rates. Not to mention server performance. Only 6 month left, I'm getting a little worried.
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
edited February 2012
<!--quoteo(post=1906520:date=Feb 24 2012, 09:18 AM:name=Raza.)--><div class='quotetop'>QUOTE (Raza. @ Feb 24 2012, 09:18 AM) <a href="index.php?act=findpost&pid=1906520"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Has it really though or are this subjective feelings? I've been doing benchmarks on ns2_tram every few builds (as repeatable as it is currently possible). Between build 162 - which was released about a year ago - and build 195, the average frame rate has increased by only 15% on my machine. UWE needs to implement some massive improvements over the next months, so that players with average PCs can achieve acceptable frame rates. Not to mention server performance. Only 6 month left, I'm getting a little worried.<!--QuoteEnd--></div><!--QuoteEEnd-->
Hmm... as always when it comes to doing performance measurements, you really have to be careful with what you measure.
Looking at server performance, back in 162 it was extremly easy to choke a high-performance server (ie get its tick rate < 10) by just building a map with a bunch of structures in t. The server would choke even with just one player.
On the same server today, I can't get the server to choke even with three-four times more structures on it - as long as there is only one player on it.
Likewise, you can (barely) play 12v12 player games on really good hardware. Back in 162, you needed really good hardware to play 6v6.
I remember the first time we PT central drilling in Mineshaft ... with atmospherics turned on, noone got more than 10 fps, and most got a lot less. On an empty map.
Same location today gives me about 50 fps (granted, single-player).
So performance has increased a lot more than the 15% you mention.
That said, yea, another factor of 2 is needed before release. But that's doable.
<!--quoteo(post=1906568:date=Feb 24 2012, 08:46 AM:name=matso)--><div class='quotetop'>QUOTE (matso @ Feb 24 2012, 08:46 AM) <a href="index.php?act=findpost&pid=1906568"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Hmm... as always when it comes to doing performance measurements, you really have to be careful with what you measure.
Looking at server performance, back in 162 it was extremly easy to choke a high-performance server (ie get its tick rate < 10) by just building a map with a bunch of structures in t. The server would choke even with just one player.
On the same server today, I can't get the server to choke even with three-four times more structures on it - as long as there is only one player on it.
Likewise, you can (barely) play 12v12 player games on really good hardware. Back in 162, you needed really good hardware to play 6v6.
I remember the first time we PT central drilling in Mineshaft ... with atmospherics turned on, noone got more than 10 fps, and most got a lot less. On an empty map.
Same location today gives me about 50 fps (granted, single-player).
So performance has increased a lot more than the 15% you mention.
That said, yea, another factor of 2 is needed before release. But that's doable.<!--QuoteEnd--></div><!--QuoteEEnd-->
i think it would be a mistake to just aim for 2x improvement, especially server side. even now only a few servers can run the game completely smoothly 6v6, the vast majority of servers will not be able to. to put it bluntly, right now it's probably as laggy on a normal server as it was on the top-end servers when the alpha hit. client side it shouldn't be a big deal, i could live with a small increase to my avg fps, 25-50%, as long as it was -stable-. that's the key, you can't expect good results/press/word of mouth if you release in a state where fps takes sharp hits during combat.
*edit* also are there any plans to up the server tickrate/update rate? 30/20 is awfully low.
A factor 2x is a very low estimation, given that even 4-5GHz machines now cannot sustain a 60FPS at all times (and we haven't even tried things like 16v16 yet). Moreover a 5GHz server is currently able to do 10v10 I guess, and hold a 20+ tickrate (guesstimating). But that needs to up to 16v16 with a dedicated core of 2.5GHz, which brings us to a factor of roughly 3-4x. And as Wheeee mentioned, we're currently stuck at the pretty low tick-rate of 30, while 60 or higher would be ideal, so make that a factor 6-8x (excuse the horrible oversimplification of what is inherently a complicated issue).
If only 2x is realised for servers, we'll still be seeing A LOT of pretty bad servers (and as such, games) on release-day.
<!--quoteo(post=1906599:date=Feb 24 2012, 09:51 AM:name=player)--><div class='quotetop'>QUOTE (player @ Feb 24 2012, 09:51 AM) <a href="index.php?act=findpost&pid=1906599"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->A factor 2x is a very low estimation, given that even 4-5GHz machines now cannot sustain a 60FPS at all times (and we haven't even tried things like 16v16 yet). Moreover a 5GHz server is currently able to do 10v10 I guess, and hold a 20+ tickrate (guesstimating). But that needs to up to 16v16 with a dedicated core of 2.5GHz, which brings us to a factor of roughly 3-4x. And as Wheeee mentioned, we're currently stuck at the pretty low tick-rate of 30, while 60 or higher would be ideal, so make that a factor 6-8x (excuse the horrible oversimplification of what is inherently a complicated issue).
If only 2x is realised for servers, we'll still be seeing A LOT of pretty bad servers (and as such, games) on release-day.<!--QuoteEnd--></div><!--QuoteEEnd-->
This is insane. Serious question and not a troll, why didn't UW use source again? Dynamic Infestation and lighting effects?
its pretty scary seeing these "basic requirements" needed to run ns2 while many of the people here can run other games on ultra settings. (for example BF3 or skyrim)
It's already been said, but we're always working on performance. Some patches will have big improvements, some will have none, and some may take a step backwards (since we're still developing a lot of game code). In Build 189 we started collecting stats on the average frame rate. This graph shows the breakdown of FPS for players, where red is between 0 and 20 FPS and green is > 45 FPS.
[attachment=36063:stats.png]
You can see that we went from 30% getting over 45 FPS in 189 to 60% getting over 45 FPS in 197. There are some baises in the data, but the should make it clear that -- despite what a some comments seem to suggest -- we do work on performance.
if you've gone to one of the pub servers that aren't top end, they're really suffering once playercount goes >4. like i said, i'm confident the client side issues will be fine, but server performance imo is the x factor. consider that if you have 20,000 unit sales at launch, you'll need at the very least 200 20-slot servers assuming that 20% of players will be online at peak times... (for example, x-box statistics show that about 25% of users are on during peak times). i'm willing to bet there aren't 200 server operators willing to run inversion-level boxes for that. that's really all i'm worried about...the game is playable, and i think if the server performance was improved to the point where you could get 40-50 ticks/updates, the rest would sort itself out.
AsranielJoin Date: 2002-06-03Member: 724Members, Playtest Lead, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow, Subnautica Playtester, Retired Community Developer
Very cool graph. Indeed the game bacame fast for me. But what would be more interesting would be fps during fights (where it matters most), which is much lower.
ScardyBobScardyBobJoin Date: 2009-11-25Member: 69528Forum Admins, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow
Part of the issue server-wise is that when server performance increases, server owners up their playercount until performance drops to close to the previous level. Server playercount has increased steadily from the 4-8 in the alpha, to the 10-14 early beta, to the current 16-20.
Comments
<!--quoteo(post=1906483:date=Feb 24 2012, 05:02 AM:name=RockyMarc)--><div class='quotetop'>QUOTE (RockyMarc @ Feb 24 2012, 05:02 AM) <a href="index.php?act=findpost&pid=1906483"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I agree, the performance has been improving a heap considering a few patches back.<!--QuoteEnd--></div><!--QuoteEEnd-->
Has it really though or are this subjective feelings?
I've been doing benchmarks on ns2_tram every few builds (as repeatable as it is currently possible). Between build 162 - which was released about a year ago - and build 195, the average frame rate has increased by only 15% on my machine. UWE needs to implement some massive improvements over the next months, so that players with average PCs can achieve acceptable frame rates. Not to mention server performance. Only 6 month left, I'm getting a little worried.
On tablet, can't embed...
I've been doing benchmarks on ns2_tram every few builds (as repeatable as it is currently possible). Between build 162 - which was released about a year ago - and build 195, the average frame rate has increased by only 15% on my machine. UWE needs to implement some massive improvements over the next months, so that players with average PCs can achieve acceptable frame rates. Not to mention server performance. Only 6 month left, I'm getting a little worried.<!--QuoteEnd--></div><!--QuoteEEnd-->
Hmm... as always when it comes to doing performance measurements, you really have to be careful with what you measure.
Looking at server performance, back in 162 it was extremly easy to choke a high-performance server (ie get its tick rate < 10) by just building a map with a bunch of structures in t. The server would choke even with just one player.
On the same server today, I can't get the server to choke even with three-four times more structures on it - as long as there is only one player on it.
Likewise, you can (barely) play 12v12 player games on really good hardware. Back in 162, you needed really good hardware to play 6v6.
I remember the first time we PT central drilling in Mineshaft ... with atmospherics turned on, noone got more than 10 fps, and most got a lot less. On an empty map.
Same location today gives me about 50 fps (granted, single-player).
So performance has increased a lot more than the 15% you mention.
That said, yea, another factor of 2 is needed before release. But that's doable.
Looking at server performance, back in 162 it was extremly easy to choke a high-performance server (ie get its tick rate < 10) by just building a map with a bunch of structures in t. The server would choke even with just one player.
On the same server today, I can't get the server to choke even with three-four times more structures on it - as long as there is only one player on it.
Likewise, you can (barely) play 12v12 player games on really good hardware. Back in 162, you needed really good hardware to play 6v6.
I remember the first time we PT central drilling in Mineshaft ... with atmospherics turned on, noone got more than 10 fps, and most got a lot less. On an empty map.
Same location today gives me about 50 fps (granted, single-player).
So performance has increased a lot more than the 15% you mention.
That said, yea, another factor of 2 is needed before release. But that's doable.<!--QuoteEnd--></div><!--QuoteEEnd-->
i think it would be a mistake to just aim for 2x improvement, especially server side. even now only a few servers can run the game completely smoothly 6v6, the vast majority of servers will not be able to. to put it bluntly, right now it's probably as laggy on a normal server as it was on the top-end servers when the alpha hit. client side it shouldn't be a big deal, i could live with a small increase to my avg fps, 25-50%, as long as it was -stable-. that's the key, you can't expect good results/press/word of mouth if you release in a state where fps takes sharp hits during combat.
*edit* also are there any plans to up the server tickrate/update rate? 30/20 is awfully low.
If only 2x is realised for servers, we'll still be seeing A LOT of pretty bad servers (and as such, games) on release-day.
If only 2x is realised for servers, we'll still be seeing A LOT of pretty bad servers (and as such, games) on release-day.<!--QuoteEnd--></div><!--QuoteEEnd-->
This is insane. Serious question and not a troll, why didn't UW use source again? Dynamic Infestation and lighting effects?
<a href="https://www.youtube.com/watch?v=1Isjgc0oX0s&feature=related" target="_blank">https://www.youtube.com/watch?v=1Isjgc0oX0s...feature=related</a>
i couldn't help it.
[attachment=36063:stats.png]
You can see that we went from 30% getting over 45 FPS in 189 to 60% getting over 45 FPS in 197. There are some baises in the data, but the should make it clear that -- despite what a some comments seem to suggest -- we do work on performance.
Indeed the game bacame fast for me. But what would be more interesting would be fps during fights (where it matters most), which is much lower.
They even take time to answer annoying questions in the forums. :p
@nuflow hi!