dePARAJoin Date: 2011-04-29Member: 96321Members, Squad Five Blue
The point is (and im repeating myself): 40fps in NS2 didnt feel like 40fps in other games. There is something wrong, cant exaplain it. They feel like 10-15.
It must be something withing the engine cause they feel "wrong" even if im alone on a server (nothing to predict) and limiting to 40 fps.
JektJoin Date: 2012-02-05Member: 143714Members, Squad Five Blue, Reinforced - Shadow
edited May 2013
If you compare using an fps limiter 20-40 fps in a source game, and 20-40 fps in NS2. You might notice the disparity in the feel of your input if you're sensitive to framerates and input delays.
ScardyBobScardyBobJoin Date: 2009-11-25Member: 69528Forum Admins, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow
It'd be interesting if people recorded 30fps video in a bunch of different first-person shooters and uploaded it to YouTube. It'd be much easier to see if 30fps in NS2 looks worse than other games in a side-by-side comparison.
JektJoin Date: 2012-02-05Member: 143714Members, Squad Five Blue, Reinforced - Shadow
I wouldn't upload to youtube if I wanted to test that to be honest. I couldn't do it with my 1960s Australian internet, but someone could - need to upload to a host that doesn't butcher the video though.
current1y, do you have FRAPS? Perhaps you could do a frametime recording of:
1) running the game while standing in the ready room of a local private server with the maxfps command and your usual graphics settings
2) running the game in a situation where you consistently get low fps with the exact same graphics settings as above, without artificially capping your fps with maxfps
40 actual fps in NS2 definitely does not feel like 10-15 fps for me. It feels kinda bad, but absolutely not that bad.
I don't know if you could watch a video of the different FPS', and visually SEE the difference. Its not really a hard thing to reproduce, start NS2 and type maxfps 50, then launch NS1 or TF2 or CSS and fps_max 50 and see/feel the difference.
Its gotta be the variance in frames, given how the fps output visually to the user is averaged out.
That's why I asked for a frametimes recording.
r_stats displays an averaged frametime (because its based off of an averaged fps) in ms just next to the FPS count.
There's other ways of doing it.. but yea, i've seen it in graphs and that's what it is, imo.
Good news is future optimizations look to make this far better, but no promises from me
Hi all,
We're all aware of the performance issues and are doing our best. In a sense, every week is a performance jam for us. We're always trying things and open to new ideas, and some things work out and some things don't. Believe me, we're not happy about the game's performance either.
--Steve
you know why ppl started to complain about performance?
cause you silently quit your open development strategy!
bring back the pivotal tracker, stop the secretiveness
and show us what you are working on right now.
screw the marketing bang (and the marketing guy,
make him a back office guy, but not a front person)
and stay honest again with your community.
then the trust MAY come back after some time.
i dont want ironhorse or scardybob
to respond to my post!
The point is (and im repeating myself): 40fps in NS2 didnt feel like 40fps in other games. There is something wrong, cant exaplain it. They feel like 10-15.
It must be something withing the engine cause they feel "wrong" even if im alone on a server (nothing to predict) and limiting to 40 fps.
I swear I have been saying this for a long time, I'm starting to feel a bit crazy every time it says 40-60 ( running mostly on low ) it still feels like 15-20 at max I really don't know what is causing it, and the crazy thing is that the game started to run worse and worse after a couple of patches after release ( can't remember when or which patches ), I'm even getting worse performance in Combat mod ( which should run smoother and it did for me even during combat, no pun intended ) around the same time.
I have tried updating drivers, installing to another HDD, verifying steam catch I even OCed my CPU from 2.4ghz up to 3.2 and with time I kept lowering my graphical options and resolution and running windowed to be able to play but still same problem.
System
Windows 7 ultimate 64bit
Intel q6600 @ 3.2ghz
4gb ram
hd5850 1gb
While, whether it's correct or not, I agree the FPS feels lower than it is presented to be in gameplay, I also believe it has been said that the FPS limiter isn't working very well and only really meant for developers to test stuff other than the "feel" or smoothness of the game, as it sort of breaks when using the Spark FPS limiter.
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
maxfps was added to simplify testing and verification for some nasty issues with low-fps situations (setting up a situation where you had enough load to actually get your fps down was easy early on ... but got progressively harder), it has basically no other use.
And while saying 40 fps feels like 15 fps in other games might be an exaggeration, there are some issues that may cause the quality of the fps to drop when the client gets loaded - mostly because frame time may start to vary a lot when there is a lot of stuff happening, so those 40 fps does not mean every frame is 25ms - one or two may clock in at 75ms while most clock in at 20ms...
I'm fairly optimistic that things will improve quite a lot in the not-to-far-future though; the latest JIT testing gave pretty good results - well, apart from the crashing :-/
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
Lol.. Yeah multiple posts from multiple devs in this thread alone on the topic of performance sure does sound "secretive" to me. :rolls eyes :
Pivotal tracker is not coming back, so recognizing this, pts like Scardybob and myself do our best to answer what we can and be highly accessible /informational. Sorry this isn't good enough. What answer do you not get from us that you think a dev would answer anyways?
P. S. I highly doubt people complain about performance due to a lack of a pivotal tracker.. ? I think they complain about it because they have poor performance.
Lol.. Yeah multiple posts from multiple devs in this thread alone on the topic of performance sure does sound "secretive" to me. :rolls eyes :
Yeah, If only the UWE Devs would just say something! like maybe in a thread like this one, on this forum, in the form of a post... maybe even two posts above this one your reading right now!
ScatterJoin Date: 2012-09-02Member: 157341Members, Squad Five Blue
In fact there have been no improvement gains, it's all a lie ! The reality is this:
DisplayedFPS = (1+ UWEclaimedFPSgain) * ActualFPS
We're all sitting on 20 fps even though it says 80 fps ! Seriously though it really does feel like it which I why I have a fit when people say "yeah I get 30 - 50 fps which Is fine why are you complaining. Oh yeah and the human eye can't see more than 60 fps BTW"
And while saying 40 fps feels like 15 fps in other games might be an exaggeration, there are some issues that may cause the quality of the fps to drop when the client gets loaded - mostly because frame time may start to vary a lot when there is a lot of stuff happening, so those 40 fps does not mean every frame is 25ms - one or two may clock in at 75ms while most clock in at 20ms...
That's exactly what I want to see with the recordings I suggested above.
I'm fairly optimistic that things will improve quite a lot in the not-to-far-future though; the latest JIT testing gave pretty good results - well, apart from the crashing :-/
Good to hear that. It would be great to get an estimate on how much of a performance boost we can expect!
I don't know if you could watch a video of the different FPS', and visually SEE the difference. Its not really a hard thing to reproduce, start NS2 and type maxfps 50, then launch NS1 or TF2 or CSS and fps_max 50 and see/feel the difference.
I wonder if this is more of an fps consistency issue than anything else. I've noticed that NS2's fps fluctuates quite a bit more than I've seen in other fps games even when it maintains a high average fps.
I liked this discussion of the FCAT/framerating method as it gets into this type of consistency issue:
Good to hear that. It would be great to get an estimate on how much of a performance boost we can expect!
Believe me i want to share these graphs, but.. *
I think its best to wait for everything to be final and stable and go through thorough testing and benchmarking over many systems before estimation numbers are publicly given..
Else you will end up with people who had their expectations high and were upset about a failed delivery etc..
That being said, i'm pretty sure you can read between the lines when @matso says he's fairly optimistic and i keep posting about whats going to be fixed (variance in fps and frametimes)
:-B
*Edit: I got permission to post the graphs from Hugh!
***DISCLAIMER**
Again.. This is very much so a WIP and does not represent the final implementation, nor represents every scenario, and it may not even reflect your computer's results.
This graph is merely to demonstrate the lack of variance in fps and "smoothness" accomplished by current LUAJIT progress - actual fps increases depend on scenarios (entity heavy) and hardware - I have seen as little as 2 fps gain for some, and 40 in others .. so take this pic with a grain of salt. It was made by @matso late last night after we tried it out in PT.
Hi all,
We're all aware of the performance issues and are doing our best. In a sense, every week is a performance jam for us. We're always trying things and open to new ideas, and some things work out and some things don't. Believe me, we're not happy about the game's performance either.
--Steve
Have you tried making a serious attack on "performance" issues that do not affect average frame rates?
Like input latency. Can you sample input more than once per frame to try and keep movement physics snappier and more precise even at lower framerates?
Like frame time variance. 60 FPS average with every other frame taking twice as long does not feel like 60 fps.
When you spin around quickly frame times increase; as it happens, spinning around quickly is something that often occurs in combat.
The framerate rate drop during combat is large, and enormous during heavy combat, and that's precisely the time you least want to drop frames. Can you strip back the most CPU demanding effects during combat, especially heavy combat?
Just being around lots of players seem to be a chief culprit, structures less so. Would some kind of LoD system allow you to get big savings on the animination system or whatever it might be that causes poor performance when lots of players are around?
can you count actual frametimes (not fraps frametimes) without what is essentially a capture card or a ridiculously high speed camera?
FRAPs frametimes are accurate so I'm not sure why you wouldn't use that. However, you can use in-game plog function and performance analyzer to do the same thing. Basically,
Start plog capture = p_logall
End plog capture = p_endlog
You plog should output to the same directory as your log.txt file (default is your appdata folder). Then go to \SteamApps\common\Natural Selection 2\utils\PerfAnalyzer\ folder and run the plog file through the perfanalyzer.py program (will need to install Python on your computer to get this to work). That will get you the graphs Ironhorse posted earlier.
The plog function is basically just outputting all the data you see from the profiler to a file you can examine later. The only downside is that there is no way (that I know yet) to convert the data into something readable by a spreadsheet program.
LuaJIT looks really promising. Thanks for sharing that IronHorse. I'm aware that this is not representative of the final product. Any performance gain is welcome, especially in heavy load scenarios.
Funnily enough I have noticed I get better FPS when I have certain settings on high than with them on low.
Mind sharing what settings you personally have on high?
I just recently upgraded my graphics card and from the experience from the two cards in this game, it depends heavily on the card in use. Just put fps counter on and fiddle with them, on my first card things like texture quality, anisotropic filtering and reflections, on this second one, even anti-aliasing and shadows on gives more fps than with them off, on the first card it was opposite with those two.
Comments
It must be something withing the engine cause they feel "wrong" even if im alone on a server (nothing to predict) and limiting to 40 fps.
1) running the game while standing in the ready room of a local private server with the maxfps command and your usual graphics settings
2) running the game in a situation where you consistently get low fps with the exact same graphics settings as above, without artificially capping your fps with maxfps
40 actual fps in NS2 definitely does not feel like 10-15 fps for me. It feels kinda bad, but absolutely not that bad.
That's why I asked for a frametimes recording.
There's other ways of doing it.. but yea, i've seen it in graphs and that's what it is, imo.
Good news is future optimizations look to make this far better, but no promises from me
you know why ppl started to complain about performance?
cause you silently quit your open development strategy!
bring back the pivotal tracker, stop the secretiveness
and show us what you are working on right now.
screw the marketing bang (and the marketing guy,
make him a back office guy, but not a front person)
and stay honest again with your community.
then the trust MAY come back after some time.
i dont want ironhorse or scardybob
to respond to my post!
I swear I have been saying this for a long time, I'm starting to feel a bit crazy every time it says 40-60 ( running mostly on low ) it still feels like 15-20 at max I really don't know what is causing it, and the crazy thing is that the game started to run worse and worse after a couple of patches after release ( can't remember when or which patches ), I'm even getting worse performance in Combat mod ( which should run smoother and it did for me even during combat, no pun intended ) around the same time.
I have tried updating drivers, installing to another HDD, verifying steam catch I even OCed my CPU from 2.4ghz up to 3.2 and with time I kept lowering my graphical options and resolution and running windowed to be able to play but still same problem.
System
Windows 7 ultimate 64bit
Intel q6600 @ 3.2ghz
4gb ram
hd5850 1gb
You don't want me to respond to your post either.
And while saying 40 fps feels like 15 fps in other games might be an exaggeration, there are some issues that may cause the quality of the fps to drop when the client gets loaded - mostly because frame time may start to vary a lot when there is a lot of stuff happening, so those 40 fps does not mean every frame is 25ms - one or two may clock in at 75ms while most clock in at 20ms...
I'm fairly optimistic that things will improve quite a lot in the not-to-far-future though; the latest JIT testing gave pretty good results - well, apart from the crashing :-/
Pivotal tracker is not coming back, so recognizing this, pts like Scardybob and myself do our best to answer what we can and be highly accessible /informational. Sorry this isn't good enough. What answer do you not get from us that you think a dev would answer anyways?
P. S. I highly doubt people complain about performance due to a lack of a pivotal tracker.. ? I think they complain about it because they have poor performance.
So there, you got a reply anyhow. ;-)
Yeah, If only the UWE Devs would just say something! like maybe in a thread like this one, on this forum, in the form of a post... maybe even two posts above this one your reading right now!
Why does UWE have to be so secretive!
DisplayedFPS = (1+ UWEclaimedFPSgain) * ActualFPS
We're all sitting on 20 fps even though it says 80 fps ! Seriously though it really does feel like it which I why I have a fit when people say "yeah I get 30 - 50 fps which Is fine why are you complaining. Oh yeah and the human eye can't see more than 60 fps BTW"
That's exactly what I want to see with the recordings I suggested above.
Good to hear that. It would be great to get an estimate on how much of a performance boost we can expect!
I liked this discussion of the FCAT/framerating method as it gets into this type of consistency issue:
if you're bottlenecked by your CPU, increasing your graphical settings may increase your FPS because it's balancing the workload a bit more.
I think its best to wait for everything to be final and stable and go through thorough testing and benchmarking over many systems before estimation numbers are publicly given..
Else you will end up with people who had their expectations high and were upset about a failed delivery etc..
That being said, i'm pretty sure you can read between the lines when @matso says he's fairly optimistic and i keep posting about whats going to be fixed (variance in fps and frametimes)
:-B
*Edit: I got permission to post the graphs from Hugh!
***DISCLAIMER**
Again.. This is very much so a WIP and does not represent the final implementation, nor represents every scenario, and it may not even reflect your computer's results.
This graph is merely to demonstrate the lack of variance in fps and "smoothness" accomplished by current LUAJIT progress - actual fps increases depend on scenarios (entity heavy) and hardware - I have seen as little as 2 fps gain for some, and 40 in others .. so take this pic with a grain of salt. It was made by @matso late last night after we tried it out in PT.
Have you tried making a serious attack on "performance" issues that do not affect average frame rates?
Like input latency. Can you sample input more than once per frame to try and keep movement physics snappier and more precise even at lower framerates?
Like frame time variance. 60 FPS average with every other frame taking twice as long does not feel like 60 fps.
When you spin around quickly frame times increase; as it happens, spinning around quickly is something that often occurs in combat.
The framerate rate drop during combat is large, and enormous during heavy combat, and that's precisely the time you least want to drop frames. Can you strip back the most CPU demanding effects during combat, especially heavy combat?
Just being around lots of players seem to be a chief culprit, structures less so. Would some kind of LoD system allow you to get big savings on the animination system or whatever it might be that causes poor performance when lots of players are around?
Start plog capture = p_logall
End plog capture = p_endlog
You plog should output to the same directory as your log.txt file (default is your appdata folder). Then go to \SteamApps\common\Natural Selection 2\utils\PerfAnalyzer\ folder and run the plog file through the perfanalyzer.py program (will need to install Python on your computer to get this to work). That will get you the graphs Ironhorse posted earlier.
The plog function is basically just outputting all the data you see from the profiler to a file you can examine later. The only downside is that there is no way (that I know yet) to convert the data into something readable by a spreadsheet program.
I just recently upgraded my graphics card and from the experience from the two cards in this game, it depends heavily on the card in use. Just put fps counter on and fiddle with them, on my first card things like texture quality, anisotropic filtering and reflections, on this second one, even anti-aliasing and shadows on gives more fps than with them off, on the first card it was opposite with those two.
I don't know what I'm looking at but I want it now.
DO WANT