bump the server tickrate to at least 60 please
TrueVeritas
Join Date: 2006-10-20 Member: 58082Members
Join Date: 2006-10-20 Member: 58082Members
alternatively let it be configurable. in all honestly i have no idea exactly what the ramifications are for this but there are already servers that are beefy enough to constantly run at 30 tickrate.
plus ns2:c exists and demands far less from servers, and would benefit greatly from 60 tick.
plus ns2:c exists and demands far less from servers, and would benefit greatly from 60 tick.
Comments
I think there are a million higher priorities right now than something like that.
that's fine. as long as we can get 6v6 in ns2:c i'm happy.
UWE already tried 60 tick, they said, actual quote from Q&A, "we have no idea if this will improve anything" and it "didn't", I think because most servers can barely handle 30 ticks above 8v8, and UWE servers can't all be 4.4ghz liquid cooled monsters (also I doubt they could tell he difference in hitregistry between updates, it used to be about 50% in beta with rifles now it's almost 80%, after doing various tests).
However, administrators and clients <b>need</b> to have more server control. Let us change our own interpolation settings, as well as give server admins the ability to change server commands such as the tickrate.
I think simply put twice the number of updates per second that a server does (which is why it drops when the servers struggle)..atleast thats my take on it...most likely totally wrong.
plus ns2:c exists and demands far less from servers, and would benefit greatly from 60 tick.<!--QuoteEnd--></div><!--QuoteEEnd-->
Would you be so kind to point me in the right direction of ns2:c? I can't seem to find it :(
edit: never mind found it
plus ns2:c exists and demands far less from servers, and would benefit greatly from 60 tick.<!--QuoteEnd--></div><!--QuoteEEnd-->
If you had any idea what you're talking about, you would know that player movement and actions are separated from the server tickrate in Spark. It doesn't work like Source/GoldSrc. Server tickrate only affects AI entities. So aside from Hydras and Whips running at 60 ticks, upping the tickrate wouldn't do much of anything.
Edit: <a href="http://www.twitch.tv/naturalselection2/b/333444319" target="_blank">http://www.twitch.tv/naturalselection2/b/333444319</a> @ 14:55
Edit: <a href="http://www.twitch.tv/naturalselection2/b/333444319" target="_blank">http://www.twitch.tv/naturalselection2/b/333444319</a> @ 14:55<!--QuoteEnd--></div><!--QuoteEEnd-->
I'm sorry but that information is slightly outdated. After the prediction multithreading and the resulting massive performance improvements it turned out that the servers couldn't handle clients sending updates each frame. So in order to avoid having the servers grind to a halt the client update rate was decoupled from fps, which means that the netcode now works very much like Source.
what's the point?
The practical difference is twice the number of server game updates a second.
The human perceivable difference is nothing.
Actually it still holds true. What changed is that the client sends updates at 30 ticks per second (but if i understand it right this can be higher in situations where its needed).
The server tickrate cannot be compared to source and increasing it to 60 would most probably not change anything.
Maybe the hit registry won't be ###### for once, and they can drop that awful high interpolation value that causes most of the network problems in the game right now. No more getting bitten from 10 feet away, no more dying instantly to one skulk bite, no more having to wait a whole second for a skulk to die after the server recognized you have killed him, think of the possibilities.
Hit registration is consistently great for me. I'm not sure what issues you're running into, but I don't think a tickrate above 30 will have the benefits you're imagining.
no, but setting interpolation to a more precise setting will.
The command is just "interp <seconds>"
so interp 0.05 for 50 ms interp.
The server enforces this value to all clients once you execute it on the server's console. I don't know why the interp seems to be reset somewhat randomly on mapchange, player joins, etc... but just use DAK if you want to always have a server on 60 ms interp etc...
NS2 has hit registration problems, even the developers have acknowledged it. They have said they have additional fixes to make after 1.0 to fix some additional problems that are a little more risky.
Increasing the client update rate is advantagous because you can then lower the interpolation, which means that your gamestate is closer to the servers, meaning less getting hit around corners, less distance between you and the skulk biting you.
If you dont understand that or its too advanced, thats not our fault, and nor is that something that takes an 'expert' to review.
NS2 has hit registration problems, even the developers have acknowledged it. They have said they have additional fixes to make after 1.0 to fix some additional problems that are a little more risky.
Increasing the client update rate is advantagous because you can then lower the interpolation, which means that your gamestate is closer to the servers, <u>meaning less getting hit around corners, less distance between you and the skulk biting you.</u>
If you dont understand that or its too advanced, thats not our fault, and nor is that something that takes an 'expert' to review.<!--QuoteEnd--></div><!--QuoteEEnd-->
What do either of those things have to do with hit registration?
NS2 has hit registration problems, even the developers have acknowledged it. They have said they have additional fixes to make after 1.0 to fix some additional problems that are a little more risky.
Increasing the client update rate is advantagous because you can then lower the interpolation, which means that your gamestate is closer to the servers, meaning less getting hit around corners, less distance between you and the skulk biting you.
If you dont understand that or its too advanced, thats not our fault, and nor is that something that takes an 'expert' to review.<!--QuoteEnd--></div><!--QuoteEEnd-->
this post is kind of ironic.
the developers have acknowledged hit reg problems because they would be talking BS if they denied it... every multiplayer FPS has hit reg problems, it's part of the game.
you must think the dev's are pretty dumb, that they had this 'magical' solution to lower interp and increase tickrate all along but decided to give the players the current settings just for the hell of it?
fair enough... but i don't think the hit registration is a priority issue.
as someone who's very familiar with hit registration, having played UT > UT2004 > TF2 competitively, and played thousands of hours on pub servers on those games as well as battlefield, COD and counterstrike 1.6, source and go, i'm saying that hit registration in NS2 (for me) is better than any of those other games except UT where it's probably about the same - so far i've had NO hit reg issues seen in 33 hours of NS2 play.
(granted the earlier games didn't even have interpolation which i think is probably the biggest cause of dodgy reg)
Iirc, increasing the tick rate also increases network traffic - an area which is already inundated for servers currently. I think the gains from changing this would have downsides unless it was just ns2c 6 v 6..
I.e. a pretty isolated scenario given the current usage of ns2.
I doubt the devs aren't aiming for increased server performance along with lowered interp values, however, so just sit tight
:)
Increasing the TICKRATE in NS2 will not have impacts like in source games, as player move updates are calculated instantly. Increasing the tickrate in NS2 would really only impact AI calculations.
Increasing the MOVERATE (aka client update rate) can/would have impacts, which is the recent change in NS2. Changing this however DOES have an impact on server performance, and is the reason they moved from an uncapped update rate to the current rate which is capped at 30.
Increasing the Server update rate would also be benifical, which is currently capped at 20. This change would allow you to lower the interpolation. This change would increase data sent to clients, increasing the choke issues which while greatly reduced, are still a problem in NS2 at times.
Saying that all games have hit detection issues so its fine to just ignore them is insanely ignorant, there are lots of multiplayer FPS games that have good hit detection. And Im not saying that they have tons of performance and bandwidth overhead and are just wanting us to suffer with lower update rates then desireable, I very well know the limits of NS2 currently. The devs are and have been working very hard on server performance and bandwidth usage, and I expect them to continue those improvements.
The other major part of this IMO is that they have made some pretty significant engine changes recently, and there are no ways for us to quantify and show hit detection issues easily to the developers, so they know better where to look for bugs.
One of the best examples of this currently is the rifle, which has a 3º spread currently, however its extremely biased towards then center, probably making it effectively 2º or less. In NS1, you had a smaller, faster skulk and the lmg had a 4º spread, and it was alot easier to kill skulks there... and I dont think I need to explain why.