So what is the cause of ridiculously high ping in this game?

  • Please make sure you are familiar with the forum rules. You can find them here: https://forums.tripwireinteractive.com/index.php?threads/forum-rules.2334636/

mattlach

FNG / Fresh Meat
Oct 20, 2011
415
134
0
Massachusetts
I have tested our servers up and down slot numbers to find what ours can handle and have settled on 38

Having this number, pings dont rise, i have a 50 ping, there is no lag.

In my opinion a lot of this has to do with people trying to have to many slots on their servers, if we go to 40 slots on our servers pings start to rise.

Three questions:

1.) Do you find this is a server hardware limitation (CPU/RAM?) or a network bandwidth limitation?

2.) Does anything else run on this server?

3.) What is your server hardware / bandwidth like?

Thanks,
Matt
 

SnowyOak

FNG / Fresh Meat
Jun 17, 2011
444
145
0
Chicago, USA
It has been suggested to me that it has something to do with game servers being insufficient to run the game.

Is that really it? If this is the case, I have yet to find a server powerful enough to run the game.

From other games I play I have come to expect the in-game latency to be pretty much the same as the game browser, or pinging a server from console.

I have come to expect - with my connection - to get between 18 and 25ms on servers in my general geographic area.

I have yet to play an RO2 server with below 80ms in-game latency (this would be considered barely playable in other games). The norm is more like 120 - 175ms, and occasionally its up in the 250-300 range.

What gives?

I feel like I', constantly shooting enemies with shots that CAN NOT MISS from relatively short distances over iron sights, and nothing...

Is there nothing that can be done to improve this? This is one of the biggest problems with the game right now IMHO.

Your ping in RO2 is calculated differently then the ping in a different online game. In the server browser you may see a ping of 35 for a server very close too you; that would be a ping comparable to other games. However, when you enter the game you'll see that ping rise as it now includes the time for the server to calculate your actions while playing.

So, for example, if a server has a 35ms ping in the browser (The actual time it takes for the signal to transport too you and back) and the server takes another 50ms to calculate all your actions, they would add up too an 85ms ping in-game. I know some other unreal engine games calculate ping in a similar fashion, Section 8:prejudice & such, so it may be unique to the engine.

The devs tell us high ping servers are a franchise of 3rd party dedicated servers, which makes perfect sense too me. TWI can afford to get the best quality hardware to host their servers (Those intel Xeon things with 4ghz turbo) while an average player won't have that available. When a clan or individual tries to host their own server, they often over estimate how powerful the server hardware is and give it a player slot limit that's simply too high. Heck, even I tried hosting a 16 slot server on a pentium 4 with very poor results. The server side for RO2 is simply very processor intensive. When the server fills up, the hardware starts to struggle with all the processing requirement and pings raise an exponential amount. This is why I avoid 64 player servers that aren't officially hosted or ones that I personally trust.

I hope that helps clear some things up.
 
Last edited:

mattlach

FNG / Fresh Meat
Oct 20, 2011
415
134
0
Massachusetts
Your ping in RO2 is calculated differently then the ping in a different online game. In the server browser you may see a ping of 35 for a server very close too you; that would be a ping comparable to other games. However, when you enter the game you'll see that ping rise as it now includes the time for the server to calculate your actions while playing.

So, for example, if a server has a 35ms ping in the browser (The actual time it takes for the signal to transport too you and back) and the server takes another 50ms to calculate all your actions, they would add up too an 85ms ping in-game. I know some other unreal engine games calculate ping in a similar fashion, Section 8:prejudice & such, so it may be unique to the engine.

This appears to be a symptom of the dedicated server code being unpotimized or it being run on inferior hardware.

The devs tell us high ping servers are a franchise of 3rd party dedicated servers, which makes perfect sense too me. TWI can afford to get the best quality hardware to host their servers (Those intel Xeon things with 4ghz turbo) while an average player won't have that available. When a clan or individual tries to host their own server, they often over estimate how powerful the server hardware is and give it a player slot limit that's simply too high.

I'm sure there is some of this. It's surprising the requirements for this server are so high though. I've been hosting various dedicated game servers on my own linux boxes since ~2001, and never have I come across a game that needed a more powerful server than I could build together using slightly used 1-2 year old spare parts.

Heck, even I tried hosting a 16 slot server on a pentium 4 with very poor results. The server side for RO2 is simply very processor intensive. When the server fills up, the hardware starts to struggle with all the processing requirement and pings raise an exponential amount.

This does not surprise me. Pentium 4's (Except for the very top end models) were pretty weak CPU's and one of Intel's biggest blunders in the history of the company.

This is why I avoid 64 player servers that aren't officially hosted or ones that I personally trust.

It has been suggested that trying 32 play servers may be a better idea. I will try this, but it removes some of the greatness of the game to me.


I guess I wonder what the requirements truly are for the dedicated server.

Is it properly multithreaded? Is it linux/windows/both?

Are the high pings based on people running really old hardware, or is the server so hardware dependent that even fairly new hardware struggles with it?
 

Floyd

Grizzled Veteran
Feb 19, 2006
4,313
725
113
Waterproof
www.ro50pc.net
Your ping in RO2 is calculated differently then the ping in a different online game. In the server browser you may see a ping of 35 for a server very close too you; that would be a ping comparable to other games. However, when you enter the game you'll see that ping rise as it now includes the time for the server to calculate your actions while playing.

So, for example, if a server has a 35ms ping in the browser (The actual time it takes for the signal to transport too you and back) and the server takes another 50ms to calculate all your actions, they would add up too an 85ms ping in-game. I know some other unreal engine games calculate ping in a similar fashion, Section 8:prejudice & such, so it may be unique to the engine.
I don't know how many times this has been explained....yet it continues to fall upon deaf ears....


The devs tell us high ping servers are a franchise of 3rd party dedicated servers, which makes perfect sense too me. TWI can afford to get the best quality hardware to host their servers (Those intel Xeon things with 4ghz turbo) while an average player won't have that available. When a clan or individual tries to host their own server, they often over estimate how powerful the server hardware is and give it a player slot limit that's simply too high. Heck, even I tried hosting a 16 slot server on a pentium 4 with very poor results. The server side for RO2 is simply very processor intensive. When the server fills up, the hardware starts to struggle with all the processing requirement and pings raise an exponential amount. This is why I avoid 64 player servers that aren't officially hosted or ones that I personally trust.

I hope that helps clear some things up.
Even with the state of the art hardware there a issues on certain maps when the server is loaded, arty is falling, tanks are battling, etc. But generally on those servers that are properly set up, while the pings rise, the gameplay is generally playable most of the map. On lessor servers, not so much. And then too, some of not-so-high-end player computers simply have problems with larger player numbers, yet they continue to play on the large capacity servers.

@krobar: I read: "Chances are..." and not a definitive "Becuase its your..."

@mattlach: the engine doesn't multi-thread well and that was a large part of the problem with servers early in the game's release. Though I'm pretty sure that most GSP's of any caliber know that by now, I wouldn't doubt that there are some that 'know' more about the game than what John has explained. And there are no Linux RO2 servers at this particular time.
 
Last edited:

Murphy

FNG / Fresh Meat
Nov 22, 2005
7,067
743
0
35
liandri.darkbb.com
I have yet to play an RO2 server with below 80ms in-game latency (this would be considered barely playable in other games).
Doesn't matter, but in my opinion 80ms is fine for pretty much every game except Quake 3 insta-gib on a fairly high level, lol. Even Counterstrike is perfectly playable with 80ms. Anyway...

The norm is more like 120 - 175ms, and occasionally its up in the 250-300 range.
I have higher latency in RO (1 and 2, btw.) than in most other games as well. No idea why, but it's always been like that so I don't really expect it to change.:(

I feel like I', constantly shooting enemies with shots that CAN NOT MISS from relatively short distances over iron sights, and nothing...
That's also due to bad hit registration over short distances. Last I heard TWI were aware of this and working on it.
 

SnowyOak

FNG / Fresh Meat
Jun 17, 2011
444
145
0
Chicago, USA
I wouldn't call the RO2/UE3 netcode unoptimized or inferior, it just calculates ping differently. Most other games would only show you the communication time between your computer and the server, which is what most people think of when they think "Ping." RO2 actually gives you more information by telling you the server processing delay as well. If it only showed the internet delay and not the hardware, would there still be a problem or would people not notice the hardware delay on overloaded servers anymore?

The server runs on 1 large thread with a few other very small threads. This is why you could get away with a single core per server, as long as that core is very powerful.

Some better info from the beta DDE thread by Mr. President himself:
Server Requirements (Beta)

16 Players: Haven't benchmarked this yet, but we don't expect it will take much CPU to run this

32 Players: CPU usage - 1 Core of a 2.6 GHZ Core I7 or equivalent (i.e. the server process will take most of 1 core, with a few smaller threads on other cores). We also tested on an Intel Core2 Quad Q8400 @ 2.66GHz and that could handle 32 players but I wouldn't try and do 32 players with a machine much slower than that.

64 Players: Goal CPU usage - 1 Core of a Intel Xeon E3-1270 3.4 GHZ (3.8 GHZ actually with Turbo enabled) or equivalent (i.e. the server process will take most of 1 core, with a few smaller threads on other cores).

A note on performance:
64 players was pushing the test machine (Intel Xeon E3-1270 3.4 GHZ (3.8 GHZ actually with Turbo enabled) pretty hard on certain maps, but did handle 64 players. Likewise 32 players pushed the Intel Core2 Quad Q8400 @ 2.66GHz pretty hard at 32 players, but did handle it. We're working on some final optimizations that might get the performance a little better, but there won't be any drastic changes from these specs.
http://forums.tripwireinteractive.com/showthread.php?t=57434http://forums.tripwireinteractive.com/showthread.php?t=57434

Also remember that while other games are very easy to host, not many include the positional effects and complicated ballistics system that RO2 has. I can host 2 killing floor servers easily off that same pentium 4 computer, which gives it 12 players connected at maximum, while just 4 more for RO2 is impossible. Even a 12 slot server would be impossible on the same computer. The difference is in Kf the bullets hit instantly, there is no calculation of the ballistic drop or travel time through arcs. I imagine spacial dependent features like this in RO2 are what make the server processing so intensive. If you think about it, the game actually preforms quite smoothly for what it does.

To my knowledge, the minimum CPU hardware required for a good quality 64 slot server is
Intel Xeon E3-1270 3.4 GHZ (3.8 GHZ actually with Turbo enabled).
If you have a server with that cpu, or one of equivalent or greater power, it should be able to run a 64 slot server pretty seamlessly. I trust TWI to do this properly, though kentucky fried goulash seems to do exceptionally well also.
 
Last edited:

the_Monk

FNG / Fresh Meat
Aug 4, 2011
286
145
0
Can you provide the ini settings on client side you mentioned. Further more, will changing these values have a negative impact when playing on servers that did not double their tickrate?


Here is the original thread posted by a member of the SWEET clan.

[url]http://forums.tripwireinteractive.com/showthread.php?t=68693&highlight=tickrate[/URL]


and no....there is no detriment to having those settings changed as they apply to "maximums" and if the current connection or server can't handle said settings it will simply fallback to lower settings.
 

the_Monk

FNG / Fresh Meat
Aug 4, 2011
286
145
0
It has been suggested that trying 32 play servers may be a better idea. I will try this, but it removes some of the greatness of the game to me.


Trust me. I thought so too, but in servers with less overall players (ie 24-32) you get to hear more of the environmental audio (which is fantastic for immersion) and the intensity is ramped way up. Playing "whack-a-mole" in a 64-person server actually seems less "great" by comparision. Try it, you'll see what I mean. :D
 
  • Like
Reactions: PUTZ and SnowyOak

vohk

FNG / Fresh Meat
Sep 15, 2011
17
11
0
The devs tell us high ping servers are a franchise of 3rd party dedicated servers, which makes perfect sense too me.

While I appreciate the detailed answer, in my personal experience this is completely wrong. On average I find the TWI servers running 64 players are just as bad as everyone else. Same high server browser pings, same lag spikes. Other players in game report the same thing. The issue is not limited to poorly configured hardware, unless TWI's qualifies.
 

SnowyOak

FNG / Fresh Meat
Jun 17, 2011
444
145
0
Chicago, USA
While I appreciate the detailed answer, in my personal experience this is completely wrong. On average I find the TWI servers running 64 players are just as bad as everyone else. Same high server browser pings, same lag spikes. Other players in game report the same thing. The issue is not limited to poorly configured hardware, unless TWI's qualifies.

I've not experienced that with the 64 slot TWI servers. When full, or very nearly full, my ping rarely goes past 120 on the official servers & on kentucky fried goulash I usually manage to stay around the 100 range. On virtually any clan hosted 64 player server my ping will surpass 200.
 

mattlach

FNG / Fresh Meat
Oct 20, 2011
415
134
0
Massachusetts
Server Requirements (Beta)

16 Players: Haven't benchmarked this yet, but we don't expect it will take much CPU to run this

32 Players: CPU usage - 1 Core of a 2.6 GHZ Core I7 or equivalent (i.e. the server process will take most of 1 core, with a few smaller threads on other cores). We also tested on an Intel Core2 Quad Q8400 @ 2.66GHz and that could handle 32 players but I wouldn't try and do 32 players with a machine much slower than that.

64 Players: Goal CPU usage - 1 Core of a Intel Xeon E3-1270 3.4 GHZ (3.8 GHZ actually with Turbo enabled) or equivalent (i.e. the server process will take most of 1 core, with a few smaller threads on other cores).

A note on performance:
64 players was pushing the test machine (Intel Xeon E3-1270 3.4 GHZ (3.8 GHZ actually with Turbo enabled) pretty hard on certain maps, but did handle 64 players. Likewise 32 players pushed the Intel Core2 Quad Q8400 @ 2.66GHz pretty hard at 32 players, but did handle it. We're working on some final optimizations that might get the performance a little better, but there won't be any drastic changes from these specs.

How I read this.

Even the Xeon E-1270 (fastest server processor money can buy for single threaded applications) is marginal for a 64 player server.

This seems to suggest that what they are trying to do can not be effectively done with current server hardware without a complete rewrite of the server engine for multithreading.

It would - however - be interesting to see how it would run on one of those watercooled overclocked core i5-2500K's at 5Ghz. It will have fewer threads and less cache than the Xeon, but it should beat it in raw single threaded performance.

Seeing that TWI did not write their own engine, but rather based the game on an existing Unreal 3 engine, the chances of seeing a multithreaded server and thus good playable 64 player servers are close to nil.

This is sad.
 

mattlach

FNG / Fresh Meat
Oct 20, 2011
415
134
0
Massachusetts
How I read this.

Even the Xeon E-1270 (fastest server processor money can buy for single threaded applications) is marginal for a 64 player server.

This seems to suggest that what they are trying to do can not be effectively done with current server hardware without a complete rewrite of the server engine for multithreading.

It would - however - be interesting to see how it would run on one of those watercooled overclocked core i5-2500K's at 5Ghz. It will have fewer threads and less cache than the Xeon, but it should beat it in raw single threaded performance.

Seeing that TWI did not write their own engine, but rather based the game on an existing Unreal 3 engine, the chances of seeing a multithreaded server and thus good playable 64 player servers are close to nil.

This is sad.


Actually, this article is interesting.

Makes you wonder why the multithreading tools were not used in the development of RO2, especially since it was known it would be CPU intensive.

I wonder what it would take to implement them retroactively. A relatively simple recompile, or a complete rewrite...
 

Faneca

FNG / Fresh Meat
Sep 16, 2010
1,150
778
0
Portugal
That should not be a concern to anyone who plays this game. As long as someone on your team gets the kill, it furthers the team goal. This game is not about individual performance. K/D ratio, number of kills, etc. don't matter. It's all about the team meeting its objectives. :rolleyes:

I actually ment that I could clearly see the bullets from my SVT-40 hit Germans in the left side of their chest and they keep running without any hit being detected by the game.
God damn fascists must have super powers.
 

LordKhaine

FNG / Fresh Meat
Dec 19, 2005
1,008
120
0
UK
I wonder what it would take to implement them retroactively. A relatively simple recompile, or a complete rewrite...

I'm assuming you don't have any experience in programming, but decently multi-threaded code isn't easy to do. In fact it's a pain in the arse. There is a reason we've had multi-core machines commonly in machines for many years without a matching increase in software utilisation.

As for the high ping thing. As mentioned, it's overloaded servers. A depressingly high number of servers are running too many slots, often a great deal too many. It's best to ignore the ping in the server browser, and check F1 the moment you join. If everyone has a 150ish ping... don't play. Change server immediately.

Very few of the 50-64 player servers seem able to actually cope with it. Similar situation as in RO1 when the player count was increased to 64 in that. And I imagine the problem will ease a little as more efficient and faster cpu's come out.
 

mattlach

FNG / Fresh Meat
Oct 20, 2011
415
134
0
Massachusetts
I'm assuming you don't have any experience in programming, but decently multi-threaded code isn't easy to do. In fact it's a pain in the arse. There is a reason we've had multi-core machines commonly in machines for many years without a matching increase in software utilisation.

I'm not a programmer, but I have some laymans knowledge and I am aware of the difficulties of writing good multithreaded code.

What I am not familiar with is how much actual coding developers who use existing engines actually do, or if most of itisscripting, model making and texturing and other artistic work.

If the unreal engine already included provisions for multithreading, were they used? Are those Unreal provisions any good? How much do they rely on the game developer to implement them, and how much of it is inherent to the engine. If so what percentage of the code in the game is unreal code vs RO2 specific code.

Can you flip a switch in the engine to enable those already existing unreal provisions (of they aren't already used) without significantly modifying the portion of the code that is RO2 specific, or do you have to have these Unreal provided multithreading provisions in mind from day one when starting a new game?

Do you see where I am going with this?
 
Last edited:

Hjanne

FNG / Fresh Meat
Dec 12, 2010
43
15
0
Here's what we found:

1. the variation in the amount of "leading" having to be done for shooting targets became much less and overall became more consistent.

2. we could swear we were able to see more bullet strikes/impacts.

3. for the lack of a better way of putting things........it seemed like bullets now "flew faster"?

All of those things made for a much more enjoyable gaming experience.

We've even had (although we are located in central Canada) people from the EU playing in our server who noticed a difference.


While we believe (as do some other clans who've also made those changes) that this has benefitted our server all of this may in fact mean nothing.......

Finally some recognition. I still cant belive every server is not running these settings by now. And as far as the results. Yes that was pretty much what we found. also smaller pingtimes are possible due to the new server updaterate.
All in all it makes the game run alot better "netcode" wise. You shoot and he dies. Not you shoot and 0.3-1 sec later he dies like it used to be.
 
Last edited:

Hjanne

FNG / Fresh Meat
Dec 12, 2010
43
15
0
Can you provide the ini settings on client side you mentioned. Further more, will changing these values have a negative impact when playing on servers that did not double their tickrate?

I made a post about how to do it a while back:
http://forums.tripwireinteractive.com/showthread.php?p=928701#post928701

The settings are there and what it does and what to look out for. GLHF

Also changeing CLIENT ConfiguredInternetSpeed to 20000. Its fine when you enter a server with a lower maxsetting then what you have since your rates will be as high as the server allows so they get lowered to whatever the server allows. Like in source engine and so on
 
Last edited: