OpenArena Message Boards

Dear All, this is my first post. Thanks for having here a great place to share and ask.

My question is very simple - I want to acquire the following three network-related performance measures:

1) Ping time
2) FPS
3) The rate of packet loss (dropped)

I know there are some commands available like \cg_drawfps 1, \cg_drawPing 1 but unfortunately it seems not work to me.
Can anybody show me all correct commands to obtain three measures? And any alternative way to obtain them?

Thank you.

Thank you so much for the reply.

For FPS, I saw it in some maps (or pk3 files?) but not in others.
For packet/snapshot drops and Ping, is there any logging way (e.g., pull them down to a file) so that I can do some statistics analysis (such as averaging) ?

Thank you!
nrz

The FPS counter should work independent of the map being played. Maybe it's an issue with your mobile version?

You could log your ping with /logfile 2; cl_shownet 3. /cl_shownet 3 will print lines like
Note that it'll really spam your console. /logfile 2 writes a log to ~/.openarena/baseoa/qconsole.log, not sure where/how you'd find that on your mobile.
As for snapshot drops, I don't know. Maybe you can extrapolate that from the shownet logs too, but I haven't looked at them in depth.

Edit: If I didn't get lost in C, the "snapshot" number should be increasing seamlessly, otherwise a snapshot was dropped.
I'm not aware of a way to print the FPS to console.

Thank you!

Lets put aside FPS first since it is not a big deal. (I can see it sometimes so it is already good enough).

For Ping, is it one-way or round-trip? My understanding of ICMP Ping is round-trip measure, but how about the way Quake 3 engine follows?

For dropped snapshot, I looked at net-related code and it did drop out-of-sequnced packets, hence for snapshots I agree with you - out-of-order snapshots will be dropped.

Moreover, I looked at UI code - they do have those statistics and plot them in the Lagometer window. But I simply do not know how to output them into a file ;)

AFAIK the ping on the stock scoreboard (Unlagged code brings g_truePing) depends on how fast clients acknowledge messages (not sure which ones) from the server (time ack minus time sent), so it's round-trip.
The latency on the lagometer is calculated differently, but not drawn as a concrete number anyways.

I don't think the regular game/engine code allows to log those numbers. You'd have to modify cgame and log them yourself, or modify the engine if you need to remain compatible with some pure server.

True for "You'd have to modify cgame and log them yourself", and that's what I tried earlier... Thanks for your replies, most are consistent with my understanding and some corrects me. I appreciate it :)

Seems there is not any bandwidth-related measure... For example how much bandwidth (probably in BytePerSecond) is required for a normal client user to play a normal game session? Let say this number is B, and the max bandwidth capacity of the server is MC, then I would assume MC/B = the max number of clients the server can serve. Am I correct? Do you have any idea or number for B and MC?

Thanks,
nrz

Some infos should be here:
(DO NOT LINK) h t t p s : / / openarena . wikia . com/wiki/Servers#Bandwidth

And maybe something also here:
(DO NOT LINK) h t t p s : / / openarena . wikia . com/wiki/Tweak

Thank you :)

Do you think Quake 3 is bandwidth-hungry game (with full features enabled such as VoIP)?

Do you know other bandwidth-hungry and latency-sensitive games?

Thanks for your time.

Unless you have sv_lanforcerate 1 (DO NOT LINK[/b]) h t t p s : / / openarena . wikia . com/wiki/Manual/Multiplayer#Optimize_for_LAN]Optimize for LAN (http://([b)) AND you are connecting in LAN, bandwidth usage is capped at 25000 bytes per seconds, IIRC (or lower, if "sv_maxrate" and "rate" values are lower than 25000... and the game can be enjoyable also with lower values). I don't know if Pelya's Android port changed something in this behavior (e.g. integrating ioquake3's new UDP autodownload code, which controls its speed with an apposite variable, distinct from sv_maxrate -maybe sv_dlrate?- which allows faster downloads than classic UDP download, although not as fast at HTTP redirection).

I can guess modern games may require much more bandwidth than OA.

Thanks. What do you mean by Modern game, could you share an example? I am stressfully testing how large bandwidth a local wifi LAN, or Bluetooth LAN or Wifi-Direct LAN setting for games can afford.

I sort of do not think games themselves consume much bandwidth. Really want to know typical values for different types of games nowadays. 

Thanks for your information.

I'm sorry, I'm not a real gamer anymore. The most modern games I played were some PlayStation 2 games such as God of War 2 and Burnout Revenge... and on PC, Quake Live (very little: I prefer OA)... (not counting some "casual games" on smartphone, such as Angy Birds series).

I'm quite outdated in games knowledge. Sorry!

I hope someone else here around may be more helpful!

PS: To answer to your question, for "modern game" I mean something published in the last few (around 3?) years.
Q3 has been released in 1999, when 56k were common (ISDN alredy was better than 56k, thanks to its lower latency)... I cannot call it a "modern" game (although I still love it). Today ADSL and fiber lines are much more common, so I can guess today's games may be optimized for such faster lines. OA bandwidth usage can be higher than Q3 one if one enables VoIP feature or "accurate physics", but still working great with a maximum of 25KB/s per user. I have no idea about the bandwidth used by a 2014 multiplayer game.

It matters a lot what kind of network model is being used. Obviously, a network where the server is asking constantly about events and actions of the players will take more bandwidth than when clients send their events and actions to the server and takes that as truth.

Beware, not every modern game uses the same network model.