Accurate Game Timing / Server Settings

For all the help you need with Armagetron!
Post Reply
User avatar
TJohnW
Posts: 7
Joined: Wed Aug 08, 2012 7:16 am
Location: NY
Contact:

Accurate Game Timing / Server Settings

Post by TJohnW »

I have already read this page here,

http://wiki.armagetronad.org/index.php? ... tion_Guide

And understand how DEDICATED_FPS affects the server and its timing capabilities. However, there are also three more settings listed right under the DEDICATED_FPS section that are CYCLE_SYNC_FF, CYCLE_SYNC_FF_STEPS and CYCLE_SYNC_INTERVAL_ENEMY. They each come with a short description but not really any explanation as to what they would each do (or acceptable values). I found the defaults for CYCLE_SYNC_FF to be 10.0 and CYCLE_SYNC_FF_STEPS to be 1.

We are trying to achieve timings that will hopefully be accurate to more than the hundredths place. Right now we have a dedicated_fps of 100, what can we do to increase the accuracy of the game time? Particularly the game time reported on a death zone or zone enter event.

Any help would be appreciated!
Thanks,
Tristan
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Re: Accurate Game Timing / Server Settings

Post by Z-Man »

None of them have any effect on timing as measured by the server. The _FF_ settings are only active on the client and influence prediction quality (and, of course, CPU cost). The two INTERVAL settings determine how often clients are informed of the state of cycles while they just go straight. Unless there is unusual jumpiness of the cycles on the clients while they go straight, you should leave them all alone.

Wall collisions and cycle turns are always determined at an accuracy far greater than 1/DEDICATED_FPS. But Zone hit times are only as accurate as the simulation time granularity, so 1/DEDICATED_FPS. I would not recommend going far over DEDICATED_FPS 100, however. It increases CPU load and, depending on other settings, may reduce accuracy of positions.
User avatar
TJohnW
Posts: 7
Joined: Wed Aug 08, 2012 7:16 am
Location: NY
Contact:

Re: Accurate Game Timing / Server Settings

Post by TJohnW »

That is interesting to hear. I had figured the zones would only get so accurate. Before on our racing servers we discovered other things, such as when a Wall and a Zones edge were touching, a player could dig against the wall and “glitch” through the zones edge. Understanding that the zone events are calculated only on DEDICATED_FPS makes this far more understandable.

Image

This raises another question..

You say that wall collisions and turns are determined at a greater accuracy, does that mean I could get a greater precision regarding the time in which a player is killed from hitting a wall? (assuming there is 0 rubber). That would be an interesting idea, we could have maps that are purely target zones, and you win by dying on a wall!

Image

In this case, the blue wall I have highlighted.
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Re: Accurate Game Timing / Server Settings

Post by Z-Man »

TJohnW wrote:You say that wall collisions and turns are determined at a greater accuracy, does that mean I could get a greater precision regarding the time in which a player is killed from hitting a wall? (assuming there is 0 rubber). That would be an interesting idea, we could have maps that are purely target zones, and you win by dying on a wall!
Possibly. I thought about it, too, but it depends on how the death report is logged. At some point, definitely, a rather precise time is available, the question is just whether it is passed along correctly.
Post Reply