Particle sync

AF · Post by AF » 16 Dec 2006, 20:52

Can we have unnecesaryily synced particles be unsynced?

For example nanospray, there's no justifiable reason for keeping it sync that does not rely on the player watching the map like a hawk while being extremely lucky to catch a single nanoparticle while being stupid enough to not already know there's something there anyway.

Lindir The Green · Post by **Lindir The Green** » 16 Dec 2006, 22:02

My understanding from the last time this was brought up is that the only thing that makes them synced is that they use the same random number generator.

AF · Post by AF » 16 Dec 2006, 22:23

They're greatly affected by network timeouts and network traffic, which shouldnt be the case for purely graphical effects.

Tobi · Post by **Tobi** » 17 Dec 2006, 01:16

Indeed there's no reason it's synced, but at the same time you will see no improvements at all if it's made unsynced, with the exception of nanoparticles possibly being affected by the maxparticles setting.

There will be no change in amount of network traffic, network lag or anything like that tho, most probably you wouldn't even notice it if it was made unsynced.

Peet · Post by **Peet** » 17 Dec 2006, 01:26

Except that people with crap computers (me) would be able to remove them from their modfiles, and possibly be able to play at >10 fps.

Tobi · Post by **Tobi** » 17 Dec 2006, 01:29

Thats what I said yes.

KDR_11k · Post by **KDR_11k** » 17 Dec 2006, 08:20

Tobi wrote:with the exception of nanoparticles possibly being affected by the maxparticles setting.

Considering the number of nanoparticle sources mods can include, the number of particles that exist per source (a few hundred I think), the lack of customizability (mods with more nanoparticle sources can't reduce the number of them, only turn them off completely and hope for the best) and the inability of players to reduce the number displayed on weaker computers that should be reason enough.

Argh · Post by **Argh** » 17 Dec 2006, 09:23

All I have to say about this is that going to the system I used for NanoBlobs, which involves a lot more code on my end (thus having to go through parser, etc.) and is much more complex... still runs a lot better than the default, at least in my experience. Just like getting rid of the default explosions, I was pleasantly surprised

Tobi · Post by **Tobi** » 17 Dec 2006, 09:31

@KDR_11k, I'm not arguing against it... I'm just pointing out that certain improvements people think they'll see from it won't be seen. Splitting up unsynced particles and synced projectiles is still on my todo list somewhere, and has been there for a long time. For entirely different reasons then those mentioned here tho.

Acidd_UK · Post by **Acidd_UK** » 17 Dec 2006, 12:44

Surely the more things can be removed from sync, the more things that can be eliminated from the possible causes of desyncs?

Post by **jcnossen** » 17 Dec 2006, 13:46

Surely the more things can be removed from sync, the more things that can be eliminated from the possible causes of desyncs?

Not really. For example, smoke particles even if synced (like with the explosion generator), dont affect anything but themself. Same for many other particle types.

AF · Post by AF » 17 Dec 2006, 16:41

with the exception of nanoparticles possibly being affected by the maxparticles setting.

I thought they already did!

This explains a lot. A lot of the time I see noticeable lag increases when lots of nanoparticles are onscreen and despite that I shouldnt ever see more than 300 particles onscreen according to my manual setting of the maxpaticles value.

Nanoparticles should be unsynced for this reason alone.

Tobi · Post by **Tobi** » 17 Dec 2006, 20:28

Of course. Now get someone to code it.

Spring RTS Engine

Particle sync

Particle sync