Engine Testing - 97.0.1-206-g8576ee8 (06. Aug 2014)
Moderator: Moderators
-
- Moderator
- Posts: 2464
- Joined: 12 Oct 2007, 09:24
Re: Engine Testing - 97.0.1-206-g8576ee8 (06. Aug 2014)
Are you sure sliding into terrain is still a problem? Impassable terrain now seems to act as a hard impassible barrier to units.
Re: Engine Testing - 97.0.1-206-g8576ee8 (06. Aug 2014)
I dunno about sliding now. Honestly I am trying to take a break from spring because if I spend more than a few hours I get all irritated about some of things that are spring specific stuff that I feel limits me. So I try to not spend too much time on the site right now, instead learning new skillsets that I didn't have(such as animation)smoth wrote:I also know in the past, when I gave doms more sliding they would slide into impassible terrain.
Re: Engine Testing - 97.0.1-206-g8576ee8 (06. Aug 2014)
I'm not sure sliding is/was a problem at all because my testing methods were imprecise. I just suspected it as a cause of CPU spikes in large-scale games because I noticed the function was being called repeated per frame under certain conditions. When I disabled the sliding code in the engine source (not using tags but actually commenting out the relevant parts of code and recompiling) or I destroyed particular units that seemed stuck between hills and units/features the spikes appeared to go away. I say "appeared to" because my testing was fairly manual and there's no guarantee I was triggering the same conditions on each run.Google_Frog wrote:Are you sure sliding into terrain is still a problem? Impassable terrain now seems to act as a hard impassible barrier to units.
Units probably don't slide onto "impassable" terrain but they can slide:
* Into a collision with stationery or moving unit/features
* Onto another slope
* Into an explosion or other impulse
I'm pretty sure when that when impulses chain like this all subsequent impulses or collisions are resolved in the same game frame. If you get a long or circular chain-reaction its probable the sim frame wont advance until it is resolved - resulting in lag. Under normal circumstances collision chain reactions should be limited by "friction" (things dont rebound/slide at 100% speed) but it's possible this doesn't happen quickly enough or under all situations or when too many units are involved. I haven't looked closely enough or recently enough to be more specific and I may also be wrong.
I didn't test any tags, my testing was mostly limited to regular BA raiders and peewees.smoth wrote:Does push resistant reduce or improve performance? or did you never investigate that tag?
- Silentwings
- Posts: 3720
- Joined: 25 Oct 2008, 00:23
Re: Engine Testing - 97.0.1-206-g8576ee8 (06. Aug 2014)
Without some indication of how long ago these test were, its quite difficult to know what to make of them. E.g. there was certainly a time, possibly quite long ago, when clusters of vibrating units caused a slowdown, but I think this problem doesn't exist anymore.
An updated version of http://springrts.com/phpbb/download/file.php?id=7567 would be pretty useful - since as I understand it thats well out of date. My self compiled spring is not suitable for doing it, else I'd make one myself.
An updated version of http://springrts.com/phpbb/download/file.php?id=7567 would be pretty useful - since as I understand it thats well out of date. My self compiled spring is not suitable for doing it, else I'd make one myself.
-
- Moderator
- Posts: 2464
- Joined: 12 Oct 2007, 09:24
Re: Engine Testing - 97.0.1-206-g8576ee8 (06. Aug 2014)
It sounds like a good idea to retest the performance of large clumps of units.
As for current engine testing I am waiting for bugs to be fixed in jK's changes to the pathfinder. http://springrts.com/mantis/view.php?id=4522
As for current engine testing I am waiting for bugs to be fixed in jK's changes to the pathfinder. http://springrts.com/mantis/view.php?id=4522
Re: Engine Testing - 97.0.1-206-g8576ee8 (06. Aug 2014)
i'm using the minimal portable windows version to test this and sometimes it hangs when trying to start a single player game. But after i close it i can try again on the same game and map and it works.
Anyone else got this?
(attached infologs)
Anyone else got this?
(attached infologs)
- Attachments
-
- infolog_hang_start_game_97.0.1-206.txt
- when it hung (i closed it through the "program not responding" popup we get on windows)
- (19.33 KiB) Downloaded 12 times
-
- infolog_working_start_game_97.0.1-206.txt
- when it worked
- (25.23 KiB) Downloaded 8 times
Re: Engine Testing - 97.0.1-206-g8576ee8 (06. Aug 2014)
oops, sorry for the endless-loop, this is the correct link:
http://springrts.com/phpbb/viewtopic.php?f=12&t=32492
http://springrts.com/phpbb/viewtopic.php?f=12&t=32492