3 registered members (NewbieZorro, TipmyPip, 1 invisible),
19,045
guests, and 8
spiders. |
Key:
Admin,
Global Mod,
Mod
|
|
|
Re: benchmark test project included
[Re: FBL]
#180936
01/31/08 18:15
01/31/08 18:15
|
Anonymous
Unregistered
|
Anonymous
Unregistered
|
I've tested it on a powerfull pc with no stuttering and on my laptop (1.8Ghz/onboard Gfx) with stuttering. The only difference (on the F11-Statistics) was a higher srv and 0.2 for rev on the laptop (rev=0 at powerfull pc).
On both systems no other forground applications where running or active.
|
|
|
Re: benchmark test project included
[Re: ]
#180937
01/31/08 18:36
01/31/08 18:36
|
Joined: Aug 2005
Posts: 1,012 germany, dresden
ulf
OP
Serious User
|
OP
Serious User
Joined: Aug 2005
Posts: 1,012
germany, dresden
|
so far i have 7 reports of stuttering movement and only 3 where it looked like a good movement...
i think i can narrow it down to happen with objects(no matter if model, sprite...) in window mode, that are moving at a constant speed and where a static camera is looking at. it doesnt seem to have to do with time_step at all, because if i move an object with
my.x += 5;
for example in the above code, i get the same stuttering result. so please test it too and post the results, maybe you find a way to get rid of this strange effect? for me this concludes that right now i'm unable to make any game that includes objects moving at constant speed in window mode. but providing window mode to players is essential.
if more could test here, i could make a new thread and ask conitec about this. thanks
|
|
|
Re: benchmark test project included
[Re: ulf]
#180938
02/01/08 09:57
02/01/08 09:57
|
Anonymous
Unregistered
|
Anonymous
Unregistered
|
I've already tested several versions (like you just with my.x+=3, fix c_move with time_step, fix c_move without time_step, vec_acc...)
The result is always the same - so it's not time_step related.
|
|
|
Re: benchmark test project included
[Re: ]
#180939
02/01/08 10:10
02/01/08 10:10
|
Joined: Aug 2005
Posts: 1,012 germany, dresden
ulf
OP
Serious User
|
OP
Serious User
Joined: Aug 2005
Posts: 1,012
germany, dresden
|
yep, thats what i found out too, it must have to do with the engine running in window mode, with objects moving at higher speed when a static camera is looking at them. i think it's not been noticed yet, because a lot of people develop in fullscreen or use a moving camera, not so high speed and so on... maybe its a rendering issue? but i doubt that because the behaviour seems to change when you mess with the fps_ or time_smooth variables, but you cant get rid of it.
i just created a bug report regarding this over in the bug reports forum. maybe this will reveal the cause of this strange behaviour.
|
|
|
Re: benchmark test project included
[Re: ulf]
#180940
02/01/08 11:09
02/01/08 11:09
|
Anonymous
Unregistered
|
Anonymous
Unregistered
|
I have also scuttering when running your app in fullscreen mode (50fps) on my laptop, but not that strong as in windowed mode (50fps) . (today lower scuttering than yesterday!) 
|
|
|
|