Partial Sphere Fall Bug [Solved]

Report any bugs here and we'll post fixes

Moderators: Sascha Willems, Thomas

Partial Sphere Fall Bug [Solved]

Postby MeltingPlastic » Sun Aug 12, 2018 3:06 pm

Hi Julio - Found another small bug. It seems to happen in single threaded and multithreaded mode equally:

Some of the Spheres do not get "purturbed" and as a result do not awake from sleep:
Last edited by MeltingPlastic on Sun Aug 12, 2018 9:12 pm, edited 1 time in total.
MeltingPlastic
 
Posts: 237
Joined: Fri Feb 07, 2014 11:30 pm

Re: Partial Sphere Fall Bug

Postby Julio Jerez » Sun Aug 12, 2018 3:49 pm

Oh that was experiment that I committed by mistake,
It will be fixed later today.
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: Partial Sphere Fall Bug

Postby Julio Jerez » Sun Aug 12, 2018 5:42 pm

ok this is some side effect of some optimization I make, I though I new what it was but I was wrong,
I try to recreate it an it seem to work here.

could you please sync, test to see if it happen and if saw please upload the demo so tha it can be debuged?
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: Partial Sphere Fall Bug

Postby MeltingPlastic » Sun Aug 12, 2018 6:07 pm

Ok - Synced to newton commit: ea5ba8131324333ee3eea8ba7113a5d5e8f0a98c
The bug is still there.
https://drive.google.com/open?id=1g53r4-hyUBneZoyZVyxUIoHFPGz2n3Pw
MeltingPlastic
 
Posts: 237
Joined: Fri Feb 07, 2014 11:30 pm

Re: Partial Sphere Fall Bug

Postby Julio Jerez » Sun Aug 12, 2018 8:46 pm

ok I think I have fixed now, please check it out, check multy threaded also.


I noticed the demo freezes at random for long period of time, the freeze is not in the newton side, it seems to be is some garbage collision or some management.
But maybe you have than under control.
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: Partial Sphere Fall Bug

Postby MeltingPlastic » Sun Aug 12, 2018 9:02 pm

Works Great Now (multithreaded and single)! The bouncing of the top most spheres is much more natural looking across all the towers as well. With the center sphere bouncing the highest and the effect gradually declining as you look at shorter towers.

I do get freezes under seemingly heavy load (although I dont get any freezes in the demo I posted) I will look into it.

I have altered Urho3D to prioritize its logical updates over its rendering so if the logical update routine becomes too long the engine will never get a chance to render. I'm not sure how I want to handle that. Maybe I'll force renders every once in a while if there is too much load.

Edit: when I set to single threaded and debug mode and I get a freeze - I hit pause on the debugger and see that it is usually doing math inside a dVector or something which is what I was expecting.
MeltingPlastic
 
Posts: 237
Joined: Fri Feb 07, 2014 11:30 pm

Re: Partial Sphere Fall Bug

Postby Julio Jerez » Sun Aug 12, 2018 9:19 pm

MeltingPlastic wrote:Edit: when I set to single threaded and debug mode and I get a freeze - I hit pause on the debugger and see that it is usually doing math inside a dVector or something which is what I was expecting.


yes but that is not what is slowing down, if you set a break point to the line after the NewtonUpdate you will see that it returns many times but there is no rendering, I am not sure what but I leave that to you unless.
There are no periodic functions in newton that execute every so often, when is slow in one frame it will be slow for many frame.
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: Partial Sphere Fall Bug [Solved]

Postby MeltingPlastic » Sun Aug 12, 2018 9:21 pm

Yes - I think its a scheduling issue on my end. Thanks!
MeltingPlastic
 
Posts: 237
Joined: Fri Feb 07, 2014 11:30 pm


Return to Bugs and Fixes

Who is online

Users browsing this forum: No registered users and 5 guests

cron