search

code-google-com / bullet

Automatically exported from code.google.com/p/bullet
0 stars 0 forks source link

Executing narrow phase in parallel with broad phase #327

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
It would be useful to be able to overlap processing of the narrow phase
with that of the broad phase, so that when each broad phase collision is
detected, the narrow phase processing for that collision begins
immediately.  This overlapping could be implemented using threads with a
custom collision/dynamics world and dispatcher.

It is of interest to me as I've created a hardware broad phase that returns
the collisions in "bunches".  This hardware replaces the Bullet broad
phase.  It would be useful to begin narrow phase processing of each bunch
once they arrive in the software.

This topic was discussed on the forum at
http://bulletphysics.org/Bullet/phpBB3/viewtopic.php?f=9&t=4571

Original issue reported on code.google.com by muiris.w...@gmail.com on 12 Jan 2010 at 10:28

GoogleCodeExporter commented 9 years ago 

We are focussing on parallelism (and OpenCL GPU support) for Bullet 3.x, so we 
will 
address and discuss this further in this context.

Original comment by erwin.coumans on 20 Jan 2010 at 4:15

GoogleCodeExporter commented 9 years ago 
Closing old issues. Bullet 3.x is under development here:
https://github.com/erwincoumans/bullet3

It still doesn't have the 'processing of narrowphase in parallel with 
broadphase' though.

Original comment by erwin.coumans on 10 Sep 2013 at 9:32