erwincoumans / experiments Goto Github PK

Hi I'm testing your OpenCL code on Macos 10.8.4 betas(but I suspect should be equal on 10.8.3) with 79xx for MAC and I get some errors on quicksort test (OpenCL_radixsort_benchmark_Apple_xcode4_debug) and perhaps for that also seems for that I get crashes on either OpenCL_gpu_rigidbody_pipeline2_Apple_xcode4_debug or OpenCL_gpu_rigidbody_pipeline2_Apple_xcode4_debug some variables get out of range values and samples crash..
So have you tested your codes on Apple+AMD GPUs or is a new bug either in new Macos builds or 79xx cards on Apple..
Also some OCL kernel compilation fails in Quadro Fermi for Mac saying something about PTX lowering error.

can you test and fix this bugs?

Original Bullet 2.x issue is here: http://code.google.com/p/bullet/issues/detail?id=336

See some research in the OpenCL branch
http://code.google.com/p/bullet/source/browse/#svn%2Fbranches%2FOpenCL
in particular
http://code.google.com/p/bullet/source/browse/#svn%2Fbranches%2FOpenCL%2FExtras%2FOpenCL%2FBatchConstraintSolver%2FShared

Original Bullet 2.x issue is here http://code.google.com/p/bullet/issues/detail?id=336

Add some GPU accelerated narrowphase collision detection for convex objects

See http://code.google.com/p/bullet/issues/detail?id=32

For some ideas, see video and discussions here:
https://mollyrocket.com/forums/viewforum.php?f=13

Original Bullet 2.x issue is here: http://code.google.com/p/bullet/issues/detail?id=152

The iterative constraint solver in Bullet 2.x doesn't handle large mass ratios well,
in other words a rigid body with large mass resting on a rigid body with small mass is unstable.
Similarly, constraints can become unstable with large mass ratios (a large arm/constraint pivot can cause a large mass ratio too).

See also http://code.google.com/p/bullet/issues/detail?id=33

For NaturalMotion forward dynamics, powering ragdoll joints using strong
motors, it is useful to have a Featherstone implementation available.

• including motors
• handling contacts

See Vangelis Kokkevis comments/paper on Bullet forums:
http://www.bulletphysics.com/Bullet/phpBB3/viewtopic.php?f=4&t=148&hilit=vangelis

We received some Featherstone implementations that need cleanup.

Original Bullet 2.x issue is here: http://code.google.com/p/bullet/issues/detail?id=336

A 3D mixed broadphase was implemented in this OpenCL branch
http://code.google.com/p/bullet/source/browse/#svn%2Fbranches%2FOpenCL%2FExtras%2FOpenCL

The GpuBoxDemo will crash on Linux, but the GpuCompoundDemo won't.

Move the original Bullet 2.x Issue http://code.google.com/p/bullet/issues/detail?id=128

See also

http://bulletphysics.org/Bullet/phpBB3/viewtopic.php?f=4&t=6863
http://bulletphysics.org/Bullet/phpBB3/viewtopic.php?f=4&t=6833

and from Pierre Terdiman's coder corner
http://www.codercorner.com/blog/?p=507
http://www.codercorner.com/blog/?p=498

Original Bullet 2.x issue is here: http://code.google.com/p/bullet/issues/detail?id=329

Some implementations are

CDTestFramework
http://code.google.com/p/bullet/source/browse/#svn%2Ftrunk%2FExtras%2FCDTestFramework
and

Blender Game Engine (wiki)
http://wiki.blender.org/index.php/Doc:Manual/Game_Engine/Logic/Object_type/Occluder http://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.49/Game_Engine

Gamekit: http://gamekit.googlecode.com

It is also discussion here with some basic demo:
http://bulletphysics.org/Bullet/phpBB3/viewtopic.php?f=9&t=7575

I just wanted to say thanks for porting fontstash to modern opengl. It really helps newbies like me getting started. Much appreciated!

I noticed in some instances like when rendering the word "font" the 'f' clips a part of the 'o'. The same happens when one set of text overlays on top of another set of text. There is a wide margin around the foreground text that the background text does not show through.

I see the text is rendered with transparency enabled and this holds true for non text objects in the background. Any idea what could be causing this?

Original topic is here: http://code.google.com/p/bullet/issues/detail?id=336