Collisions appear to work, but particles get into the arm. This has to do with the fact that the arm motion is too fast, and also, the arm rotates "over" the particles. Not clear how to deal with particles that are within the object by more than the boundary separation distance.
…y is not happening.
cloud_omega is the rotation vector with magnitude equal to angular velocity. cloud_xg is the fixed point about which the rotation occurs (not the center of gravity). Am still debugging cloud_velocity.cl OpenCL kernel.
classed-based and does not depends on Microsoft extensions.
…nes. I am about to remove _cloud_ from variable names.
…proved. Buffer with added size accessors.
…idParamsScaled. No more kernel errors (I use the same Bitonic Sort kernel for Cloud and SPH. Do now know why there was a kernel error. should not have happened.
Added methods to get nb elements and nb of bytes of device arrays.
All functions of a class must remain in a SINGLE FILE. If there are special circumstances, all files related to the same class MUST be in the same directory. (helps with tools like grep, etc.) Still have a Kernel error. Cleaning up code is the best way to track these errors.
…functions using cloud class.
…k. But am not yet calling cloud methods from Cloud class.
All within the rtps namespace. Will now clean up the code, and add some setters/getters to make sph, grid and other parameters.
…conditions. For now, the cloud velocity is constant and hardcoded in the collision routine, although I have created velocity arrays for the cloud points.
is not ok. Reason unknown. Added debug print statement printDevArray in SPH.cpp . Adding cloud point velocity vector to the boundary conditions.