Multi-device OpenCL kernel load balancer and pipeliner API for C#. Uses shared-distributed memory model to keep GPUs updated fast while using same kernel on all devices(for simplicity).
GNU General Public License v3.0
93
stars
10
forks
source link
Complete device to device pipeline stage initialization kernel execution #31
so it doesn't start with garbage, especially for hidden buffers per stage (hiddens also needs testing)