CUDA for Wave Simulation - Part 3: Efficient CUDA
Non-trivial CUDA The current snippet of code calling CUDA is: cuda_step<<< 1, 1 >>> This uses only one CUDA thread, and is probably extremely inefficient. To verify this, let’s add some way to time t
Non-trivial CUDA The current snippet of code calling CUDA is: cuda_step<<< 1, 1 >>> This uses only one CUDA thread, and is probably extremely inefficient. To verify this, let’s add some way to time t
There’s a good first CUDA tutorial here: https://developer.nvidia.com/blog/even-easier-introduction-cuda/. Then there are the docs here: https://docs.nvidia.com/cuda/cuda-c-programming-guide/#. The docs describe three levels of the thread hierarchy —...
