Performance being a main feature of Blaze, it has to be for Blaze CUDA too. Therefore benchmarking will be necessary at some point to make sure we reach that goal.
Base files are here, some basic benchmarks have been added to make sure Blaze CUDA behaves well (no performance loss, computations being memory bound etc.). Everything's good so far.
Performance being a main feature of Blaze, it has to be for Blaze CUDA too. Therefore benchmarking will be necessary at some point to make sure we reach that goal.