Closed omlins closed 3 years ago
@luraess recently observed that for small problems the computation of kernel launch parameters (when using @parallel <kernel-to-launch>) was not performance-negligable. This PR fixes this performance issue.
@parallel <kernel-to-launch>
@luraess recently observed that for small problems the computation of kernel launch parameters (when using
@parallel <kernel-to-launch>
) was not performance-negligable. This PR fixes this performance issue.