! Configure the kernel launch. * * \param grid The thread grid dimensions for the launch. * \param block The thread block dimensions for the launch. * \param smem The amount of shared memory to dynamically allocate, in * bytes. * \param stream The CUDA stream to launch the kernel in. */
| 3001 | * \param stream The CUDA stream to launch the kernel in. |
| 3002 | */ |
| 3003 | inline KernelLauncher configure(dim3 grid, dim3 block, unsigned int smem = 0, |
| 3004 | cudaStream_t stream = 0) const { |
| 3005 | return KernelLauncher(*this, grid, block, smem, stream); |
| 3006 | } |
| 3007 | /*! Configure the kernel launch with a 1-dimensional block and grid chosen |
| 3008 | * automatically to maximise occupancy. |
| 3009 | * |
nothing calls this directly
no test coverage detected