No, all kernels must be launched from the host. However, kernel launches are asynchronous, so you can queue them up to run in sequence without further host intervention.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Sharing a single counter (variable) across multiple thread(s) block(s) | 13 | 4377 | December 27, 2017 | |
| Writes in same memory location Cant add numbers from different threads? | 46 | 25719 | July 5, 2007 | |
| Newbie - Need to use shared mem? | 27 | 15017 | December 17, 2008 | |
| Texture Memory / Large Data / Global Memory Advice | 14 | 10754 | May 18, 2010 | |
| __syncthreads and shared memory | 21 | 4411 | June 15, 2011 | |
| [SOLVED] Code his own shared memory with device memory! | 15 | 2604 | October 7, 2015 | |
| Threads, branching and writing to global memory | 3 | 663 | October 24, 2018 | |
| Writing results into global array for only some threads | 5 | 1730 | April 6, 2009 | |
| Wishlist Place your considered suggestions here | 201 | 204644 | April 13, 2009 | |
| cuda shared memory usage + no reduction with threads | 5 | 1125 | April 23, 2012 |