![]() |
Fleet
0.0.9
Inference in the LOT
|


Go to the source code of this file.
Functions | |
| int | mpi_rank () |
| int | mpi_size () |
| bool | is_mpi_head () |
| template<typename T > | |
| void | mpi_return (T &x) |
| Return my results via MPI. More... | |
| template<typename T > | |
| std::vector< T > | mpi_gather () |
| Reads all the MPI returns from mpi_return. NOTE that the output does not come with any order guarantees. This waits until all are finished. NOTE: This has NO robustness to losing nodes. https://mpitutorial.com/tutorials/dynamic-receiving-with-mpi-probe-and-mpi-status/. More... | |
| template<typename T > | |
| std::vector< T > | mpi_map (const size_t n) |
| Send an int n to each request. Each requestor can use this to index into some data which they are assumed to know already. More... | |
Variables | |
| const size_t | MPI_HEAD_RANK = 0 |
| bool is_mpi_head | ( | ) |
| std::vector<T> mpi_gather | ( | ) |
Reads all the MPI returns from mpi_return. NOTE that the output does not come with any order guarantees. This waits until all are finished. NOTE: This has NO robustness to losing nodes. https://mpitutorial.com/tutorials/dynamic-receiving-with-mpi-probe-and-mpi-status/.
| std::vector<T> mpi_map | ( | const size_t | n | ) |
Send an int n to each request. Each requestor can use this to index into some data which they are assumed to know already.
| int mpi_rank | ( | ) |
| void mpi_return | ( | T & | x | ) |
Return my results via MPI.
| x |
| int mpi_size | ( | ) |
| const size_t MPI_HEAD_RANK = 0 |
1.8.13