| Sets callback that is called when event
| is finished
|
| Initializes event
|
| Waits till operation is fully finished,
| can be called concurrently from multiple
| threads
|
| Queries current status of operation,
| can be called concurrently from multiple
| threads
|
| Called on event to signal that CPU part
| of operation is finished,
|
| Optionally accepts error message from
| CPU part.
|
| Should be called no more than once per
| event
|
| Waits and returns as soon as possible
| in order schedule next operation, e.g.
| for CUDA->CUDA waits only for CPU part
| of CUDA op, for CUDA->CPU waits till
| the CUDA op is fully completed.
|
| Prepares context to synchronize device
| part of operation.
|
| Can be called concurrently from multiple
| threads
|