Definitions and functionality related to CUDA events (not including the event wrapper type event_t itself)
More...
|
using | handle_t = CUipcEventHandle |
| The concrete value passed between processes, used to tell the CUDA Runtime API which event is desired. More...
|
|
Definitions and functionality related to CUDA events (not including the event wrapper type event_t itself)
◆ handle_t
The concrete value passed between processes, used to tell the CUDA Runtime API which event is desired.
The CUDA driver's raw handle for events passed between processes.
◆ export_()
Enable use of an event which this process created by other processes.
- Parameters
-
event | the event to share with other processes |
- Returns
- the handle to pass directly to other processes with which they may obtain a proper CUDA event
◆ import() [1/2]
Obtain a proper CUDA event, corresponding to an event created by another process, using a handle communicated via operating-system inter-process communications.
- Note
- IMHO, the CUDA runtime API should allow for obtaining the device from an event handle (or otherwise - have a handle provide both an event handle and a device ID), but that is not currently the case.
- Parameters
-
event_ipc_handle | the handle obtained via inter-process communications |
device | the device with which the imported event is associated |
◆ import() [2/2]
- Parameters
-
context | the device-context with which the imported event is associated |
event_ipc_handle | The handle created by another process, to be imported |
- Returns
- An event usable in the current process