You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Firstly, thank you very much for this library, I'm new to both Rust and OpenCL and find this library with its tests and examples as a great resource.
I'm trying to use set_event_callback from opencl3/cl3 but cannot find any documentation or examples on how this can be done. Can you please point me to a doc/test if one is available for this?
My goal is for my code to be able to run kernel executions repeatedly while utilizing the GPUs to the maximum possible. So my plan was to avoid blocking actions in the execution path and use event callbacks in order to determine when the kernel execution has been completed. Once it completes, I can flag it and allow my code to start a new execution again. As I mentioned I'm new to this, so maybe what I'm trying to achieve is not possible.
The text was updated successfully, but these errors were encountered:
Thank you Itay, I'm glad to hear that you find the library useful.
I'm sorry about the lack of a call to set_event_callback from opencl3, that was an oversight by me.
I've added a new method set_callback to the opencl3Event struct and I've also added an example in test_event to show how it can be used to call a callback function. They are in the develop branch. Please let me know if you have any problems with them.
Thank you for your quick response, works as expected, however I'm reading more about `wait_for_events` and I can see the point you made about being easier and more useful.
Firstly, thank you very much for this library, I'm new to both Rust and OpenCL and find this library with its tests and examples as a great resource.
I'm trying to use
set_event_callback
fromopencl3/cl3
but cannot find any documentation or examples on how this can be done. Can you please point me to a doc/test if one is available for this?My goal is for my code to be able to run kernel executions repeatedly while utilizing the GPUs to the maximum possible. So my plan was to avoid blocking actions in the execution path and use event callbacks in order to determine when the kernel execution has been completed. Once it completes, I can flag it and allow my code to start a new execution again. As I mentioned I'm new to this, so maybe what I'm trying to achieve is not possible.
The text was updated successfully, but these errors were encountered: