IExecutionContext.execute() vs IExecuteContext.execute_async()


I am using TensorRT in image classification. What is the difference between IExecutionContext.execute() vs IExecuteContext.execute_async()?

I have read the documentation available in the webpage, but I do not fully understand it.


The difference is that CPU is blocked until inference job finished in IExecutionContext.execute().
But CPU won’t wait for GPU task but only launch in IExecuteContext.execute_async().