Concurrent instances of TensorRT

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Dear all,

I’m posting two questions for which I didn’t manage to find an answer on the documentation or in other threads of this forum.

QUESTION 1)
Is
nvinfer1::IExecutionContext::execute
implicitly creating a stream, which is then used to execute the inference?

QUESTION 2)
Consider a TX2 platform.
Suppose that two CPU threads create two different contexts (for two different networks),
and then call IExecutionContext::execute on their corresponding context.

Will the two inference tasks be serialized or will run concurrently (sharing the GPU in some way)?

Thanks in advance for any answer.