About dla core engine built

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

I just want to know about the dla core usage.
when we built an engine, we specify the dlaCore number.
does it mean that it only use that tensorcore even my gpu have two or more tensor core?

2

Hi @017 ,
Hope this helps.

Thanks

3

ok, after weeks of reading and practicing, I can answer myself as below.
an engine built with dlacore 0 will always use that dlacore 0. so we can build two instances of engine and run two engines in two threads parallelly. in this case, the overall throughput may be increased.
2. dlacore is different from tensor cores. dlacore is a logic unit while tensor core is physical unit .

correct me if I am wrong. Thanks.