Python Boost Performance

I am working on a project where I am experimenting with boost python. While learning how to organize my python interface, I came across a comment stating that python boost issues are related to performance. Is there any real problem with his work?

In this case, I am working with a large project, and we want to expose part of it in python. I find boost python makes it easy to display classes that I already have. Therefore, I would prefer to stick with python methods with increasing class exposure, because it is so simple. If someone has no alternative that is just as easy to use and execute.

+5
source share
2 answers

We use boost :: python to integrate a large computer vision library into a highly customizable software package for researchers in other fields. We did not face problems and problems. However, recently we have not conducted comparative tests.

+2
source

If your use case requires a lot of calls between Python and C ++ in a tight loop, then Boost.Python might be a performance issue, at least with respect to hand wrappers that use the Python C-API directly, It's much harder to guess if it will work worse than something like that, convenient for the user, for example, SWIG.

, - API, ++/Python, , , , , , . Python ++ Python , , Python-to-++ .

+2

All Articles