2 ms·
More realistically, as it happened in ML/AI scene, the knowledgeable people will write the complex libraries and will hand these down to scientists and other le
by bayindirh 1y ago
More realistically, as it happened in ML/AI scene, the knowledgeable people will write the complex libraries and will hand these down to scientists and other less experienced, or risk-averse developers (which is not a bad thing).
With the critical mass Python acquired over the years, GIL becomes a very sore bottleneck in some cases. This is why I decided to learn Go, for example. Properly threaded (and green threaded) programming language which is higher level than C/C++, but lower than Python which allows me to do things which I can't do with Python. Compilation is another reason, but it was secondary with respect to threading.
- bgwalter 1y agoKnowledgeable people? Pytorch has memory leaks by design, it uses std::shared_ptr for a graph with cycles. It also has threading issues.