4 ms·
Interesting. I wonder how it manages the threading with the GIL and all that.
by weberc2 8y ago
Interesting. I wonder how it manages the threading with the GIL and all that.
- jashmatthews 8y agoThread per request works because the GIL is released during blocking IO. You only need one process per core and each process runs multiple threads. It saves a massive amount of RAM compared to trying to cram 50 (g)unicorns onto a server.