: I will have lots of threads downloading files using Net::FTP,
There is a limit to how much benefit you can derive from running downloads concurrently. Namely, the bandwidth of your internet connection. Even if you're sitting on the end of a very fat pipe downloading from independant, throttling servers, I'd seriously doubt you'll see much improvement beyond low 10s of threads. Unless all the servers are on a local Gigabit network.
Not all of the per-thread memory allocation is stack. Most of it is cloned internal datastructures that you cannot do much about.
You could try playing with the stack size per the documentation and see what if any difference it makes:
use threads ('stack_size' => VALUE);This sets the default per-thread stack size at the start of the application.
Perl doesn't use the (C) stack for much. Especially since the regex engine was change from recursive to iterative.
Personally, I modify the header of my perl executables to reduce the per-thread stack allocation to a minimum (4k), and I've never encountered a problem with this. I don't know if this is possible with *nix executables.
In reply to Re^19: Monitor list of threads running
by BrowserUk
in thread Monitor list of threads running
by mr_p
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |