I would see the benefit of this fine-grained multi-threading only if we had massively parallelised CPU architecture as well (e.g. one of those funky asynchronous chips with many many pipelines). Not only that, the OS has to allow the individual process itself manage its own threads in a very very lightweight way.
Right now, however, even on multi-core processors, the OS still has to do a fair bit of work to establish and tear down a thread. Unless perl maintains a pool of threads that can be near-instantly reused for tiny tiny tasks, this is not too useful as a pragma.
Nonetheless, functional programming is a tasty tasty concept that I wish had more real-life applicability in my work.
In reply to Re: RFC: Implicit Parallelization Pragma
by Anonymous Monk
in thread RFC: Implicit Parallelization Pragma
by hardburn
For: | Use: | ||
& | & | ||
< | < | ||
> | > | ||
[ | [ | ||
] | ] |