On Fri, Jul 30, 2010 at 9:36 PM, Evan Cofsky <evan@theunixman.com> wrote:
On 07/27 11:48, Maciej Fijalkowski wrote:
Right now, no. But there are ways in which you can experiment. Truly concurrent threads (depends on implicit vs explicit shared memory) might require a truly concurrent GC to achieve performance. This is work (although not as big as removing refcounting from CPython for example).
Would starting to remove the GIL then be a useful project for someone (like me, for example) to undertake? It might be a good start to experimentation with other kinds of concurrency. I've been interested in Software Transactional Memory (http://en.wikipedia.org/wiki/Software_transactional_memory).
-- Evan Cofsky <evan@theunixman.com>
I think removing GIL is not a good place to start. It's far too complex without knowing codebase (it's fairly complex with knowing codebase). There are many related projects, which are smaller in size and eventually might lead to having some idea how to remove the GIL. If you're interested, come to #pypy on IRC to discuss. Cheers, fijal