jessenoller.com

⁴⁴ So, with the recent fur­vor over the GIL, one of the things GvR asked for was for some­one to pro­vide some bench­mark num­bers of Sin­gle v. Multi v. Other in var­i­ous tests.

I started work­ing on this a night or so ago, and a lot of things have fallen out of it — not the least of which is my burn­ing desire to really look at the GIL, thread­ing in python and the ecosys­tem of alternatives.

A thing of Note: I’ve spawned an ini­tial google code project to explore the alternatives/benchmarks. Read about it here.

The first test script I wrote was a sim­ple one: take a func­tion which cal­cu­lated a num­ber of fibonacci num­bers and run it in a loop with larger and larger sets. I then:

• Ran it twice, synchronusly.
• Ran it twice, once in each thread (2 threads).
• Ran it twice, but used the pro­cess­ing mod­ule to spawn to processes, run­ning it once in each.
• Ran it twice, but used the par­al­lel python mod­ule to spawn a pool of work­ers for 2 proces­sors, run­ning it once in each.

Yes. I know that fib cal­cu­la­tions are a processor-only activ­ity, and that’s all I wanted to do at first. I want to add tests for file access, net­work access, and I want to run the tests (as allowed by the exten­stion mod­ules) in PyPy, Jython and Iron­Python⁴⁴.

In the long run — I want to make some­thing of a test suite to not only demo the var­i­ous methods/alternatives but to also to work towards a future where some­thing like par­al­lel python or the pro­cess­ing mod­ule can make it into the stan­dard python library.

Not to men­tion, for those look­ing for alter­na­tives to basic thread­ing in python (or more infor­ma­tion on thread­ing in python). Yes, things that have to be taken into account:

• Shared vs. unshared mem­ory and state.
• Ease of use/API
• Amount of Fun.
• Can the solu­tion spread across machines?
• Is it, or can it be, use­ful for web appli­ca­tions? (i.e: can it be used to spread load across processors)

If you have a test or a sug­ges­tion — drop me an email or post a com­ment. Heck, this might just turn out to be a learn­ing expe­ri­ence for me or it could turn out to be some­thing bigger.

Update: Wow, you guys are awe­some. The pri­vate and pub­lic com­ments I’ve got­ten on this have left me with a lot to think about — and hell — a lot to learn. I’m still in the toying/planning phase, so please keep feed­ing me infor­ma­tion. When I get more time this week­end (if I get time) I will try to put up some­thing cohe­sive, includ­ing a sub­ver­sion repos­i­tory and a maybe a wiki.
Update 2: Hello Red­dit.4

1. Yes, this is a gratu­ti­tous pic­ture of my daughter444
2. Yes, Daniel Watkins has started this process as well444

4

44 So, with the recent fur­vor over the GIL, one of the things GvR asked for was for some­one to pro­vide some bench­mark num­bers of Sin­gle v. Multi v. Other...