Compute optimal interpreter check interval for Zope
Computes the optimal interpreter check interval for the machine it is run on. The formula is as follows:
<result of pystone benchmark> / 50
Installation and use
The checkinterval script must be installed into the Python interpreter used to run Zope. This is most easily achieved by adding a part to your buildout.cfg:
[buildout] parts = checkinterval [checkinterval] recipe = zc.recipe.egg eggs = jarn.checkinterval
After re-running buildout, type:
$ ./bin/checkinterval 1305
The number you see is the recommended check interval for this machine; put it into your zope.conf file:
Now restart Zope and bask in the glow.
The Python Library Reference on the topic of check interval: “This integer value determines how often the interpreter checks for periodic things such as thread switches and signal handlers. The default is 100, meaning the check is performed every 100 Python virtual instructions. Setting it to a larger value may increase performance for programs using threads.”
Now, the Zope application server is such a program, and it benefits greatly from setting the right check interval. If the value is too low, Zope threads are interrupted unnecessarily, causing a noticable performance hit on today’s multi-cpu hardware.
Where’s the 50 coming from?
The constant 50 in the formula was determined by benchmarks performed at Zope Corporation and has become part of the “Zope lore” (See e.g. this post by Matt Kromer). Going beyond pystone/50 produced no further benefits.
The value may well be meaningless for applications other than Zope and platforms other than Intel.
More on check intervals and the GIL from David Beazly.
For those back from the Beazly talk: Zope uses long running threads and asyncore, making it (more) independent from OS scheduling issues. Still, the interruption argument holds.
1.0 - 2009-06-13
- Initial release