Richard Jones' Log: NeedForSpeed[-1]

Sat, 27 May 2006

We're half-way through the 6th and final day of the sprint. I've just committed the new-style exceptions branch to the trunk.


Overall, my laptop (1.5GHz G4) sees an increase in pystones from 22935.8 in 2.4 maint (r46459) to 23148.1 in 2.5 trunk (r46453). This is pretty much expected as the sprint hasn't really touched anything that pystone focuses on.

On the other hand pybench will see improvements in a number of its tests. In particular the TryRaiseExcept test speeds up about 30% on 2.4, and 80% on 2.5a2 thanks to the new exceptions implementation that I wrote and Georg Brandl (or georgbot as we've just started calling him) helped clean up. I'll look into including some pybench results when Sean has finished tweaking the pybench tool.

Sean has yet more photos up from yesterday and today.