Overnight pmichaud rewrote .chomp in PIR to optimize it, leading to huge improvements in performance. The .lines version of the code has gone from 55s to 21.5s; the .get version from 44s to 7.2s. Weirdly enough, the .slurp version has gone from 5.6s to 9s. (No idea why.)
The good news is that the performance of the .get version is clearly much, much better. (Still with plenty of room for improvement!) The bad news is that the penalty for using the iterator-based version is now much more significant — it adds nearly 200% to the read time!
But then, that’s another potential spot for optimization…
August 21, 2010 at 12:11 pm |
[...] Rakudo It I Never Metaop I Didn't Like « Factorial and Memoizing Timing Text File Reads, Part 2 [...]