Hello Bearophine,
On Tuesday, 24 December 2013 at 23:13:59 UTC, bearophile wrote:
Some ides to speed up your code:
- Try to use parse instead of split + to!size_t
- Try to use byLineFast, in the attach here (the code is not
mine): https://d.puremagic.com/issues/attachment.cgi?id=1305
- Try to disable the GC before the associative array creation
and re-enable it when it's built. (This could double your max
memory usage or more). To disable it import GC from
core.memory and call GC.disable; and GC.enable;
This code is almost 2.5X faster on my PC (but I am using only
300_000 lines of text):
You're just too fast for me...
After incorporating your three suggestions, the entire file (11M
lines) loads in just 25 seconds (down from 3.5 minutes).
AMAZING!
Many thanks,
-gordon
