Make mk-log-parser faster

[GoogleCodeExporter]

I've attached the results of parsing an 8GB file with mk-log-parser r2565

The following command is useful:

zcat smallprof.out.gz | sort -k 2nr,2 | head -n 30

Original issue reported on code.google.com by baron.schwartz on 30 Nov 2008 at 2:13

Attachments:

• smallprof.out.gz

[GoogleCodeExporter]

There are some log-parsing speed and accuracy (and simplicity!) improvements in 
r2588
and the few revisions before it.  The biggest abuser is still
QueryRewriter::fingerprint().

Original comment by baron.schwartz on 7 Dec 2008 at 11:53

• Added labels: Priority-High
• Removed labels: Priority-Medium

[GoogleCodeExporter]

The changes in r2600 need to be profiled.

Original comment by baron.schwartz on 11 Dec 2008 at 7:48

[GoogleCodeExporter]

The slowest part is still fingerprinting (by a very large margin.)  See issue 
133.

Original comment by baron.schwartz on 14 Dec 2008 at 10:45

[GoogleCodeExporter]

Features targeted for query review improvements

Original comment by baron.schwartz on 16 Dec 2008 at 9:27

• Added labels: Milestone-query_review

[GoogleCodeExporter]

I'm calling it quits as of r2668.  See the profiling results in r2669.

Original comment by baron.schwartz on 25 Dec 2008 at 11:18

[GoogleCodeExporter]

Changes are rolled into mk-log-parser as of r2670.

Original comment by baron.schwartz on 25 Dec 2008 at 11:22

• Changed state: Fixed