[CDBI] Make name_lc go fast

Brad Bowman list at bereft.net
Tue Apr 3 14:12:26 BST 2007


Aaron Trevena wrote:
> Is this 5% difference ( give/take a margin of error almost as large)
> worth the trouble of investigating?

As far as Class::DBI is concerned, if a 5% speed up can be gained without
obscuring and complicating the code, it may be worthwhile.  My benchmarks
support the 5% but my use of Class::DBI is atypical, SQLite doesn't have
the overhead of client-server databases.  If there's a 5% difference running
the tests or another benchmark suite using another database then
it's something worth considering.  If any difference is lost in the
traffic noise then no bother.

> I'd suggest running some benchmarks against a realistic dataset on a
> network/server/database setup resembling the production environment
> you plan and seeing where real bottlenecks are.

The benchmarks I'm running are quite close to my production use,
although instead of 2 minutes it can take 2 hours.  Even then, 5% is
no big deal.

Brad

-- 
 When one is not capable of true intelligence, it is good to consult with
 someone of good sense. -- Hagakure http://bereft.net/hagakure/



More information about the ClassDBI mailing list