On 7/25/06, Anton Luht <[EMAIL PROTECTED]> wrote:

Hello,

I'd like to suggest people that know some algorithmic tricks to look
at the corresponding areas of classlib. Maybe some of the community
members had (un)lucky experience of tuning performance of some
application that led them to a deep knowledge of some very specific
area has good approaches but not too widely known to an average
developer. It might also be an interest for students and other people
from institutes and universities.

What I'm talking about is not related to 'premature optimization' that
is now being discussed in another thread but something very narrow,
limited to maybe one method and not influencing anything else.

For example, consider current implementation of
java.util.BitSet.cardinality() . It just checks all bits one by one
and increments count. [1] Gives an overview of algorithms for checking
set bit count. The fastest algorithms are with table lookup, which
requires additional memory, but there are algorithms that are several
times faster than iteration (in C) and don't require tables.


+1.

I believe
such local optimizations with good comments and links to corresponding
resources won't do any harm and are a good point to start with.

Does it make sence?


Yes. But I don't think it's time to shift too much efforts to this area now.
As you mentioned  above, we are glad to see if anyone volunteers to do such
kind of things.

I also suggest to focus on proper codes, which are hot or core. Never waste
time on seldom executed codes. :)

Thanks.

[1] http://www-db.stanford.edu/~manku/bitcount/bitcount.html

--
Regards,
Anton Luht,
Intel Middleware Products Division

---------------------------------------------------------------------
Terms of use : http://incubator.apache.org/harmony/mailing.html
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]




--
Andrew Zhang
China Software Development Lab, IBM

Reply via email to