Hi all!
This is what's new on the trunk since r2000 which was my last report. This
should be message No. 1,001 which is a palindrome number and we won't have the
next such number until message No. 1,111.
But here goes:
1. I did more identifier cleanup - whitelisted more identifiers, went over the
existing ones, merged identifiers, etc. Now the command
{{{
ruby scripts/find-ids.rb *.[ch]
}}}
has an empty output.
2. The various structures of Freecell Solver used to have several allocators /
memory pools / whatever you'll want to call them to handle stuff. In some
cases, several instances of the same kind where available. Now they are merged
into one implementation and only one instance where needed.
3. I converted several ugly macros to inline functions.
4. I merged in the DLLEXPORT branch and enabled -fvisibility=hidden. What it
does is make sure non-static global identifiers are not exported from the
DLL/shared-library by default (unless explicitly specified to) and so the
shared library is smaller, and loads faster.
5. Got rid of compiling preset.c while FCS_FREECELL_ONLY is enabled.
6. I merged some duplicate code in fcs_hash.c.
7. Added the FCS_WITH_CONTEXT_VARIABLE compilation option which gets rid of
context variable in the internal hash's comparison functions (which was just
NULL). Strangely, it makes the serial benchmark worse and the thread-enabled
benchmark a bit better.
-----------------
Some possible future directions:
1. Set up unions in the instance/hard_thread/soft_thread structs for the
various scan types (to save space).
2. Regarding finding short solutions: create a meta-scan that will optimise on
the lowest maximal solution length of a given scan (at each iteration of the
quota allocation) instead of on the lowest average solution length.
Regards,
Shlomi Fish
--
-----------------------------------------------------------------
Shlomi Fish http://www.shlomifish.org/
http://www.shlomifish.org/humour/ways_to_do_it.html
God gave us two eyes and ten fingers so we will type five times as much as we
read.
Received on Sun Jun 21 2009 - 03:56:52 IDT