Advice wanted on parallel processing
Brandon S. Allbery KF8NH
allbery at ece.cmu.edu
Wed Mar 18 19:19:09 EDT 2009
On 2009 Mar 18, at 16:59, Colin Paul Adams wrote:
>>>>>> "Brandon" == Brandon S Allbery KF8NH <allbery at ece.cmu.edu>
>>> The array has 12 elements.
> Brandon> How many times do you call it? Perhaps the real
> Brandon> optimization you need is to memoize.
> Very many times indeed. But it is a different array on most occasions
> (I am not updating in place).
> It represents one rank of the board on which the game is played. The
> move generator produces a new board for each move.
It might be helpful for you to show more of the program.
One thing to keep in mind about parallelization is that the level at
which happens matters: if individual calculations across the list are
fast, all you gain by parallelizing it is MP synchronization/locking
overhead. On the other hand, it's possible that if the caller can be
rearranged so that rank_value is computed on another CPU while it's
doing something else, you could gain quite a bit. Or maybe the caller
is itself at the right level to parallelize.
brandon s. allbery [solaris,freebsd,perl,pugs,haskell] allbery at kf8nh.com
system administrator [openafs,heimdal,too many hats] allbery at ece.cmu.edu
electrical and computer engineering, carnegie mellon university KF8NH
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 195 bytes
Desc: This is a digitally signed message part
Url : http://www.haskell.org/pipermail/glasgow-haskell-users/attachments/20090318/4072456f/PGP.bin
More information about the Glasgow-haskell-users