|
From: Bill H. <goo...@go...> - 2008-08-01 08:57:33
|
I should be able to upgrade to a 64 bit kernel tomorrow. That will also speed things up slightly. It's not clear to me that we actually need to use threads. We could just run the same program 4 times starting at different points each time. So I don't know if we really need OpenMP or not. If we don't use it, we don't have to worry about threads having to wait for each other. It's up to you of course. Lot's of fun to play with OpenMP anyway. The suggestions you've given do seem like they should speed things up quite a bit. Bill. 2008/8/1 Peter Shrimpton <ps...@gm...>: > Hello > > I have written some code to do the 2-SPRP test Lucas test and Pocklington > code in parallel (See attached). I am fairly certain it is working but > unfortunately it is pretty slow (and currently only dealing with numbers up > to 10^10 because it is only 32bits). I think the following will speed it up: > > 1)Do multiple Pocklington tests at once > 2)Redo the parallel sieving part: When a thread reaches the end of the > current sieved chunk it starts sieving the next chunk. However all of the > other threads have to wait until it has finished sieving before they can > continue. > 3)Store part of the Lucas chain. Currently every time it does a Lucas test > it calculates the Lucas chain from the beginning which is clearly not > necessary. > > Peter > > ------------------------------------------------------------------------- > This SF.Net email is sponsored by the Moblin Your Move Developer's challenge > Build the coolest Linux based applications with Moblin SDK & win great > prizes > Grand prize is a trip for two to an Open Source event anywhere in the world > http://moblin-contest.org/redirect.php?banner_id=100&url=/ > _______________________________________________ > flint-devel mailing list > fli...@li... > https://lists.sourceforge.net/lists/listinfo/flint-devel > > |