pulsar2009-9a.exe release
Posted: 22 Jun 2009, 17:06
hi,
I posted on talkchess.com but i know some dont read there or may not go there as frequently so i'm going to respost, word for word, ( sorrry for the redundancy ) here.
at http://www.adam16mr.org/pulsar.html
Most of the work was done in crazyhouse. its a stronger crazyhouse engine. An issue with it flagging with time controls like 40 moves in x minutes should be fixed. Supports analysis mode in atomic, crazyhouse, 3checks, losers, giveaway, and regular chess. I had very few people testing that but it appeared to be working to me. New variants added, fischer random and shatranj. It has allready had ways to handicap its search depth to play like only 2 ply move or 3 ply move( pulsar page has info on the arguments to use for different wilds to limit depth), like in regular chess you can add the program argument mxO-2 to make it play a depth 2 search, ( supports mxO-1 through mxO-5 i think), that's O as in octupus, the new feature is it in addition to this can support sleep1 through sleep5, which means after it moves, it will delay or sleep that many seconds before making its move. For example with -fcp "pulsar2009-9a.exe mxO-3 sleep2" it would search up to ply 3 but would take 2 seconds a move. In losers and giveaway it should now move instantly if there is only one legal move ( a forced capture).
I figure there could be some bug reports. If they are not serious i may not release again. If there are some flaws i didnt notice ( typically stuff not broken in 2007 release ), i may consider a pulsar2007-9b release in a few weeks or month. Possibly sooner depending on how serious, i'm hoping its been tested enough.
Mike
I posted on talkchess.com but i know some dont read there or may not go there as frequently so i'm going to respost, word for word, ( sorrry for the redundancy ) here.
at http://www.adam16mr.org/pulsar.html
Most of the work was done in crazyhouse. its a stronger crazyhouse engine. An issue with it flagging with time controls like 40 moves in x minutes should be fixed. Supports analysis mode in atomic, crazyhouse, 3checks, losers, giveaway, and regular chess. I had very few people testing that but it appeared to be working to me. New variants added, fischer random and shatranj. It has allready had ways to handicap its search depth to play like only 2 ply move or 3 ply move( pulsar page has info on the arguments to use for different wilds to limit depth), like in regular chess you can add the program argument mxO-2 to make it play a depth 2 search, ( supports mxO-1 through mxO-5 i think), that's O as in octupus, the new feature is it in addition to this can support sleep1 through sleep5, which means after it moves, it will delay or sleep that many seconds before making its move. For example with -fcp "pulsar2009-9a.exe mxO-3 sleep2" it would search up to ply 3 but would take 2 seconds a move. In losers and giveaway it should now move instantly if there is only one legal move ( a forced capture).
I figure there could be some bug reports. If they are not serious i may not release again. If there are some flaws i didnt notice ( typically stuff not broken in 2007 release ), i may consider a pulsar2007-9b release in a few weeks or month. Possibly sooner depending on how serious, i'm hoping its been tested enough.
Mike