Very hard to say which Toga is strongest, but personally I prefer Toga 1.4.1SE. Up to this point there was clear continuous improvement. After this point it's only mess with dozens of forks but without practically any improvement.
So personally I consider 1.4.1SE as a final version of Toga. If someone seriously continuous from here, I strongly suggest to change name to avoid confusion.
So personally I consider 1.4.1SE as a final version of Toga. If someone seriously continuous from here, I strongly suggest to change name to avoid confusion.
I agree.
Unfortunately 1.4.1SE has a flawed MP implementation. Instead of YBW it uses completely independent search threads which share the hash table. This gives great nps but little improvement in playing strength.
As far as I know none of the many forks of Toga actually tried to fix Toga's MP implementation.
I see there are many "Merge" versions, the latest, Toga II Merge_b15_09, released yesterday. I would think the "latest" should be the "greatest", but with so many forks, who knows?
Too many version released is not the problem. Toga has an MP bug which causes it to crash almost every time when using 2-4 or more cores.....
It gives the Toga Team a huge headache at the moment, since they are unable to find out what is really wrong with it, but they are doing their best to try and fix it. Why do think there are so many versions released ?
Too many version released is not the problem. Toga has an MP bug which causes it to crash almost every time when using 2-4 or more cores.....
I thought 2 threads was ok (there was one book bug which I fixed).
I wonder if the hash table might be the problem. Last time I looked Toga did not lock or protect in any way the hash table while it is being accessed by multiple threads.