Page 2 of 2

Re: Stockfish 7 and partial 6 piece syzygy problem?

Posted: Thu Sep 01, 2016 9:14 pm
by mcostalba
syzygy wrote:
Jouni wrote:I see there has been syzygy rewriting lately. May be this can be fixed at the same time?
The SF rewrite will only slow down the probing code with no benefit (I've already shown how to do the rewrite properly, but there is no interest).

However, I have just fixed it here, because it is a bug to do anything with TBScore if the probe was not successful:
https://github.com/syzygy1/Cfish/blob/e ... rch.c#L886

Even if this were fixed in SF, I make no guarantee that probing an incomplete set of tables does what you would like it to do.
Thanks to all people involved, for reporting the bug, finding the fix and confirming the bug.

I will open a PR and fix SF, possibly this week end.

Re: Stockfish 7 and partial 6 piece syzygy problem?

Posted: Sat Sep 03, 2016 6:23 am
by mcostalba
Bug has been fixed now.

Re: Stockfish 7 and partial 6 piece syzygy problem?

Posted: Sat Sep 03, 2016 5:12 pm
by Jouni
Seems to work now OK without any obvious drawback!