I cannot find a problem as easily with tables that do not need the 16-bit compression code (max DTZ < 580 plies).
I also cannot find a problem as easily with KRBNvKQN, which was generated earlier.
So for the moment I am guessing that the problem lies in the NUMA adaptations of the 16-bit compression code (which should have had almost no effect on that part of the code). Alternatively, it is the parallel compressed I/O code.
If someone finds an inconsistency in some other table, please let me know.
7-men Syzygy attempt
Moderators: hgm, Rebel, chrisw
-
- Posts: 5569
- Joined: Tue Feb 28, 2012 11:56 pm
-
- Posts: 560
- Joined: Sun Nov 08, 2015 11:10 pm
Re: 7-men Syzygy attempt
Ok, this is what I will do, I will regenerate this table with the non-NUMA generator, hopefully it only affects 16-bit ones.
-
- Posts: 5569
- Joined: Tue Feb 28, 2012 11:56 pm
Re: 7-men Syzygy attempt
That will tell us if the problem was introduced by one of the recent changes.noobpwnftw wrote: ↑Sat May 12, 2018 5:06 pm Ok, this is what I will do, I will regenerate this table with the non-NUMA generator, hopefully it only affects 16-bit ones.
KQNNvKQN does not need 16-bit compression and seems OK.
KQBNvKQB does need 16-bit compression and has a problem already for the position with max DTZ:
https://lichess.org/analysis/standard/1 ... 6/8/B1K5_b_-_-
-
- Posts: 5569
- Joined: Tue Feb 28, 2012 11:56 pm
Re: 7-men Syzygy attempt
I have found the problem. It only affects the .rtbz tables with max DTZ >= 580 ply that were generated with the NUMA-aware generator. I am testing a fix now.syzygy wrote: ↑Sat May 12, 2018 5:45 pmThat will tell us if the problem was introduced by one of the recent changes.noobpwnftw wrote: ↑Sat May 12, 2018 5:06 pm Ok, this is what I will do, I will regenerate this table with the non-NUMA generator, hopefully it only affects 16-bit ones.
The affected tables seem to be:
- KQBNvKQB (660 ply)
- KRRNvKRR (581 ply)
- KQBNvKQN (633 ply)
The KRBNvKQN table that was generated earlier cannot have been affected by this problem.
If you have already regenerated one of these tables with the pre-NUMA generator (or with current master), then it is correct.
-
- Posts: 560
- Joined: Sun Nov 08, 2015 11:10 pm
Re: 7-men Syzygy attempt
Master branch(with parallel compression) seems to produce correct results.syzygy wrote: ↑Sat May 12, 2018 5:45 pmThat will tell us if the problem was introduced by one of the recent changes.noobpwnftw wrote: ↑Sat May 12, 2018 5:06 pm Ok, this is what I will do, I will regenerate this table with the non-NUMA generator, hopefully it only affects 16-bit ones.
KQNNvKQN does not need 16-bit compression and seems OK.
KQBNvKQB does need 16-bit compression and has a problem already for the position with max DTZ:
https://lichess.org/analysis/standard/1 ... 6/8/B1K5_b_-_-
I'm uploading regenerated KQBNvKQB now.
-
- Posts: 5569
- Joined: Tue Feb 28, 2012 11:56 pm
Re: 7-men Syzygy attempt
Yes, it should.noobpwnftw wrote: ↑Sun May 13, 2018 3:16 amMaster branch(with parallel compression) seems to produce correct results.syzygy wrote: ↑Sat May 12, 2018 5:45 pmThat will tell us if the problem was introduced by one of the recent changes.noobpwnftw wrote: ↑Sat May 12, 2018 5:06 pm Ok, this is what I will do, I will regenerate this table with the non-NUMA generator, hopefully it only affects 16-bit ones.
KQNNvKQN does not need 16-bit compression and seems OK.
KQBNvKQB does need 16-bit compression and has a problem already for the position with max DTZ:
https://lichess.org/analysis/standard/1 ... 6/8/B1K5_b_-_-
I'm uploading regenerated KQBNvKQB now.
I have fixed the numa branch now. I have tested it by forcing 16-bit compression of KRBvKNN (incorrect before the fix, correct now). Unfortunately the problem did not occur for 5-piece tables (just too small, it seems), which is why it slipped through.
-
- Posts: 560
- Joined: Sun Nov 08, 2015 11:10 pm
Re: 7-men Syzygy attempt
I'm regenerating affected tables(KQBNvKQN,KRRNvKRR) with fixed NUMA generator.syzygy wrote: ↑Sun May 13, 2018 3:31 amYes, it should.noobpwnftw wrote: ↑Sun May 13, 2018 3:16 amMaster branch(with parallel compression) seems to produce correct results.syzygy wrote: ↑Sat May 12, 2018 5:45 pm
That will tell us if the problem was introduced by one of the recent changes.
KQNNvKQN does not need 16-bit compression and seems OK.
KQBNvKQB does need 16-bit compression and has a problem already for the position with max DTZ:
https://lichess.org/analysis/standard/1 ... 6/8/B1K5_b_-_-
I'm uploading regenerated KQBNvKQB now.
I have fixed the numa branch now. I have tested it by forcing 16-bit compression of KRBvKNN (incorrect before the fix, correct now). Unfortunately the problem did not occur for 5-piece tables (just too small, it seems), which is why it slipped through.
-
- Posts: 300
- Joined: Mon Apr 30, 2018 11:51 pm
-
- Posts: 560
- Joined: Sun Nov 08, 2015 11:10 pm
Re: 7-men Syzygy attempt
I have also removed those, and some regenerated files has been uploaded, please re-check and make sure they are correct now.
-
- Posts: 5569
- Joined: Tue Feb 28, 2012 11:56 pm
Re: 7-men Syzygy attempt
As expected, the new KQBNvKQN.rtbz solves the problems I noticed.noobpwnftw wrote: ↑Sun May 13, 2018 11:15 amI have also removed those, and some regenerated files has been uploaded, please re-check and make sure they are correct now.