Three mismatch positions from perft(7) work unit 484

Discussion of chess software programming and technical issues.

Moderators: hgm, Dann Corbit, Harvey Williamson

User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Three mismatch positions from perft(7) work unit 484

Post by sje »

Three mismatch positions from perft(7) work unit 484

Code: Select all

22930c22930
< rn1qkbnr/ppp1ppp1/7p/3p4/4P3/2NQ3b/PPPP1PPP/R1B1KBNR b KQkq - 3 4 24 37326031982 895824767568
---
> rn1qkbnr/ppp1ppp1/7p/3p4/4P3/2NQ3b/PPPP1PPP/R1B1KBNR b KQkq - 3 4 24 37326032161 895824771864
22997c22997
< rn1qkbnr/ppp1ppp1/7p/3p4/4P3/3Q2Pb/PPPP1P1P/RNB1KBNR b KQkq - 0 4 24 34544134235 829059221640
---
> rn1qkbnr/ppp1ppp1/7p/3p4/4P3/3Q2Pb/PPPP1P1P/RNB1KBNR b KQkq - 0 4 24 34544134593 829059230232
45813c45813
< rn1qkbnr/ppp1pppp/3p4/1N6/4b3/4PP2/PPPP2PP/R1BQKBNR b KQkq - 2 4 24 17617685736 422824457664
---
> rn1qkbnr/ppp1pppp/3p4/1N6/4b3/4PP2/PPPP2PP/R1BQKBNR b KQkq - 2 4 24 17617686720 422824481280
Top
AndrewGrant
Posts: 1660
Joined: Tue Apr 19, 2016 6:08 am
Location: U.S.A
Full name: Andrew Grant

Re: Three mismatch positions from perft(7) work unit 484

Post by AndrewGrant »

rn1qkbnr/ppp1ppp1/7p/3p4/4P3/2NQ3b/PPPP1PPP/R1B1KBNR b KQkq - 3 4 37326032161

rn1qkbnr/ppp1ppp1/7p/3p4/4P3/3Q2Pb/PPPP1P1P/RNB1KBNR b KQkq - 0 4 34544134593

rn1qkbnr/ppp1pppp/3p4/1N6/4b3/4PP2/PPPP2PP/R1BQKBNR b KQkq - 2 4 17617686720

No guarantees that my program is bug free :)
Talkchess is dead without moderation. If you want my attention, contact me via andrew@grantnet.us
Top
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Re: Three mismatch positions from perft(7) work unit 484

Post by sje »

Thank you for your calculations. Your results match Symbolic's with its 128 bit signatures.

So far, Symbolic has matched close to ten million results generated by Ankan's 64 bit signature program. Of the few dozen mismatches, Symbolic's results have been independently verified. This is a good indication that Symbolic's output is error free. The quality of this indication will improve if the upcoming thirty million results also match or are otherwise verified.

The most likely error source? Transient, undetected memory or processor corruption due to cosmic rays or local radioisotope decay; this occurring while running a work unit outside the 400-799 range and so without independent verification.
Top
User avatar
cdani
Posts: 2204
Joined: Sat Jan 18, 2014 10:24 am
Location: Andorra

Re: Three mismatch positions from perft(7) work unit 484

Post by cdani »

Same results for Andscacs:

Code: Select all

Andscacs 0.86205 by Daniel Jose
pf rn1qkbnr/ppp1ppp1/7p/3p4/4P3/2NQ3b/PPPP1PPP/R1B1KBNR b KQkq - 3 4
perft 7
37326032161
ms&#58; 457803  n/s&#58; 81532956
pf rn1qkbnr/ppp1ppp1/7p/3p4/4P3/3Q2Pb/PPPP1P1P/RNB1KBNR b KQkq - 0 4
perft 7
34544134593
ms&#58; 441715  n/s&#58; 78204576
pf rn1qkbnr/ppp1pppp/3p4/1N6/4b3/4PP2/PPPP2PP/R1BQKBNR b KQkq - 2 4
perft 7
17617686720
ms&#58; 228852  n/s&#58; 76982882
Top
pkumar
Posts: 99
Joined: Tue Oct 15, 2013 5:45 pm

Re: Three mismatch positions from perft(7) work unit 484

Post by pkumar »

perft D7 rn1qkbnr/ppp1ppp1/7p/3p4/4P3/2NQ3b/PPPP1PPP/R1B1KBNR b KQkq - 3 4
D7: 37326032161

perft D7 rn1qkbnr/ppp1ppp1/7p/3p4/4P3/3Q2Pb/PPPP1P1P/RNB1KBNR b KQkq - 0 4
D7: 34544134593

perft D7 rn1qkbnr/ppp1pppp/3p4/1N6/4b3/4PP2/PPPP2PP/R1BQKBNR b KQkq - 2 4
D7: 17617686720
Top

Return to “Computer Chess Club: Programming and Technical Discussions”