What is causing this problem?

Discussion of chess software programming and technical issues.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
Michael Sherwin
Posts: 3041
Joined: Fri May 26, 2006 1:00 am
Location: WY, USA
Full name: Michael Sherwin

What is causing this problem?

Post by Michael Sherwin » Wed Aug 16, 2017 1:35 pm

FEN: rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1

RomiChess:
1 00:00 40 40 +0.44 c2c4
2 00:00 224 224 +0.36 e2e4 g8f6
3 00:00 367 367 +0.66 e2e4 g8f6
4 00:00 946 946 +0.29 e2e4 g8f6 b1c3 e7e5
5 00:00 2,394 2,394 +0.79 e2e4 d7d5
6 00:00 8,682 8,682 +0.29 e2e4
7 00:00 17,418 17,418 +0.67 e2e4 d7d5 b1c3 g8f6 e4d5 f6d5 f1c4
8 00:00 39,877 3,987,700 +0.29 e2e4 d7d5 b1c3 g8f6 e4d5 f6d5 f1c4 c8e6
9 00:00 83,661 4,183,050 +0.45 e2e4 d7d5 e4d5 g8f6 b1c3 c8g4 g1f3 f6d5 f1c4
10 00:00 220,952 5,523,800 +0.44 e2e4 d7d5 e4d5 d8d5 b1c3 d5e5 f1e2 e5g5 e2f3 e7e5 d2d3
11 00:00 418,643 5,233,037 +0.46 e2e4 d7d5 e4d5 g8f6 b1c3 f6d5 f1c4 d5c3
12 00:00 968,144 5,378,577 +0.60 e2e4 d7d5 e4d5 g8f6 f1c4 c8f5 b1c3 b8d7 d2d4 d7b6 c4b5
13 00:00 1,871,936 5,348,388 +0.63 e2e4 d7d5 e4d5 d8d5 b1c3 d5e6 f1e2 e6b6 g1f3 g8f6 e1g1 c8f5 f3e5
14 00:00 2,761,971 5,211,266 +0.60 e2e4 d7d5 e4d5 g8f6 f1c4 c8g4 f2f3 g4f5 b1c3 b8d7 d2d4 d7b6 c4b5
15 00:00 4,000,588 5,128,958 +0.60 e2e4 d7d5 e4d5 g8f6 f1c4 c8g4 f2f3 g4f5 b1c3 b8d7 d2d4 d7b6 c4b5 f5d7 b5d7 d8d7 c1f4
16 00:01 6,159,631 4,927,704 +0.60 e2e4 d7d5 e4d5 g8f6 f1c4 c8g4 f2f3 g4f5 b1c3 b8d7 d2d4 d7b6 c4b5 f5d7 b5d7 d8d7 c1f4 b6d5
17 00:02 15,361,050 5,242,679 +0.48 e2e4 d7d5 e4d5 g8f6 f1c4 f6d5 b1c3 c8e6 d1h5 c7c6 d2d4 b8d7 c1d2 g7g6 h5g5 d8b6 c3d5
18 00:05 29,132,374 5,306,443 +0.38 e2e4 d7d5 e4d5 g8f6 f1c4 f6d5 b1c3 c8e6 d1h5 c7c6 d2d3 d5f6 h5h4 e6c4 h4c4 b8d7 c1f4
19 00:10 56,513,436 5,387,362 +0.40 e2e4 d7d5 e4d5 g8f6 f1c4 f6d5 b1c3 c8e6 d1h5 c7c6 c3d5 e6d5 c4d5 d8d5 h5d5 c6d5 g1f3 b8d7 e1g1
20 00:22 125,239,896 5,505,050 +0.44 e2e4 d7d5 e4d5 g8f6 f1c4 f6d5 b1c3 c8e6 d1h5 c7c6 c4d5 e6d5 c3d5 d8d5
21 01:33 533,308,224 5,719,736 +0.43 e2e4 e7e6 b1c3 d7d5 d2d4 g8f6 e4e5 f6d7 c1e3 c7c5 f1b5 b8c6 g1f3 f8e7

Is it normal to take so long when the second move changes?
I hate if statements. Pawns demand if statements. Therefore I hate pawns.

Sven
Posts: 3826
Joined: Thu May 15, 2008 7:57 pm
Location: Berlin, Germany
Full name: Sven Schüle
Contact:

Re: What is causing this problem?

Post by Sven » Wed Aug 16, 2017 6:19 pm

What is the exact meaning of the numbers in column 3 and 4 of each row?

User avatar
pedrox
Posts: 992
Joined: Fri Mar 10, 2006 5:07 am
Location: Basque Country (Spain)
Contact:

Re: What is causing this problem?

Post by pedrox » Wed Aug 16, 2017 7:31 pm

By logic are nodes and nodes per second.

Not bad almost 6 M of nodes per second for Romichess with a single thread, this is 10 times more nps than my engine or perhaps I should buy new computer.

bob
Posts: 20555
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

Re: What is causing this problem?

Post by bob » Wed Aug 16, 2017 8:12 pm

If the second move changes, that means move ordering was broken below that node. Not that unusual. Only thing I would do is compare the two PVs to make sure something is not broken to make you expend that kind of effort needlessly.

User avatar
JVMerlino
Posts: 1003
Joined: Wed Mar 08, 2006 9:15 pm
Location: San Francisco, California

Re: What is causing this problem?

Post by JVMerlino » Wed Aug 16, 2017 8:19 pm

bob wrote:If the second move changes, that means move ordering was broken below that node. Not that unusual. Only thing I would do is compare the two PVs to make sure something is not broken to make you expend that kind of effort needlessly.
Indeed. The time to reach depth 21 was 4.2x the previous depth. Not something you'd want for every iteration, but definitely not something to be concerned about. Fail lows at the root can take 10x (or much more) to resolve. But, as Bob says, making sure that the search worked as intended is not unreasonable.

Michael Sherwin
Posts: 3041
Joined: Fri May 26, 2006 1:00 am
Location: WY, USA
Full name: Michael Sherwin

Re: What is causing this problem?

Post by Michael Sherwin » Wed Aug 16, 2017 9:32 pm

pedrox wrote:By logic are nodes and nodes per second.

Not bad almost 6 M of nodes per second for Romichess with a single thread, this is 10 times more nps than my engine or perhaps I should buy new computer.
Yes, nodes and nodes per second. Output is from Arena.

My computer is an i7 6 core overclocked to 3.8 gHz, water cooled.
I hate if statements. Pawns demand if statements. Therefore I hate pawns.

Michael Sherwin
Posts: 3041
Joined: Fri May 26, 2006 1:00 am
Location: WY, USA
Full name: Michael Sherwin

Re: What is causing this problem?

Post by Michael Sherwin » Wed Aug 16, 2017 9:37 pm

bob wrote:If the second move changes, that means move ordering was broken below that node. Not that unusual. Only thing I would do is compare the two PVs to make sure something is not broken to make you expend that kind of effort needlessly.
Seems fine then. Match in progress. Romi vs Yace 0.99.87 +30 -30 =24. Romi usually performs sub par against Yace so equal is fantastic for Romi! :)
I hate if statements. Pawns demand if statements. Therefore I hate pawns.

Post Reply