I made some perft tests with my new Loop-code:
Conditions:
========
CPU = AMD-FX60
Compiler = MS-VC-2005
Binaries = LOOP-64BIT
FEN-Position = r3r1k1/1pq2pp1/2p2n2/1PNn4/2QN2b1/6P1/3RPP2/2R3KB b - -
Results:
======
Depth = 5
Nodes = 323.787.902
Time = 4.188
Nps = 77.313.252
Depth = 6
Nodes = 15.097.513.050
Time = 211.328
Nps = 71.441.139
Are these simple performance-results ok???
perft results
Moderators: hgm, Rebel, chrisw
Re: perft results
Now I want to find out, if it is possible to beat these results with magic bitboards.
The new results will follow in a few days...
The new results will follow in a few days...
-
- Posts: 10282
- Joined: Thu Mar 09, 2006 12:37 am
- Location: Tel-Aviv Israel
Re: perft results
Your numbers for perft are correct.LoopList wrote:I made some perft tests with my new Loop-code:
Conditions:
========
CPU = AMD-FX60
Compiler = MS-VC-2005
Binaries = LOOP-64BIT
FEN-Position = r3r1k1/1pq2pp1/2p2n2/1PNn4/2QN2b1/6P1/3RPP2/2R3KB b - -
Results:
======
Depth = 5
Nodes = 323.787.902
Time = 4.188
Nps = 77.313.252
Depth = 6
Nodes = 15.097.513.050
Time = 211.328
Nps = 71.441.139
Are these simple performance-results ok???
Movei can calculate them faster on the same hardware.
You can download movei and print in console mod
setboard r3r1k1/1pq2pp1/2p2n2/1PNn4/2QN2b1/6P1/3RPP2/2R3KB b - -
perft 6 and see it.
Note that movei is using 256 mbytes hash for calculating perft.
Movei's times for calculating perft on a slow pIV:
1)4.844 seconds for perft 5.
2)126.531 seconds for perft 6.
Uri
-
- Posts: 27794
- Joined: Fri Mar 10, 2006 10:06 am
- Location: Amsterdam
- Full name: H G Muller
Re: perft results
Seems correct:
(With my specialized perft on a 2.8GHz Pentium IV)
Code: Select all
$ ./qperft 6 "r3r1k1/1pq2pp1/2p2n2/1PNn4/2QN2b1/6P1/3RPP2/2R3KB b"
- - - - - - - - - - - -
- - - - - - - - - - - -
- - r . . . r . k . - -
- - . p q . . p p . - -
- - . . p . . n . . - -
- - . P N n . . . . - -
- - . . Q N . . b . - -
- - . . . . . . P . - -
- - . . . R P P . . - -
- - . . R . . . K B - -
- - - - - - - - - - - -
- - - - - - - - - - - -
Quick Perft by H.G. Muller
Perft mode: No hashing, bulk counting in horizon nodes
perft(1)= 56 ( 0.000 sec)
perft(2)= 2594 ( 0.000 sec)
perft(3)= 137198 ( 0.000 sec)
perft(4)= 6391735 ( 0.078 sec)
perft(5)= 323787902 ( 3.625 sec)
perft(6)= 15097513050 (183.063 sec)
$