Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Pentium 3 Optimized

Author: Aaron Gordon

Date: 03:04:48 09/28/01

Go up one level in this thread


On September 28, 2001 at 04:43:49, Bernhard Bauer wrote:

>On September 27, 2001 at 15:50:23, Aaron Gordon wrote:
>
>>Not sure how much this will help but it's pretty much the same stuff as the
>>other exe's but profiled on a P3. Lemme know how it goes..
>>
>>ftp://65.10.79.9/pub/crafty/CraftyP3.zip
>
>For the following position which is treated in another thread
>
>[D]R7/5prp/4pk2/1p2P3/5p1P/1B1n1P2/P4Pr1/5R1K b - -
>
>Bob's wcrafty gives:
>                9->   1.05  -5.20   1. ... Nxe5 2. Bd1 Nd3 3. Be2 Nxf2+
>                                    4. Rxf2 Rxf2 5. Bxb5 Rgg2 6. Rh8
>               10     1.66  -5.26   1. ... Nxe5 2. Bd1 Nd3 3. Bc2 Nxf2+
>                                    4. Rxf2 Rxf2 5. Be4 Rf1+ 6. Kh2 b4
>                                    7. Rb8 Rf2+ 8. Kh1 Rb2
>               10->   1.77  -5.26   1. ... Nxe5 2. Bd1 Nd3 3. Bc2 Nxf2+
>                                    4. Rxf2 Rxf2 5. Be4 Rf1+ 6. Kh2 b4
>                                    7. Rb8 Rf2+ 8. Kh1 Rb2
>               11     3.11  -5.59   1. ... Nxe5 2. Bd1 Nd3 3. Bc2 Nxf2+
>                                    4. Rxf2 Rxf2 5. Be4 Rg3 6. Ra5 Rh3+
>                                    7. Kg1 Rb2 8. Bxh7 Rxh4
>               11->   3.44  -5.59   1. ... Nxe5 2. Bd1 Nd3 3. Bc2 Nxf2+
>                                    4. Rxf2 Rxf2 5. Be4 Rg3 6. Ra5 Rh3+
>                                    7. Kg1 Rb2 8. Bxh7 Rxh4
>               12     6.98     ++   1. ... Nxe5!!
>               12    22.19  -7.63   1. ... Nxe5 2. Bd5 exd5 3. Ra6+ Kf5
>                                    4. Ra3 b4 5. Rb3 d4 6. Rxb4 Nxf3 7.
>                                    Rb5+ Ke4 8. Ra1
>               12->  22.75  -7.63   1. ... Nxe5 2. Bd5 exd5 3. Ra6+ Kf5
>                                    4. Ra3 b4 5. Rb3 d4 6. Rxb4 Nxf3 7.
>                                    Rb5+ Ke4 8. Ra1
>               13    34.69     ++   1. ... Nxe5!!
>               13     2:51 -Mat07   1. ... Nxe5 2. Bd5 R2g6 3. Bxe6 Nxf3
>                                    4. Bg4 Rxg4 5. Ra6+ Kf5 6. Rh6 Rg1+
>                                    7. Rxg1 Rxg1#
>               13->   2:53 -Mat07   1. ... Nxe5 2. Bd5 R2g6 3. Bxe6 Nxf3
>                                    4. Bg4 Rxg4 5. Ra6+ Kf5 6. Rh6 Rg1+
>                                    7. Rxg1 Rxg1#
>              time=2:53  cpu=203%  mat=0  n=98308065  fh=97%  nps=568k
>              ext-> chk=8750353 cap=174933 pp=225234 1rep=1464970 mate=174639
>              predicted=0  nodes=98308065  evals=1877708
>              endgame tablebase-> probes done=0  successful=0
>              SMP->  split=693  stop=47  data=9/32  cpu=5:52  elap=2:53
>
>While your CraftyP3 gives:
>
>                9->   0.49  -5.03   1. ... Kxe5 2. Ra5 Nxf2+ 3. Rxf2 Rxf2
>                                    4. Rxb5+ Kf6 5. Rb7 Rg3 6. a4 Rgxf3
>               10     0.86  -5.42   1. ... Kxe5 2. Ra5 Nxf2+ 3. Rxf2 Rxf2
>                                    4. Rxb5+ Kf6 5. Rg5 Rxg5 6. hxg5+ Kxg5
>                                    7. a4 Rxf3 8. Bc4
>         (2)   10->   1.08  -5.42   1. ... Kxe5 2. Ra5 Nxf2+ 3. Rxf2 Rxf2
>                                    4. Rxb5+ Kf6 5. Rg5 Rxg5 6. hxg5+ Kxg5
>                                    7. a4 Rxf3 8. Bc4
>               11     2.22     ++   1. ... Kxe5!!
>         (2)   11->   2.46  -5.81   1. ... Kxe5 2. Ra5 Nxf2+ 3. Rxf2 Rxf2
>                                    4. Rxb5+ Kf6 5. Rg5 Rxg5 6. hxg5+ Kxg5
>                                    7. a4 Rxf3 8. Bc4
>               12     4.71  -5.78   1. ... Kxe5 2. Ra5 Nxf2+ 3. Rxf2 Rxf2
>                                    4. Rxb5+ Kf6 5. Rb4 Rxf3 6. Kh2 Rf2+
>                                    7. Kh1 h5 8. Rb5 e5 9. a4
>               12    10.24     ++   1. ... Nxe5!!
>               12    21.60 -16.17   1. ... Nxe5 2. Bd5 R2g6 3. Bxe6 Nxf3
>                                    4. Bg4 Rxg4 5. Ra6+ Ke5 6. Rg6 R4xg6
>                                    7. Re1+ Nxe1 8. a3 Rg1+ 9. Kh2 R7g2+
>                                    10. Kh3 Rxf2
>               12->  21.63 -16.17   1. ... Nxe5 2. Bd5 R2g6 3. Bxe6 Nxf3
>                                    4. Bg4 Rxg4 5. Ra6+ Ke5 6. Rg6 R4xg6
>                                    7. Re1+ Nxe1 8. a3 Rg1+ 9. Kh2 R7g2+
>                                    10. Kh3 Rxf2
>               13    22.71     ++   1. ... Nxe5!!
>               13    47.72 -Mat07   1. ... Nxe5 2. Bd5 R2g6 3. Ra4 bxa4
>                                    4. Kh2 exd5 5. Kh3 Nxf3 6. h5 Rg3+
>                                    7. fxg3 Rxg3#
>               13->  47.77 -Mat07   1. ... Nxe5 2. Bd5 R2g6 3. Ra4 bxa4
>                                    4. Kh2 exd5 5. Kh3 Nxf3 6. h5 Rg3+
>                                    7. fxg3 Rxg3#
>              time=48.28  cpu=210%  mat=0  n=25624646  fh=95%  nps=530k
>              ext-> chk=1967221 cap=52863 pp=91939 1rep=335874 mate=121960
>              predicted=0  nodes=25624646  evals=1688740
>              endgame tablebase-> probes done=0  successful=0
>              hashing-> trans/ref=36%  pawn=99%  used=65%
>              SMP->  split=506  stop=79  data=9/32  cpu=1:41  elap=48.28
>
>For the bench command I get
>
>   Wcrafty          CraftyP3
>95 385 463 nodes  89 453 914
>   404 175  nps      428 009
>       236 elapsed       209
>   2.71186   SMP     3.0622
>
>All on a dual P3-450.
>
>How did you compile?
>
>Kind regards
>Bernhard

Thats cool, your scores are close to a single CPU P3-800 box. Friend of mine
benchmarked his P3-800 and got..

Total nodes: 104161030
Raw nodes per second: 493654
Total elapsed time: 211
SMP time-to-ply measurement: 3.033175

I just compiled with...
CFLAGS = -O2 -Qipo -G6 -Qunroll -Qprof_gen -Qprof_dir c:\opt

Then ran the program for a while & benchmarked it on a P3 then recompiled with..
CFLAGS = -O2 -Qipo -G6 -Qunroll -Qprof_use -Qprof_dir c:\opt

I've been trying new compiler flags but so far this seems to be fastest (for
me).. I'll post new exe's if I find something faster.



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.