Author: Shaun Brewer
Date: 02:49:51 04/17/03
Go up one level in this thread
On April 16, 2003 at 15:02:45, Aaron Gordon wrote:
>On April 16, 2003 at 14:14:03, Robert Hyatt wrote:
>
>>On April 16, 2003 at 06:01:59, Aaron Gordon wrote:
>>
>>>On April 16, 2003 at 05:54:21, Javier Ros Padilla wrote:
>>>
>>>>What is the difference in chess programs between DDR and PC133 SDRAM?
>>>>
>>>>Javier Ros
>>>
>>>With smaller hash sizes (1-64mb) you'll see no speed/strength increase in chess
>>>programs. With very large hash sizes however (and depending on the program) you
>>>should see a 2-7% increase in Kn/s. I did the tests a while back, if I can find
>>>the results I'll post them here for you.
>>
An observation I posted a while back compared KT133A (SDRAM) with KT333 PC2100
and KT400 PC2700.
All PCs used XP2000+
SDRAM was faster. (not much)
Reasons (I believe).
SDRAM was decent CAS2 the PC2100 and PC2700 were generic CAS2.5(3) ?
I now have a new nforce2 board with PC2700 cas 2. I did a test adjusting my new
processor to XP2000+ 266fsb.
This new board with reasonable DDR outperformed the KT133A SDRAM (from memory
the improvement was about 5% (I dont have access at the moment MB has died)) -
obviously once I went up to 333fsb (10x166 rather than 12.5x133)the perfromance
gap increased further ?% sorry can't remember.
Shaun
Old Data:
wcrafty-18.15.exe from Bob's ftp site
crafty.rc
hash 256M
hashp 64M
read a.pgn
analyze
a.pgn
[White ""]
[Black ""]
[WhiteElo "0"]
[BlackElo "0"]
[Result "*"]
[Date "????.??.??"]
[Event ""]
[Site ""]
[Round "?"]
[ECO "B33"]
[Annotator ""]
[Source ""]
[Remark "Sicilian defence Sveshnikov"]
1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4 Nc6 5.Nb5 Nf6 6.N1c3 d6 7.Bf4
e5 8.Bg5 a6 9.Na3 b5 10.Bxf6 gxf6 11.Nd5 f5 12.c3 Bg7 13.exf5
Bxf5 *
Here are the results:
My machine KT133A 512mb SDRAM cas 2
hash table memory = 192M bytes.
pawn hash table memory = 48M bytes.
clearing hash tables
time surplus 0.00 time limit 30.00 (3:30)
nss depth time score variation (1)
6-> 0.16 0.22 14. Bd3 Be6 15. Be4 Bd7 16. c4 Nd4
7 0.25 0.31 14. Bd3 Be6 15. Be4 Rc8 16. O-O O-O
17. Qf3
7-> 0.35 0.31 14. Bd3 Be6 15. Be4 Rc8 16. O-O O-O
17. Qf3
8 0.82 0.16 14. Bd3 e4 15. Qh5 Bg6 16. Qe2 f5 17.
O-O O-O
8 1.24 0.30 14. Qf3 Be6 15. Bd3 O-O 16. O-O Rc8
17. Bf5 Qg5
8-> 1.49 0.30 14. Qf3 Be6 15. Bd3 O-O 16. O-O Rc8
17. Bf5 Qg5
9 2.80 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. Bc2 Ne7
17. Qh5 f5 18. O-O-O Nxd5 19. Rxd5
9-> 4.19 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. Bc2 Ne7
17. Qh5 f5 18. O-O-O Nxd5 19. Rxd5
10 7.17 0.21 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Kf8 17. Bd3 Kg7 18. O-O Rg8 19. Qe4
h6
10-> 16.94 0.21 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Kf8 17. Bd3 Kg7 18. O-O Rg8 19. Qe4
h6
11 23.05 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. O-O Ne7
17. Nxe7+ Qxe7 18. Be4 Rad8 19. Qh5
f5 20. Bd5+ Be6
11-> 34.52 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. O-O Ne7
17. Nxe7+ Qxe7 18. Be4 Rad8 19. Qh5
f5 20. Bd5+ Be6
12 48.01 0.17 14. Qf3 Bd7 15. Bd3 O-O 16. O-O f5
17. Bc2 e4 18. Qh5 Be8 19. Qh3 Bg6
12-> 1:56 0.17 14. Qf3 Bd7 15. Bd3 O-O 16. O-O f5
17. Bc2 e4 18. Qh5 Be8 19. Qh3 Bg6
13 3:03 0.18 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Ke7 17. Bd3 Qd7 18. Qe4 d5 19. Qb4+
Qd6 20. Qxd6+ Kxd6 21. Rd1 Bg4 22.
f3 Bd7
13-> 4:07 0.18 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Ke7 17. Bd3 Qd7 18. Qe4 d5 19. Qb4+
Qd6 20. Qxd6+ Kxd6 21. Rd1 Bg4 22.
f3 Bd7
14 6:12 0.13 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Kf8 17. Be2 d5 18. O-O Be7 19. Nc2
Qc8 20. Qxc8+ Rxc8 21. a4 d4 22. axb5
axb5 23. Rfc1
14 21:18 0.17 14. Nc2 O-O 15. a4 Ne7 16. Bd3 Bxd3
17. Qxd3 Nxd5 18. Qxd5 bxa4 19. Rxa4
Qd7 20. Rh4 Qb5 21. Nb4 Qc5
14-> 25:22 0.17 14. Nc2 O-O 15. a4 Ne7 16. Bd3 Bxd3
17. Qxd3 Nxd5 18. Qxd5 bxa4 19. Rxa4
Qd7 20. Rh4 Qb5 21. Nb4 Qc5
15 43:49 0.15 14. Nc2 O-O 15. Nce3 Be6 16. Bd3 f5
17. Bc2 Rc8 18. Qh5 e4 19. O-O Ne7
20. Rad1 <HT>
15-> 59:18 0.15 14. Nc2 O-O 15. Nce3 Be6 16. Bd3 f5
17. Bc2 Rc8 18. Qh5 e4 19. O-O Ne7
20. Rad1 <HT>
16 139:35 0.37 14. Nc2 O-O 15. Nce3 Bd7 16. Qc2 Ne7
17. Nxe7+ Qxe7 18. Bd3 h6 19. O-O-O
Be6 20. Nf5 Bxf5 21. Bxf5 b4 22. c4
b3 23. Qxb3
16-> 155:58 0.37 14. Nc2 O-O 15. Nce3 Bd7 16. Qc2 Ne7
17. Nxe7+ Qxe7 18. Bd3 h6 19. O-O-O
Be6 20. Nf5 Bxf5 21. Bxf5 b4 22. c4
b3 23. Qxb3
17 246:38 0.28 14. Nc2 O-O 15. Nce3 Be6 16. Bd3 f5
17. Qh5 e4 18. Bc2 Ne7 19. Bb3 Nxd5
<HT>
17-> 323:00 0.28 14. Nc2 O-O 15. Nce3 Be6 16. Bd3 f5
17. Qh5 e4 18. Bc2 Ne7 19. Bb3 Nxd5
<HT>
White(14): quit
execution complete.
KT400 512mb PC2700 cas 2.5
hash table memory = 192M bytes.
pawn hash table memory = 48M bytes.
clearing hash tables
time surplus 0.00 time limit 30.00 (3:30)
nss depth time score variation (1)
6-> 0.17 0.22 14. Bd3 Be6 15. Be4 Bd7 16. c4 Nd4
7 0.28 0.31 14. Bd3 Be6 15. Be4 Rc8 16. O-O O-O
17. Qf3
7-> 0.39 0.31 14. Bd3 Be6 15. Be4 Rc8 16. O-O O-O
17. Qf3
8 0.86 0.16 14. Bd3 e4 15. Qh5 Bg6 16. Qe2 f5 17.
O-O O-O
8 1.26 0.30 14. Qf3 Be6 15. Bd3 O-O 16. O-O Rc8
17. Bf5 Qg5
8-> 1.51 0.30 14. Qf3 Be6 15. Bd3 O-O 16. O-O Rc8
17. Bf5 Qg5
9 2.86 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. Bc2 Ne7
17. Qh5 f5 18. O-O-O Nxd5 19. Rxd5
9-> 4.36 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. Bc2 Ne7
17. Qh5 f5 18. O-O-O Nxd5 19. Rxd5
10 7.43 0.21 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Kf8 17. Bd3 Kg7 18. O-O Rg8 19. Qe4
h6
10-> 18.23 0.21 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Kf8 17. Bd3 Kg7 18. O-O Rg8 19. Qe4
h6
11 24.50 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. O-O Ne7
17. Nxe7+ Qxe7 18. Be4 Rad8 19. Qh5
f5 20. Bd5+ Be6
11-> 36.22 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. O-O Ne7
17. Nxe7+ Qxe7 18. Be4 Rad8 19. Qh5
f5 20. Bd5+ Be6
12 49.98 0.17 14. Qf3 Bd7 15. Bd3 O-O 16. O-O f5
17. Bc2 e4 18. Qh5 Be8 19. Qh3 Bg6
12-> 1:59 0.17 14. Qf3 Bd7 15. Bd3 O-O 16. O-O f5
17. Bc2 e4 18. Qh5 Be8 19. Qh3 Bg6
13 3:08 0.18 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Ke7 17. Bd3 Qd7 18. Qe4 d5 19. Qb4+
Qd6 20. Qxd6+ Kxd6 21. Rd1 Bg4 22.
f3 Bd7
13-> 4:13 0.18 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Ke7 17. Bd3 Qd7 18. Qe4 d5 19. Qb4+
Qd6 20. Qxd6+ Kxd6 21. Rd1 Bg4 22.
f3 Bd7
14 6:20 0.13 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Kf8 17. Be2 d5 18. O-O Be7 19. Nc2
Qc8 20. Qxc8+ Rxc8 21. a4 d4 22. axb5
axb5 23. Rfc1
14 21:40 0.17 14. Nc2 O-O 15. a4 Ne7 16. Bd3 Bxd3
17. Qxd3 Nxd5 18. Qxd5 bxa4 19. Rxa4
Qd7 20. Rh4 Qb5 21. Nb4 Qc5
14-> 26:05 0.17 14. Nc2 O-O 15. a4 Ne7 16. Bd3 Bxd3
17. Qxd3 Nxd5 18. Qxd5 bxa4 19. Rxa4
Qd7 20. Rh4 Qb5 21. Nb4 Qc5
15 44:58 0.15 14. Nc2 O-O 15. Nce3 Be6 16. Bd3 f5
17. Bc2 Rc8 18. Qh5 e4 19. O-O Ne7
20. Rad1 <HT>
15-> 60:42 0.15 14. Nc2 O-O 15. Nce3 Be6 16. Bd3 f5
17. Bc2 Rc8 18. Qh5 e4 19. O-O Ne7
20. Rad1 <HT>
16 142:09 0.37 14. Nc2 O-O 15. Nce3 Bd7 16. Qc2 Ne7
17. Nxe7+ Qxe7 18. Bd3 h6 19. O-O-O
Be6 20. Nf5 Bxf5 21. Bxf5 b4 22. c4
b3 23. Qxb3
16-> 158:47 0.37 14. Nc2 O-O 15. Nce3 Bd7 16. Qc2 Ne7
17. Nxe7+ Qxe7 18. Bd3 h6 19. O-O-O
Be6 20. Nf5 Bxf5 21. Bxf5 b4 22. c4
b3 23. Qxb3
17 250:48 0.28 14. Nc2 O-O 15. Nce3 Be6 16. Bd3 f5
17. Qh5 e4 18. Bc2 Ne7 19. Bb3 Nxd5
<HT>
17-> 328:20 0.28 14. Nc2 O-O 15. Nce3 Be6 16. Bd3 f5
17. Qh5 e4 18. Bc2 Ne7 19. Bb3 Nxd5
<HT>
White(14): quit
execution complete.
KT333 1GB PC2100 cas ?
hash table memory = 192M bytes.
pawn hash table memory = 48M bytes.
clearing hash tables
time surplus 0.00 time limit 30.00 (3:30)
nss depth time score variation (1)
6-> 1.66 0.22 14. Bd3 Be6 15. Be4 Bd7 16. c4 Nd4
7 1.75 0.31 14. Bd3 Be6 15. Be4 Rc8 16. O-O O-O
17. Qf3
7-> 1.86 0.31 14. Bd3 Be6 15. Be4 Rc8 16. O-O O-O
17. Qf3
8 2.98 0.16 14. Bd3 e4 15. Qh5 Bg6 16. Qe2 f5 17.
O-O O-O
8 3.38 0.30 14. Qf3 Be6 15. Bd3 O-O 16. O-O Rc8
17. Bf5 Qg5
8-> 3.69 0.30 14. Qf3 Be6 15. Bd3 O-O 16. O-O Rc8
17. Bf5 Qg5
9 5.04 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. Bc2 Ne7
17. Qh5 f5 18. O-O-O Nxd5 19. Rxd5
9-> 6.50 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. Bc2 Ne7
17. Qh5 f5 18. O-O-O Nxd5 19. Rxd5
10 9.61 0.21 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Kf8 17. Bd3 Kg7 18. O-O Rg8 19. Qe4
h6
10-> 19.81 0.21 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Kf8 17. Bd3 Kg7 18. O-O Rg8 19. Qe4
h6
11 26.14 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. O-O Ne7
17. Nxe7+ Qxe7 18. Be4 Rad8 19. Qh5
f5 20. Bd5+ Be6
11-> 38.08 0.19 14. Qf3 Bd7 15. Bd3 O-O 16. O-O Ne7
17. Nxe7+ Qxe7 18. Be4 Rad8 19. Qh5
f5 20. Bd5+ Be6
12 52.17 0.17 14. Qf3 Bd7 15. Bd3 O-O 16. O-O f5
17. Bc2 e4 18. Qh5 Be8 19. Qh3 Bg6
12-> 2:09 0.17 14. Qf3 Bd7 15. Bd3 O-O 16. O-O f5
17. Bc2 e4 18. Qh5 Be8 19. Qh3 Bg6
13 3:24 0.18 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Ke7 17. Bd3 Qd7 18. Qe4 d5 19. Qb4+
Qd6 20. Qxd6+ Kxd6 21. Rd1 Bg4 22.
f3 Bd7
13-> 4:31 0.18 14. Qf3 Be6 15. Nf6+ Bxf6 16. Qxc6+
Ke7 17. Bd3 Qd7 18. Qe4 d5 19. Qb4+
Qd6 20. Qxd6+ Kxd6 21. Rd1 Bg4 22.
f3 Bd7
>>
>>The interesting thing is that in the testing I have done, the SDRAM provides
>>lower
>>latency, which is pretty important. Not _way_ lower, but 20ns or so...
>
>Probably depends on the chipset. On DDR chipset I have (KT333) gives 2.7gb/s
>bandwidth at 200fsb(400DDR) and 95ns latency while the Nforce2 I have gets
>3.2gb/s and 75ns latency.
>
>I don't know the ins and outs of chess programming so I'm not sure why this is..
>but when I tested ~180ns 1.6gb/s vs 75ns 3.2gb/s I saw no speed difference (with
>a smaller hash table).
>
>Normally I think 100fsb(400DDR) on an nforce2 is ~125ns but as soon as you go
>asynchronous the latency goes to hell heh. So, never run asynchronous (mem/bus)
>on any board if you have the option available.
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.