Hiarcs TB test - HD v Flash

This forum is for general discussions and questions, including Collectors Corner and anything to do with Computer chess.

Moderators: Harvey Williamson, Steve B, Watchman

Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
Shaun Brewer
Member
Posts: 185
Joined: Tue Aug 28, 2007 10:12 pm
Location: Brighton

Hiarcs TB test - HD v Flash

Post by Shaun Brewer »

You might find this post interesting

TB tests - HD v Flash including Hiarcs

Shaun
Shaun Brewer
Member
Posts: 185
Joined: Tue Aug 28, 2007 10:12 pm
Location: Brighton

Post by Shaun Brewer »

I could not post the graphs here hence the link...

Shaun
User avatar
AMD64inside
Member
Posts: 20
Joined: Tue Jul 31, 2007 7:31 pm

Post by AMD64inside »

Hi Shaun

I have also found big speed increases in ply reached and Kn/s when testing my HDD v my 8GB Corsair Voyager GT. (When using complete 5 man)

On average a 40% higher Kn/s rate acheived when using the flash compared to HDD. (In test postitions with 6 or 7 remaining men)

Regards

Darrell
"The game itself is nothing, the playing of it is everything"
User avatar
AMD64inside
Member
Posts: 20
Joined: Tue Jul 31, 2007 7:31 pm

Post by AMD64inside »

I just carried out a quick test on the position you gave in your link.

[fen]8/3k4/p6r/5R2/PP5p/7K/8/8[/fen]

8/3k4/p6r/5R2/PP5p/7K/8/8 w - - 0 1

Using Fritz 10 with 256Mb hash and 32Mb tablebase hash and using infinite analysis mode for a 3.00 min think.

Here is the analysis outputs, first is with complete 5 man TB on my flash drive, second is with the complete 5 man TB on my Seagate 750Gb drive.

(PC is an AMD dual core 4200+ with 1GB Ram)

New game, 3 mins analysis FLASH Drive
8/3k4/p6r/5R2/PP5p/7K/8/8 w - - 0 1

Analysis by Fritz 10:

1.Ra5 Kc7 2.Rd5 Rd6 3.Rg5 Rh6 4.a5 Rh8
= (0.16) Depth: 9/20 00:00:00 120kN, tb=12
1.Rc5
= (0.16) Depth: 9/20 00:00:00 128kN, tb=13
1.Rc5 Rc6 2.Rd5+ Rd6 3.Rh5 Rd3+ 4.Kh2 Rd2+ 5.Kg1 Rd4 6.Rh7+ Kd6 7.Rh6+ Kd7 8.b5 Rxa4
= (0.08) Depth: 10/22 00:00:00 244kN, tb=43
1.a5
= (0.08) Depth: 10/22 00:00:00 280kN, tb=52
1.a5 Kc6 2.Rf6+ Rxf6
= (0.09) Depth: 10/22 00:00:00 298kN, tb=55
1.Rf4
= (0.10) Depth: 10/22 00:00:00 307kN, tb=59
1.b5
= (0.10) Depth: 10/22 00:00:00 439kN, tb=103
1.b5 axb5 2.Rxb5 Kc6 3.Kg4 h3
= (0.11) Depth: 10/22 00:00:00 470kN, tb=126
1.Rg5
= (0.11) Depth: 10/22 00:00:00 594kN, tb=151
1.Rf1
= (0.11) Depth: 10/22 00:00:00 671kN, tb=155
1.Rf2
= (0.12) Depth: 10/22 00:00:00 722kN, tb=156
1.Rf2 Kc7 2.Rd2 Rd6
= (0.13) Depth: 10/22 00:00:00 753kN, tb=158
1.Rf2 Kc7 2.Rf3 Kd7 3.Rd3+ Rd6 4.Rb3 Rh6 5.Rb2 Kc7 6.Re2 Kb6
= (0.13) Depth: 11/22 00:00:00 913kN, tb=162
1.Rf4
= (0.13) Depth: 11/22 00:00:00 967kN, tb=202
1.Rf4 Kd6 2.Rd4+ Kc7 3.Rc4+ Kb6 4.Rc5 Rc6 5.a5+ Kb7 6.Rg5 Rc4 7.Rg7+ Kc6 8.Rg6+ Kb5 9.Rb6+ Ka4
= (0.16) Depth: 11/22 00:00:00 1046kN, tb=212
1.Rf4 Kd6 2.Rd4+ Kc7 3.Rc4+ Kb6 4.Rc5 Rc6 5.a5+ Kb7 6.Rg5 Rc4 7.Rg7+ Kc6 8.Rg6+ Kb5 9.Rb6+ Ka4
= (0.16) Depth: 12/26 00:00:00 1268kN, tb=324
1.Rf4 Kd6 2.Rd4+ Ke5 3.Rd2 Rb6 4.Rb2 Kd5 5.a5 Rb7 6.Kxh4 Kc4 7.Kg5 Rxb4
= (0.04) Depth: 13/26 00:00:01 1862kN, tb=693
1.Rg5
= (0.05) Depth: 13/26 00:00:01 2405kN, tb=976
1.Rg5 Kc7 2.Re5 Kd7 3.Re3 Kc7 4.Rd3 Rc6 5.Rd4 Rh6 6.Rc4+ Kb6 7.Rg4 Kc7 8.Rg5 Kd7 9.b5
= (0.14) Depth: 13/26 00:00:02 2980kN, tb=1128
1.Rg5 Kc7 2.Re5 Kd7 3.Re3 Kc7 4.Rf3 Kd7
= (0.13) Depth: 14/28 00:00:03 4573kN, tb=2061
1.Rg5 Kc7 2.Re5 Kd6 3.Rc5
= (0.12) Depth: 15/37 00:00:07 11171kN, tb=5547
1.Rg5 Kd6 2.Ra5 Kd7 3.b5 axb5 4.Ra7+ Kc8 5.axb5 Rb6 6.Ra5 Kb7 7.Kxh4 Re6 8.Ra4 Re1
= (0.00) Depth: 16/43 00:00:14 22935kN, tb=13486
1.Rf4
= (0.01) Depth: 16/43 00:00:15 24012kN, tb=14150
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Rh8
= (0.12) Depth: 16/43 00:00:16 25307kN, tb=14813
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4
= (0.00) Depth: 17/40 00:00:18 29268kN, tb=20100
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6
= (0.00) Depth: 18/32 00:00:32 50196kN, tb=40552
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5
= (0.00) Depth: 19/35 00:00:39 59929kN, tb=59144
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5 Kd5
= (0.00) Depth: 20/44 00:00:53 79253kN, tb=94889
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5 Kd5 10.Kh6
= (0.00) Depth: 21/53 00:01:19 118mN, tb=164875
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.Rg5 Kc6 5.Rc5+ Kd7 6.Rc4 Kd6 7.b5
= (0.00) Depth: 22/50 00:02:07 192mN, tb=298580

(, 24.10.2007)


New game, 3 Mins Analysis HDD
8/3k4/p6r/5R2/PP5p/7K/8/8 w - - 0 1

Analysis by Fritz 10:

1.Ra5 Kc7 2.Rd5 Rd6 3.Rg5 Rh6 4.a5 Rh8
= (0.16) Depth: 9/20 00:00:00 120kN, tb=12
1.Rc5
= (0.16) Depth: 9/20 00:00:00 128kN, tb=13
1.Rc5 Rc6 2.Rd5+ Rd6 3.Rh5 Rd3+ 4.Kh2 Rd2+ 5.Kg1 Rd4 6.Rh7+ Kd6 7.Rh6+ Kd7 8.b5 Rxa4
= (0.08) Depth: 10/22 00:00:00 244kN, tb=43
1.a5
= (0.08) Depth: 10/22 00:00:00 280kN, tb=52
1.a5 Kc6 2.Rf6+ Rxf6
= (0.09) Depth: 10/22 00:00:00 298kN, tb=55
1.Rf4
= (0.10) Depth: 10/22 00:00:00 307kN, tb=59
1.b5
= (0.10) Depth: 10/22 00:00:00 439kN, tb=103
1.b5 axb5 2.Rxb5 Kc6 3.Kg4 h3
= (0.11) Depth: 10/22 00:00:00 470kN, tb=126
1.Rg5
= (0.11) Depth: 10/22 00:00:00 594kN, tb=151
1.Rf1
= (0.11) Depth: 10/22 00:00:00 671kN, tb=155
1.Rf2
= (0.12) Depth: 10/22 00:00:01 722kN, tb=156
1.Rf2 Kc7 2.Rd2 Rd6
= (0.13) Depth: 10/22 00:00:01 753kN, tb=158
1.Rf2 Kc7 2.Rf3 Kd7 3.Rd3+ Rd6 4.Rb3 Rh6 5.Rb2 Kc7 6.Re2 Kb6
= (0.13) Depth: 11/22 00:00:01 913kN, tb=162
1.Rf4
= (0.13) Depth: 11/22 00:00:01 967kN, tb=202
1.Rf4 Kd6 2.Rd4+ Kc7 3.Rc4+ Kb6 4.Rc5 Rc6 5.a5+ Kb7 6.Rg5 Rc4 7.Rg7+ Kc6 8.Rg6+ Kb5 9.Rb6+ Ka4
= (0.16) Depth: 11/22 00:00:01 1046kN, tb=212
1.Rf4 Kd6 2.Rd4+ Kc7 3.Rc4+ Kb6 4.Rc5 Rc6 5.a5+ Kb7 6.Rg5 Rc4 7.Rg7+ Kc6 8.Rg6+ Kb5 9.Rb6+ Ka4
= (0.16) Depth: 12/26 00:00:01 1268kN, tb=324
1.Rf4 Kd6 2.Rd4+ Ke5 3.Rd2 Rb6 4.Rb2 Kd5 5.a5 Rb7 6.Kxh4 Kc4 7.Kg5 Rxb4
= (0.04) Depth: 13/26 00:00:02 1862kN, tb=693
1.Rg5
= (0.05) Depth: 13/26 00:00:03 2405kN, tb=976
1.Rg5 Kc7 2.Re5 Kd7 3.Re3 Kc7 4.Rd3 Rc6 5.Rd4 Rh6 6.Rc4+ Kb6 7.Rg4 Kc7 8.Rg5 Kd7 9.b5
= (0.14) Depth: 13/26 00:00:04 2980kN, tb=1128
1.Rg5 Kc7 2.Re5 Kd7 3.Re3 Kc7 4.Rf3 Kd7
= (0.13) Depth: 14/28 00:00:07 4573kN, tb=2061
1.Rg5 Kc7 2.Re5 Kd6 3.Rc5
= (0.12) Depth: 15/37 00:00:16 11171kN, tb=5547
1.Rg5 Kd6 2.Ra5 Kd7 3.b5 axb5 4.Ra7+ Kc8 5.axb5 Rb6 6.Ra5 Kb7 7.Kxh4 Re6 8.Ra4 Re1
= (0.00) Depth: 16/43 00:00:32 22935kN, tb=13486
1.Rf4
= (0.01) Depth: 16/43 00:00:33 24012kN, tb=14150
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Rh8
= (0.12) Depth: 16/43 00:00:34 25307kN, tb=14813
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4
= (0.00) Depth: 17/40 00:00:41 29268kN, tb=20100
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6
= (0.00) Depth: 18/32 00:01:06 50196kN, tb=40552
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5
= (0.00) Depth: 19/35 00:01:21 59929kN, tb=59144
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5 Kd5
= (0.00) Depth: 20/44 00:01:48 79253kN, tb=94889
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5 Kd5 10.Kh6
= (0.00) Depth: 21/53 00:02:32 118mN, tb=164875
(, 24.10.2007)

As you can see the flash setup completed the 21/53 depth search in 1:19 mins compared to the HDD of 2:32 mins.

I noted on watching the Kn/s the flash setup was allowing Fritz to run at 1300 - 1400 Kn/s compared to the HDD setup of around 700 Kn/s.

Regards Darrell
"The game itself is nothing, the playing of it is everything"
KarmaZen&Oliver
Member
Posts: 66
Joined: Fri Aug 31, 2007 9:48 am
Contact:

Post by KarmaZen&Oliver »

AMD64inside wrote:I just carried out a quick test on the position you gave in your link.

[fen]8/3k4/p6r/5R2/PP5p/7K/8/8[/fen]

8/3k4/p6r/5R2/PP5p/7K/8/8 w - - 0 1

Using Fritz 10 with 256Mb hash and 32Mb tablebase hash and using infinite analysis mode for a 3.00 min think.

Here is the analysis outputs, first is with complete 5 man TB on my flash drive, second is with the complete 5 man TB on my Seagate 750Gb drive.

(PC is an AMD dual core 4200+ with 1GB Ram)

New game, 3 mins analysis FLASH Drive
8/3k4/p6r/5R2/PP5p/7K/8/8 w - - 0 1

Analysis by Fritz 10:

1.Ra5 Kc7 2.Rd5 Rd6 3.Rg5 Rh6 4.a5 Rh8
= (0.16) Depth: 9/20 00:00:00 120kN, tb=12
1.Rc5
= (0.16) Depth: 9/20 00:00:00 128kN, tb=13
1.Rc5 Rc6 2.Rd5+ Rd6 3.Rh5 Rd3+ 4.Kh2 Rd2+ 5.Kg1 Rd4 6.Rh7+ Kd6 7.Rh6+ Kd7 8.b5 Rxa4
= (0.08) Depth: 10/22 00:00:00 244kN, tb=43
1.a5
= (0.08) Depth: 10/22 00:00:00 280kN, tb=52
1.a5 Kc6 2.Rf6+ Rxf6
= (0.09) Depth: 10/22 00:00:00 298kN, tb=55
1.Rf4
= (0.10) Depth: 10/22 00:00:00 307kN, tb=59
1.b5
= (0.10) Depth: 10/22 00:00:00 439kN, tb=103
1.b5 axb5 2.Rxb5 Kc6 3.Kg4 h3
= (0.11) Depth: 10/22 00:00:00 470kN, tb=126
1.Rg5
= (0.11) Depth: 10/22 00:00:00 594kN, tb=151
1.Rf1
= (0.11) Depth: 10/22 00:00:00 671kN, tb=155
1.Rf2
= (0.12) Depth: 10/22 00:00:00 722kN, tb=156
1.Rf2 Kc7 2.Rd2 Rd6
= (0.13) Depth: 10/22 00:00:00 753kN, tb=158
1.Rf2 Kc7 2.Rf3 Kd7 3.Rd3+ Rd6 4.Rb3 Rh6 5.Rb2 Kc7 6.Re2 Kb6
= (0.13) Depth: 11/22 00:00:00 913kN, tb=162
1.Rf4
= (0.13) Depth: 11/22 00:00:00 967kN, tb=202
1.Rf4 Kd6 2.Rd4+ Kc7 3.Rc4+ Kb6 4.Rc5 Rc6 5.a5+ Kb7 6.Rg5 Rc4 7.Rg7+ Kc6 8.Rg6+ Kb5 9.Rb6+ Ka4
= (0.16) Depth: 11/22 00:00:00 1046kN, tb=212
1.Rf4 Kd6 2.Rd4+ Kc7 3.Rc4+ Kb6 4.Rc5 Rc6 5.a5+ Kb7 6.Rg5 Rc4 7.Rg7+ Kc6 8.Rg6+ Kb5 9.Rb6+ Ka4
= (0.16) Depth: 12/26 00:00:00 1268kN, tb=324
1.Rf4 Kd6 2.Rd4+ Ke5 3.Rd2 Rb6 4.Rb2 Kd5 5.a5 Rb7 6.Kxh4 Kc4 7.Kg5 Rxb4
= (0.04) Depth: 13/26 00:00:01 1862kN, tb=693
1.Rg5
= (0.05) Depth: 13/26 00:00:01 2405kN, tb=976
1.Rg5 Kc7 2.Re5 Kd7 3.Re3 Kc7 4.Rd3 Rc6 5.Rd4 Rh6 6.Rc4+ Kb6 7.Rg4 Kc7 8.Rg5 Kd7 9.b5
= (0.14) Depth: 13/26 00:00:02 2980kN, tb=1128
1.Rg5 Kc7 2.Re5 Kd7 3.Re3 Kc7 4.Rf3 Kd7
= (0.13) Depth: 14/28 00:00:03 4573kN, tb=2061
1.Rg5 Kc7 2.Re5 Kd6 3.Rc5
= (0.12) Depth: 15/37 00:00:07 11171kN, tb=5547
1.Rg5 Kd6 2.Ra5 Kd7 3.b5 axb5 4.Ra7+ Kc8 5.axb5 Rb6 6.Ra5 Kb7 7.Kxh4 Re6 8.Ra4 Re1
= (0.00) Depth: 16/43 00:00:14 22935kN, tb=13486
1.Rf4
= (0.01) Depth: 16/43 00:00:15 24012kN, tb=14150
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Rh8
= (0.12) Depth: 16/43 00:00:16 25307kN, tb=14813
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4
= (0.00) Depth: 17/40 00:00:18 29268kN, tb=20100
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6
= (0.00) Depth: 18/32 00:00:32 50196kN, tb=40552
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5
= (0.00) Depth: 19/35 00:00:39 59929kN, tb=59144
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5 Kd5
= (0.00) Depth: 20/44 00:00:53 79253kN, tb=94889
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5 Kd5 10.Kh6
= (0.00) Depth: 21/53 00:01:19 118mN, tb=164875
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.Rg5 Kc6 5.Rc5+ Kd7 6.Rc4 Kd6 7.b5
= (0.00) Depth: 22/50 00:02:07 192mN, tb=298580

(, 24.10.2007)


New game, 3 Mins Analysis HDD
8/3k4/p6r/5R2/PP5p/7K/8/8 w - - 0 1

Analysis by Fritz 10:

1.Ra5 Kc7 2.Rd5 Rd6 3.Rg5 Rh6 4.a5 Rh8
= (0.16) Depth: 9/20 00:00:00 120kN, tb=12
1.Rc5
= (0.16) Depth: 9/20 00:00:00 128kN, tb=13
1.Rc5 Rc6 2.Rd5+ Rd6 3.Rh5 Rd3+ 4.Kh2 Rd2+ 5.Kg1 Rd4 6.Rh7+ Kd6 7.Rh6+ Kd7 8.b5 Rxa4
= (0.08) Depth: 10/22 00:00:00 244kN, tb=43
1.a5
= (0.08) Depth: 10/22 00:00:00 280kN, tb=52
1.a5 Kc6 2.Rf6+ Rxf6
= (0.09) Depth: 10/22 00:00:00 298kN, tb=55
1.Rf4
= (0.10) Depth: 10/22 00:00:00 307kN, tb=59
1.b5
= (0.10) Depth: 10/22 00:00:00 439kN, tb=103
1.b5 axb5 2.Rxb5 Kc6 3.Kg4 h3
= (0.11) Depth: 10/22 00:00:00 470kN, tb=126
1.Rg5
= (0.11) Depth: 10/22 00:00:00 594kN, tb=151
1.Rf1
= (0.11) Depth: 10/22 00:00:00 671kN, tb=155
1.Rf2
= (0.12) Depth: 10/22 00:00:01 722kN, tb=156
1.Rf2 Kc7 2.Rd2 Rd6
= (0.13) Depth: 10/22 00:00:01 753kN, tb=158
1.Rf2 Kc7 2.Rf3 Kd7 3.Rd3+ Rd6 4.Rb3 Rh6 5.Rb2 Kc7 6.Re2 Kb6
= (0.13) Depth: 11/22 00:00:01 913kN, tb=162
1.Rf4
= (0.13) Depth: 11/22 00:00:01 967kN, tb=202
1.Rf4 Kd6 2.Rd4+ Kc7 3.Rc4+ Kb6 4.Rc5 Rc6 5.a5+ Kb7 6.Rg5 Rc4 7.Rg7+ Kc6 8.Rg6+ Kb5 9.Rb6+ Ka4
= (0.16) Depth: 11/22 00:00:01 1046kN, tb=212
1.Rf4 Kd6 2.Rd4+ Kc7 3.Rc4+ Kb6 4.Rc5 Rc6 5.a5+ Kb7 6.Rg5 Rc4 7.Rg7+ Kc6 8.Rg6+ Kb5 9.Rb6+ Ka4
= (0.16) Depth: 12/26 00:00:01 1268kN, tb=324
1.Rf4 Kd6 2.Rd4+ Ke5 3.Rd2 Rb6 4.Rb2 Kd5 5.a5 Rb7 6.Kxh4 Kc4 7.Kg5 Rxb4
= (0.04) Depth: 13/26 00:00:02 1862kN, tb=693
1.Rg5
= (0.05) Depth: 13/26 00:00:03 2405kN, tb=976
1.Rg5 Kc7 2.Re5 Kd7 3.Re3 Kc7 4.Rd3 Rc6 5.Rd4 Rh6 6.Rc4+ Kb6 7.Rg4 Kc7 8.Rg5 Kd7 9.b5
= (0.14) Depth: 13/26 00:00:04 2980kN, tb=1128
1.Rg5 Kc7 2.Re5 Kd7 3.Re3 Kc7 4.Rf3 Kd7
= (0.13) Depth: 14/28 00:00:07 4573kN, tb=2061
1.Rg5 Kc7 2.Re5 Kd6 3.Rc5
= (0.12) Depth: 15/37 00:00:16 11171kN, tb=5547
1.Rg5 Kd6 2.Ra5 Kd7 3.b5 axb5 4.Ra7+ Kc8 5.axb5 Rb6 6.Ra5 Kb7 7.Kxh4 Re6 8.Ra4 Re1
= (0.00) Depth: 16/43 00:00:32 22935kN, tb=13486
1.Rf4
= (0.01) Depth: 16/43 00:00:33 24012kN, tb=14150
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Rh8
= (0.12) Depth: 16/43 00:00:34 25307kN, tb=14813
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4
= (0.00) Depth: 17/40 00:00:41 29268kN, tb=20100
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6
= (0.00) Depth: 18/32 00:01:06 50196kN, tb=40552
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5
= (0.00) Depth: 19/35 00:01:21 59929kN, tb=59144
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5 Kd5
= (0.00) Depth: 20/44 00:01:48 79253kN, tb=94889
1.Rf4 Kd6 2.Rc4 Kd5 3.Rc5+ Kd6 4.a5 Re6 5.Rc4 Re3+ 6.Kxh4 Kd5 7.Rg4 Re4 8.Rxe4 Kxe4 9.Kh5 Kd5 10.Kh6
= (0.00) Depth: 21/53 00:02:32 118mN, tb=164875
(, 24.10.2007)

As you can see the flash setup completed the 21/53 depth search in 1:19 mins compared to the HDD of 2:32 mins.

I noted on watching the Kn/s the flash setup was allowing Fritz to run at 1300 - 1400 Kn/s compared to the HDD setup of around 700 Kn/s.

Regards Darrell
OK. flahs are better for read. but for write not are very good, are poor.

I like more if you test FLASH verus Western DIGITAL RAPTOR 150 or 74 GB, that are HDD 10.000 rmp. and are very fast to read files IO...

I have raptors, but not have Flash

you have flash, but not have RAPTORs... ;-)
User avatar
Harvey Williamson
Site Admin
Posts: 6079
Joined: Sun Jul 29, 2007 6:57 am
Location: Media City, UK
Contact:

Post by Harvey Williamson »

Hi,

These tests are interesting. I have just been experimenting with changing tablebase depth for Hiarcs from the default 2. I have tried 3,4 & 5 has anyone else experimented with this?

Best Wishes,

Harvey
Shaun Brewer
Member
Posts: 185
Joined: Tue Aug 28, 2007 10:12 pm
Location: Brighton

Post by Shaun Brewer »

AMD64inside wrote:Hi Shaun

I have also found big speed increases in ply reached and Kn/s when testing my HDD v my 8GB Corsair Voyager GT. (When using complete 5 man)

On average a 40% higher Kn/s rate acheived when using the flash compared to HDD. (In test postitions with 6 or 7 remaining men)

Regards

Darrell
Hi Darrell,

thanks for your results - the more the better!!!

Shaun
Shaun Brewer
Member
Posts: 185
Joined: Tue Aug 28, 2007 10:12 pm
Location: Brighton

Post by Shaun Brewer »

KarmaZen&Oliver wrote:
OK. flahs are better for read. but for write not are very good, are poor.

I like more if you test FLASH verus Western DIGITAL RAPTOR 150 or 74 GB, that are HDD 10.000 rmp. and are very fast to read files IO...

I have raptors, but not have Flash

you have flash, but not have RAPTORs... ;-)
Sorry no Raptors here :(, although I suspect flash would still outperform a single Raptor, and for TB access write speed is not an issue...

I have a new flash drive in the post - once that arrives I will run some additional tests.

Now if someone had both a fast Raid array and flash we could do with some more results ;)

Shaun
Shaun Brewer
Member
Posts: 185
Joined: Tue Aug 28, 2007 10:12 pm
Location: Brighton

Post by Shaun Brewer »

Harvey Williamson wrote:Hi,

These tests are interesting. I have just been experimenting with changing tablebase depth for Hiarcs from the default 2. I have tried 3,4 & 5 has anyone else experimented with this?

Best Wishes,

Harvey
Hi Harvey,

Hiarcs seems already very aggressive with its TB access as can be seen from the graph.

For 5men TBs I think flash is a must but it must have fast random read.

For 6 men a combination of flash and fast raid would probably be optimal.

One thing I am glad is that nodes/sec holds even though time to ply does not when running 2/4 threads.

I will also run the node counts without TB for an additional comparison, my guess is that slow TB access could impact performance more than not having the TBs - however a lot of games would be required to prove that.

Once I have any more info I will post here.

Shaun

P.S. Good luck for Leiden
Shaun Brewer
Member
Posts: 185
Joined: Tue Aug 28, 2007 10:12 pm
Location: Brighton

Post by Shaun Brewer »

More results/graphs and raw data

here

Shaun
bob
Member
Posts: 29
Joined: Sat Oct 13, 2007 1:24 am

Re: Hiarcs TB test - HD v Flash

Post by bob »

Shaun Brewer wrote:You might find this post interesting

TB tests - HD v Flash including Hiarcs

Shaun
It isn't so interesting since it is really biased toward the flash drive. 7200 rpm drive (non-SCSI) are not known for high-performance.

I run 15K U320 SCSI drives in a raid-0 (striping) mode and I have not found a flash drive that can keep up yet...
Shaun Brewer
Member
Posts: 185
Joined: Tue Aug 28, 2007 10:12 pm
Location: Brighton

Re: Hiarcs TB test - HD v Flash

Post by Shaun Brewer »

bob wrote:
Shaun Brewer wrote:You might find this post interesting

TB tests - HD v Flash including Hiarcs

Shaun
It isn't so interesting since it is really biased toward the flash drive. 7200 rpm drive (non-SCSI) are not known for high-performance.

I run 15K U320 SCSI drives in a raid-0 (striping) mode and I have not found a flash drive that can keep up yet...
Please please post some results - it would be fantastic if people could add additional raw data from a variety of hardware then I can combine the results...

The whole purpose of my post is to share results and the more info the better.

Shaun
User avatar
Harvey Williamson
Site Admin
Posts: 6079
Joined: Sun Jul 29, 2007 6:57 am
Location: Media City, UK
Contact:

Post by Harvey Williamson »

I have all my 4 and 5 man tbs on 1 of these:

http://www.overclockers.co.uk/showprodu ... =MY-131-CS

Image
Shaun Brewer
Member
Posts: 185
Joined: Tue Aug 28, 2007 10:12 pm
Location: Brighton

Post by Shaun Brewer »

Hi Harvey,

A nice flash drive - it may be faster - the price put me off - I think this drive is similar for TB access but write is slower and seems a little flimsy but should be okay for TBs. I am going to buy a second one soon.

Am I correct in thinking you have raid on a couple of your machines???

Any chance you could run a couple of test with the position I used (5 men only) with raid v flash - I can then add your data...

Shaun
User avatar
Harvey Williamson
Site Admin
Posts: 6079
Joined: Sun Jul 29, 2007 6:57 am
Location: Media City, UK
Contact:

Post by Harvey Williamson »

Shaun Brewer wrote:Hi Harvey,

A nice flash drive - it may be faster - the price put me off - I think this drive is similar for TB access but write is slower and seems a little flimsy but should be okay for TBs. I am going to buy a second one soon.

Am I correct in thinking you have raid on a couple of your machines???

Any chance you could run a couple of test with the position I used (5 men only) with raid v flash - I can then add your data...

Shaun
Hi Shaun,

I do not use Raid on any of my machines so am unable to test this.

Best Wishes,

Harvey
Post Reply