Hard disk performance

Discussion about development of draughts in the time of computer and Internet.
Post Reply
Krzysztof Grzelak
Posts: 886
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Hard disk performance

Post by Krzysztof Grzelak » Mon Jul 06, 2020 15:19

I would like to buy a hard disk Samsung 970 Pro 1TB M.2. I have a question. Is it worth buying it and what will be the disk performance for draughts such as Kingsrow, Damage, Dragon, Scan, Maximus , Sjende Blyn, Moby Dam and other draughts programs. I'm also interested in what the performance will be between this disk and a regular disk SSD. Thank you in advance for any feedback.

Krzysztof.

Joost Buijs
Posts: 325
Joined: Wed May 04, 2016 11:45
Real name: Joost Buijs

Re: Hard disk performance

Post by Joost Buijs » Mon Jul 06, 2020 20:01

Krzysztof Grzelak wrote:
Mon Jul 06, 2020 15:19
I would like to buy a hard disk Samsung 970 Pro 1TB M.2. I have a question. Is it worth buying it and what will be the disk performance for draughts such as Kingsrow, Damage, Dragon, Scan, Maximus , Sjende Blyn, Moby Dam and other draughts programs. I'm also interested in what the performance will be between this disk and a regular disk SSD. Thank you in advance for any feedback.

Krzysztof.
I have a few of these NVME drives in use (although 2TB), the tranfer-rate (random and sequential) is about 5 to 6 times as high as what you get from regular IDE SSDs. They are very useful for endgame databases, for other applications I don't see much difference. You are probably better off with the 970 Evo Plus, it is a lot cheaper and it has a somewhat higher random read-performance too (600.000 IOPS vs 500.000 IOPS).

Joost

BertTuyt
Posts: 1417
Joined: Wed Sep 01, 2004 19:42

Re: Hard disk performance

Post by BertTuyt » Mon Jul 06, 2020 22:42

It also depends if all databases are preloaded in the DB cache at the start.
If this is the case, you might not see the advantage at all (only a faster initialization).
Damage has an option for preload of all 7P DBs (when cache is sufficient).
For Kingsrow with the 8P DB, then a fast SSD has certainly advantages.

Bert

Krzysztof Grzelak
Posts: 886
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Hard disk performance

Post by Krzysztof Grzelak » Tue Jul 07, 2020 00:09

Joost Buijs wrote:
Mon Jul 06, 2020 20:01
I have a few of these NVME drives in use (although 2TB), the tranfer-rate (random and sequential) is about 5 to 6 times as high as what you get from regular IDE SSDs. They are very useful for endgame databases, for other applications I don't see much difference. You are probably better off with the 970 Evo Plus, it is a lot cheaper and it has a somewhat higher random read-performance too (600.000 IOPS vs 500.000 IOPS).
Joost

Thank you for your answer Joost. You disappointed me a little with this news. I wanted to see the difference when looking for a move and calculating a variant, for example during 1 second. Joost, what's your variant NVME on your motherboard?

Krzysztof Grzelak
Posts: 886
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Hard disk performance

Post by Krzysztof Grzelak » Tue Jul 07, 2020 00:15

BertTuyt wrote:
Mon Jul 06, 2020 22:42
It also depends if all databases are preloaded in the DB cache at the start.
If this is the case, you might not see the advantage at all (only a faster initialization).
Damage has an option for preload of all 7P DBs (when cache is sufficient).
For Kingsrow with the 8P DB, then a fast SSD has certainly advantages.

Bert
I would be happier if such a disk calculated more variants and movements than ordinary disks SSD.

Joost Buijs
Posts: 325
Joined: Wed May 04, 2016 11:45
Real name: Joost Buijs

Re: Hard disk performance

Post by Joost Buijs » Tue Jul 07, 2020 10:01

Krzysztof Grzelak wrote:
Tue Jul 07, 2020 00:09
Joost Buijs wrote:
Mon Jul 06, 2020 20:01
I have a few of these NVME drives in use (although 2TB), the tranfer-rate (random and sequential) is about 5 to 6 times as high as what you get from regular IDE SSDs. They are very useful for endgame databases, for other applications I don't see much difference. You are probably better off with the 970 Evo Plus, it is a lot cheaper and it has a somewhat higher random read-performance too (600.000 IOPS vs 500.000 IOPS).
Joost

Thank you for your answer Joost. You disappointed me a little with this news. I wanted to see the difference when looking for a move and calculating a variant, for example during 1 second. Joost, what's your variant NVME on your motherboard?
The version of NVME is 1.3 the bus they are on is PCIe 4.0 x4, but the 970 Pro/Evo does not support that. Like Bert already said, they are useful for large EGDBs, but they wont increase your search speed during the opening/midgame when you're not accessing the EGDB.

If you want to calculate higher nodes per second, you have to switch to a faster CPU. I think you use the TR-1950X, recently I built a machine with a TR-3970X, on this machine with precision boost on (using 8 threads) Kingsrow 1.61 runs around 95 MNPS, so you can compare it with the TR-1950X. My own draughts program (using 32 threads) several times surpassed 500 MNPS on this machine, not that it matters much because with draughts at this level most games will end in a draw anyway.

Krzysztof Grzelak
Posts: 886
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Hard disk performance

Post by Krzysztof Grzelak » Tue Jul 07, 2020 11:49

Thank you for your answer Joost. It doesn't make sense to buy such a disk. And how does it relate to chess Joost.

Joost Buijs
Posts: 325
Joined: Wed May 04, 2016 11:45
Real name: Joost Buijs

Re: Hard disk performance

Post by Joost Buijs » Tue Jul 07, 2020 12:16

Krzysztof Grzelak wrote:
Tue Jul 07, 2020 11:49
Thank you for your answer Joost. It doesn't make sense to buy such a disk. And how does it relate to chess Joost.
For chess it's the same, it only helps with EGDBs. I have a few of these 2TB drives with Nalimov 6P (1.2 TB), syzygy 6P and Kingsrows EGDBs on them. In practice I only use Nalimov because it is depth to mate. EGDBs don't add much in strength, nowadays chess engines are so strong that they will find the right move in 99% of the cases without EGDB anyway. I guess it's the same with Draughts.

Krzysztof Grzelak
Posts: 886
Joined: Thu Jun 20, 2013 17:16
Real name: Krzysztof Grzelak

Re: Hard disk performance

Post by Krzysztof Grzelak » Tue Jul 07, 2020 13:07

Thank you Joost. I have a huge request to you Joost. You wouldn't have made the Nalimov (6P) database available online.

Joost Buijs
Posts: 325
Joined: Wed May 04, 2016 11:45
Real name: Joost Buijs

Re: Hard disk performance

Post by Joost Buijs » Tue Jul 07, 2020 20:14

Krzysztof Grzelak wrote:
Tue Jul 07, 2020 13:07
Thank you Joost. I have a huge request to you Joost. You wouldn't have made the Nalimov (6P) database available online.
Krzysztof, I don't think my provider likes it when I upload 1.2 terabyte. I'm willing to put Nalimov 6P on a 1.5 TB portable hard-drive, but than you have to sent me one, or I can buy one for you locally, a WD Elements 1.5 TB costs 61 euro's at Amazon. If you are interested you may send me a PM.

As an alternative you can download them from here: http://kirill-kryukov.com/chess/tablebases-online/ Of course this will take ages, even with a very fast internet connection.

Joost

Post Reply