7 Man Syzygy and SSD

[Nordlandia]

I use game adjudication settings based on the TCEC settings, with minor adjustments. Draw at 35 instead of 40 moves. I tried 7-piece adjudication earlier today in cutechess gui but it doesn't seem to work. 6-men work as long WDL and DTZ are stored in one folder. 7-men adjudication will save alot of time for engine tournaments. As of now i use 6-piece adjudication. I'm considering buying WD 20TB My Book Duo Desktop for 7-piece adjudication.

On a large scale, 7-piece adjudication will definitely save much more time compared to 6-piece. HDD is fine for simple static adjudication.

[MikeB]

I use game adjudication settings based on the TCEC settings, with minor adjustments. Draw at 35 instead of 40 moves. I tried 7-piece adjudication earlier today in cutechess gui but it doesn't seem to work. 6-men work as long WDL and DTZ are stored in one folder. 7-men adjudication will save alot of time for engine tournaments. As of now i use 6-piece adjudication. I'm considering buying WD 20TB My Book Duo Desktop for 7-piece adjudication.

On a large scale, 7-piece adjudication will definitely save much more time compared to 6-piece.

If you are only doing one probe and only when its down to 7 pieces, that should work fine. If the engine starts probing an HHD drive when there are more than 7 pieces on the board, the search will crawl to a virtual stop. Current SF does that today - McCain is modified so that it does one probe at 7 pieces and that's it. Meaning that if you are running engine tournaments with 7 piece syzygy that are on HDD, you have to set to probe limit to 6 ( not use 7 pieces at all) or simply point to the SSD drives which generally will just have 6 and some 7 piece syzygy. I found using the partial set of 7- pieces syzygy unsatisfactory, so I loaded the remaining syzygy files three 8TB HDD drives to have the full set. I ended up repurposing the CD & CDR bays in the 2010 Mac Pro for two of the 8TB Drives.

So altogether , the Mac Pro houses 6 drives - 3 8TB drives and 3SD one 2 TB and two 1 TB - 28TB in total. I use the excess space in the 8TB as part of my backup and recovery plan. In the snapshot below , the 1TB SD drive tabled OWC 960 is dedicated to syzygy. with all 6 piece syzygy and below and some 7 men syzygy. The balance of the 7 pieces syzygy is split between the S1, S2, and S3 8TB drives. I had to give drives ( and folders labels ) short 2 character labels since not all GUIs can handle characters required ( string length) when you normal type labels for multiple drives and paths. One thing you might want to consider is placing the rtbw and rtbz files on separate drives - at a minimum each set of paired files kxxxkxx rtbw and rtbz, when using an HHD drive, should be on separate physical drives - since they are both usually accessed in close proximity to each other timewise. The 2 TB SSD drive is used strictly for backup purposes. That was a personal choice I made since I do use Apple's time machine app and it provides me quick access to saved snapshots - file folders. The sizes indicated in the snapshot below tell me how much of the drive is being used.

[Nordlandia]

MikeB: if you have time please comfirm whether cutechess gui adjourn 7-piece positions.

[MikeB]

MikeB: if you have time please comfirm whether cutechess gui adjourn 7-piece positions.

Sorry , I'm not familiar with cutechess at all. Never used it.

[Nordlandia]

It's not witchcraft. Tools -> Settings -> General -> Syzygy tablebase path:

https://pasteboard.co/ImPn422.png

[Nordlandia]

Ain't the partial 7-piece set as listed sufficient for most general usage Mike.

I'm downloading KBNPvKRB for 16th match game endgame of the Kasparov Karpov 1990 match.

After all these years and extensive investigations, Black can draw with best play from this position. That is the consensus with present engines and 6-men syzygy is draw. I don't know if 7-piece give more insight.

This endgame is won when consulting 7-pieces.

[d]8/8/p2k1p2/1p1p3p/1P1P4/P4PK1/6N1/1b6 b - - 0 48

Maybe KBNPvKRB table give definite answer of this position.

[d]6k1/6p1/4p1n1/3p4/2b3P1/5P1R/1B5K/8 b - - 0 42

[MikeB]

Ain't the partial 7-piece set as listed sufficient for most general usage Mike.

I'm downloading KBNPvKRB for 16th match game endgame of the Kasparov Karpov 1990 match.

After all these years and extensive investigations, Black can draw with best play from this position. That is the consensus with present engines and 6-men syzygy is draw. I don't know if 7-piece give more insight.

This endgame is won when consulting 7-pieces.

[d]8/8/p2k1p2/1p1p3p/1P1P4/P4PK1/6N1/1b6 b - - 0 48

Maybe KBNPvKRB table give definite answer of this position.

[d]6k1/6p1/4p1n1/3p4/2b3P1/5P1R/1B5K/8 b - - 0 42

The first position appears to a win , the second position is less clear, but if black makes a mistake it will lose for sure.

I believe partial 7-piece set as listed sufficient for most general usage. One thing that is very important to check is to make sure you do not have any corrupt database files. Early on some which were released were corrupt. They have been fixed since then, but you should still check. They had used something to md5 checksum, but I believe in the end they ended up using something else. So you should check your all checksums for downloaded files to the master or you may get incorrect results.

[jdart]

MikeB: if you have time please comfirm whether cutechess gui adjourn 7-piece positions.

It doesn't. See:

https://github.com/cutechess/cutechess/issues/499

[Ozymandias]

MikeB: if you have time please comfirm whether cutechess gui adjourn 7-piece positions.

It doesn't. See:

https://github.com/cutechess/cutechess/issues/499

It was already pointed out in another issue:

https://github.com/cutechess/cutechess/issues/461

Unfortunately, cutechess hasn't been maintained for half a year now.

[Nordlandia]

Which of the pawnless tables is good to download ?

The partial syzygy set consists of pawnful tables.

http://tablebase.lichess.ovh/tables/standard/7/