5 men Syzygy on USB 3.0 Flash Drive

[ernest]

With a lowly flash drive, having such speeds:

Is the access time of that flash drive, which is really what counts, just as lowly (meaning of the order of 1 ms or more) ?

0.59 ms

So on this account, that flash drive is not bad at all !
(of course it ain't SSD...)

[zullil]

The NPS is also higher by 7%, explaining the Elo difference. I was under impression that 5 men Syzygy on HDD would perform almost identically to flash or even SSD, as all the files are cached into RAM.

At very short time control it could make a difference. That should disappear if you type

Code: Select all

type *.rtbw >nul

before the test (obviously from the directory with the files that are going to be accessed by that test).

If I am not doing something wrong, still seems 6% faster on flash.

You entered this command twice? Once in the directory that contains the .rtbw files on the flash drive, and then again in the directory that contains the same files on your hard drive?

[Laskos]

The NPS is also higher by 7%, explaining the Elo difference. I was under impression that 5 men Syzygy on HDD would perform almost identically to flash or even SSD, as all the files are cached into RAM.

At very short time control it could make a difference. That should disappear if you type

Code: Select all

type *.rtbw >nul

before the test (obviously from the directory with the files that are going to be accessed by that test).

If I am not doing something wrong, still seems 6% faster on flash.

You entered this command twice? Once in the directory that contains the .rtbw files on the flash drive, and then again in the directory that contains the same files on your hard drive?

No, just on hard drive.

[zullil]

The NPS is also higher by 7%, explaining the Elo difference. I was under impression that 5 men Syzygy on HDD would perform almost identically to flash or even SSD, as all the files are cached into RAM.

At very short time control it could make a difference. That should disappear if you type

Code: Select all

type *.rtbw >nul

before the test (obviously from the directory with the files that are going to be accessed by that test).

If I am not doing something wrong, still seems 6% faster on flash.

You entered this command twice? Once in the directory that contains the .rtbw files on the flash drive, and then again in the directory that contains the same files on your hard drive?

No, just on hard drive.

Weird. Then shouldn't any tablebase probe by the HDD-Stockfish read its info directly from RAM? Why wouldn't this be faster than reading from the flash drive? Then again, I have no idea how Windows handles stuff like this.

[Laskos]

The NPS is also higher by 7%, explaining the Elo difference. I was under impression that 5 men Syzygy on HDD would perform almost identically to flash or even SSD, as all the files are cached into RAM.

At very short time control it could make a difference. That should disappear if you type

Code: Select all

type *.rtbw >nul

before the test (obviously from the directory with the files that are going to be accessed by that test).

If I am not doing something wrong, still seems 6% faster on flash.

You entered this command twice? Once in the directory that contains the .rtbw files on the flash drive, and then again in the directory that contains the same files on your hard drive?

No, just on hard drive.

Weird. Then shouldn't any tablebase probe by the HDD-Stockfish read its info directly from RAM? Why wouldn't this be faster than reading from the flash drive? Then again, I have no idea how Windows handles stuff like this.

Yes, I don't understand either.

[syzygy]

The NPS is also higher by 7%, explaining the Elo difference. I was under impression that 5 men Syzygy on HDD would perform almost identically to flash or even SSD, as all the files are cached into RAM.

At very short time control it could make a difference. That should disappear if you type

Code: Select all

type *.rtbw >nul

before the test (obviously from the directory with the files that are going to be accessed by that test).

If I am not doing something wrong, still seems 6% faster on flash.

You entered this command twice? Once in the directory that contains the .rtbw files on the flash drive, and then again in the directory that contains the same files on your hard drive?

No, just on hard drive.

Weird. Then shouldn't any tablebase probe by the HDD-Stockfish read its info directly from RAM? Why wouldn't this be faster than reading from the flash drive? Then again, I have no idea how Windows handles stuff like this.

It should, but I also don't know how Windows handles this. There should be no HDD accesses.

On Linux this trick works perfectly fine.

[Ozymandias]

Another option is to just load 5-men into a RAM drive, and point one of the engines to that path. The other one should read from the flash drive, and this time be slower.

[Laskos]

The NPS is also higher by 7%, explaining the Elo difference. I was under impression that 5 men Syzygy on HDD would perform almost identically to flash or even SSD, as all the files are cached into RAM.

At very short time control it could make a difference. That should disappear if you type

I checked time to depth 25 on some 300 endgame positions, single thread, fixed hash, average used time 7s per position, not too short, and the effect all but disappeared:

Flash: 34:21
HDD: 34:30

So it only appears at ultrafast. I am unable to figure out what happens with your tip to force *.rtbw into RAM, anyway, if done correctly or with RAM Disk, an improvement comes by definition, so I am not too curious to see the result.

[Norm Pollock]

When using tablebases. particularly large files, defragmenting them is important.

[syzygy]

When using tablebases. particularly large files, defragmenting them is important.

I wonder in what shape bittorrent typically leaves large downloaded files.