New asmFish released

[Dann Corbit]

He wants to run it on a Mac, not a Windows box or on Linux.

For CFish, you just compile it.
For the assembly programs you will have to rewrite it.
On my machine, Cfish is a lot faster than SF.
I guess that will also be true on a Mac.
The speed AsmFish and Cerebellum on a Mac is zero, since they do not run on it.

[Milos]

He wants to run it on a Mac, not a Windows box or on Linux.

For CFish, you just compile it.
For the assembly programs you will have to rewrite it.
On my machine, Cfish is a lot faster than SF.
I guess that will also be true on a Mac.
The speed AsmFish and Cerebellum on a Mac is zero, since they do not run on it.

Brainfish is pure SF why it doesn't run on Mac, you just need to compile it and as I said compiling it is as easy as compiling SF, it compiles nicely on anything. Cfish only compiles cleanly under MinGW, under different GCC flavour of Mac it's quite some work to make it.
On my machine no matter how I compile Cfish using latest MinGW with 6.2.0 GCC it is still slower than Brainfish numa and just a little bit faster 1-2% than official SF compile.

[Milos]

Ehm, the most invasive change of the "unix version" of brainfish is the insertion of the "author"'s own name, unless you run it on a NUMA machine (for which it seems to use the patch prepared by Mohammed based on - I think - Texel's code).

NUMA patch provides noticeable speed up even on non-NUMA machines. If you tried it you'd know it. I don't know if it is Moha's patch (I am not his buddy like you), but it's not the same as the one Texel uses.

In the windows version large pages is off by default, so any speed increase of brainfish (when not enabling large pages) is due to the better/different compile.

This is just plain wrong. The fact you are obviously a Linux guy and don't know Windows well doesn't make it less BS. You can turn large pages on in the program as explained in MSDN and that is exactly how it is done in Brainfish_numa (Houdart has been doing it this way since H3).

[shrapnel]

I don't know, but I'm not too impressed with Asmfish.
It's no doubt very fast and very good for Blitz and STC, but SOMETHING seems to go slightly wrong in LTC matches.
Handling of 6-men Syzygy is also dicey.
Others may agree or not, but I'm sticking to Brainfish or SDV or Stockfish BYO....more reliable.

[shrapnel]

Can't put my finger on it, but something is definitely wrong about ASMfish.
Its like Music. A guy may not know anything about Music but be still able to catch a false Note.

[Lyudmil Tsvetkov]

3000 GAME GAUNTLET
i3 380M | 2 @ 2.53 | Win7 x64 | 6 GB Ram | Gui~ CuteChess GUI 0.9.4
Draw Adjudication~120 / 40 / 2 | Resign~3 / 4.8 | Cores~1 | HT~On
Book~TopGM_2move.pgn | Hash~8MB | Concurrency~3
Ponder~No | TB's~No | Elo~Ordo 1.2.6 | Time~20" + .2"

ASM +45 over Stockfish 8
Stockfish 8 +74 over Komodo 10.2
Komodo 10.2 +10 over Stockfish 7

Code: Select all

   # PLAYER         :  RATING  ERROR  PLAYED     W     D    L   (%)  D(%)
   1 SF8 ASMPD      :    3273      7    3000  1175  1457  368  63.5  48.6
   2 Stockfish 8    :    3228     13    1000   140   592  268  43.6  59.2
   3 Komodo 10.2    :    3154     14    1000   153   367  480  33.6  36.7
   4 Stockfish 7    :    3144     14    1000    75   498  427  32.4  49.8

45 elo over SF at 20'', 20 elo at 60'', according to Pohl's list, probably just 10 elo at 5 min. and 1 elo at TC larger than half an hour.

As is very well known, fast games are excellent tool to gather some stats, but all the best moves are found at LTC.

Looking forward to an engine that will score +100 elo at VLTC in comparison to a minute or so.

[corres]

[quote="syzygy"]
In the windows version large pages is off by default, so any speed increase of brainfish (when not enabling large pages) is due to the better/different compile.
[/quote]
I agree.
The speed of Brainfish is corresponding to Stockfish8 from ultimaq web page and my binary made with mingw530posix-gcc4.9.3.

[corres]

[quote="shrapnel"]
Can't put my finger on it, but something is definitely wrong about ASMfish.
Its like Music. A guy may not know anything about Music but be still able to catch a false Note.
[/quote]
Please try Pedantfish. As Mr. Ronald de Man had written to me Pedantfish is identical to Stockfish on the other hand asmFish is a faster but less precise version of Stockfish.

[syzygy]

Ehm, the most invasive change of the "unix version" of brainfish is the insertion of the "author"'s own name, unless you run it on a NUMA machine (for which it seems to use the patch prepared by Mohammed based on - I think - Texel's code).

NUMA patch provides noticeable speed up even on non-NUMA machines. If you tried it you'd know it.

But if you understood it...

In the windows version large pages is off by default, so any speed increase of brainfish (when not enabling large pages) is due to the better/different compile.

This is just plain wrong. The fact you are obviously a Linux guy and don't know Windows well doesn't make it less BS. You can turn large pages on in the program as explained in MSDN and that is exactly how it is done in Brainfish_numa (Houdart has been doing it this way since H3).

You cannot read English?

"In the windows version large pages is off by default,"

Code: Select all

  o&#91;"Large Pages"&#93;           << Option&#40;false, on_large_pages&#41;;

"so any speed increase of brainfish (when not enabling large pages) is due to the better/different compile."

[APassionForCriminalJustic]

I don't know, but I'm not too impressed with Asmfish.
It's no doubt very fast and very good for Blitz and STC, but SOMETHING seems to go slightly wrong in LTC matches.
Handling of 6-men Syzygy is also dicey.
Others may agree or not, but I'm sticking to Brainfish or SDV or Stockfish BYO....more reliable.

I don't get it; what is this something that you are referring to? I've been using asmFish for VLTC for months. I've played hundreds of games with it - and that includes full-blown Syzygy use. There are absolutely NO issues. You need to provide evidence that proves that something isn't correct. If not then one should rightly assume that it must be in your head. If the engine can play very fast and good at blitz then why would it not be very good at VLTC?

This handling of the six-men tables being dicey is definitely not what I see. I see as expected; it is a perfect handling. Remember the days when complaints were made that Stockfish had issues with rook and pawn endgames with Syzygy?! There have always been complaints about the fish. We need evidence Shrapnel.