Skip to content

Commit 44d0d34

Browse files
ianfabianfab
committed
Diverge readme from official stockfish
1 parent 6cd272f commit 44d0d34

2 files changed

Lines changed: 4 additions & 314 deletions

File tree

‎.gitattributes‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1 @@
1+
README.md merge=ours

‎README.md‎

Lines changed: 3 additions & 314 deletions
Original file line numberDiff line numberDiff line change
@@ -107,331 +107,20 @@ The javascript binding [ffish.js](https://www.npmjs.com/package/ffish) contribut
107107

108108
For in-browser use a [port of Fairy-Stockfish to WebAssembly](https://github.com/fairy-stockfish/fairy-stockfish.wasm) is available at [npm](https://www.npmjs.com/package/fairy-stockfish-nnue.wasm). It is e.g. used for local analysis on [pychess.org](https://www.pychess.org/analysis/chess). Also see the [Fairy-Stockfish WASM demo](https://github.com/ianfab/fairy-stockfish-nnue-wasm-demo) available at https://fairy-stockfish-nnue-wasm.vercel.app/.
109109

110-
# Stockfish
111-
## Overview
112-
113-
[![Build Status](https://github.com/official-stockfish/Stockfish/actions/workflows/stockfish.yml/badge.svg)](https://github.com/official-stockfish/Stockfish/actions)
114-
[![Build Status](https://ci.appveyor.com/api/projects/status/github/official-stockfish/Stockfish?branch=master&svg=true)](https://ci.appveyor.com/project/mcostalba/stockfish/branch/master)
115-
116-
[Stockfish](https://stockfishchess.org) is a free, powerful UCI chess engine
117-
derived from Glaurung 2.1. Stockfish is not a complete chess program and requires a
118-
UCI-compatible graphical user interface (GUI) (e.g. XBoard with PolyGlot, Scid,
119-
Cute Chess, eboard, Arena, Sigma Chess, Shredder, Chess Partner or Fritz) in order
120-
to be used comfortably. Read the documentation for your GUI of choice for information
121-
about how to use Stockfish with it.
122-
123-
The Stockfish engine features two evaluation functions for chess, the classical
124-
evaluation based on handcrafted terms, and the NNUE evaluation based on efficiently
125-
updatable neural networks. The classical evaluation runs efficiently on almost all
126-
CPU architectures, while the NNUE evaluation benefits from the vector
127-
intrinsics available on most CPUs (sse2, avx2, neon, or similar).
128-
129-
130-
## Files
131-
132-
This distribution of Stockfish consists of the following files:
133-
134-
* [Readme.md](https://github.com/official-stockfish/Stockfish/blob/master/README.md), the file you are currently reading.
135-
136-
* [Copying.txt](https://github.com/official-stockfish/Stockfish/blob/master/Copying.txt), a text file containing the GNU General Public License version 3.
137-
138-
* [AUTHORS](https://github.com/official-stockfish/Stockfish/blob/master/AUTHORS), a text file with the list of authors for the project
139-
140-
* [src](https://github.com/official-stockfish/Stockfish/tree/master/src), a subdirectory containing the full source code, including a Makefile
141-
that can be used to compile Stockfish on Unix-like systems.
142-
143-
* a file with the .nnue extension, storing the neural network for the NNUE
144-
evaluation. Binary distributions will have this file embedded.
145-
146-
## The UCI protocol and available options
147-
148-
The Universal Chess Interface (UCI) is a standard protocol used to communicate with
149-
a chess engine, and is the recommended way to do so for typical graphical user interfaces
150-
(GUI) or chess tools. Stockfish implements the majority of it options as described
151-
in [the UCI protocol](https://www.shredderchess.com/download/div/uci.zip).
152-
153-
Developers can see the default values for UCI options available in Stockfish by typing
154-
`./stockfish uci` in a terminal, but the majority of users will typically see them and
155-
change them via a chess GUI. This is a list of available UCI options in Stockfish:
156-
157-
* #### Threads
158-
The number of CPU threads used for searching a position. For best performance, set
159-
this equal to the number of CPU cores available.
160-
161-
* #### Hash
162-
The size of the hash table in MB. It is recommended to set Hash after setting Threads.
163-
164-
* #### Clear Hash
165-
Clear the hash table.
166-
167-
* #### Ponder
168-
Let Stockfish ponder its next move while the opponent is thinking.
169-
170-
* #### MultiPV
171-
Output the N best lines (principal variations, PVs) when searching.
172-
Leave at 1 for best performance.
173-
174-
* #### Use NNUE
175-
Toggle between the NNUE and classical evaluation functions. If set to "true",
176-
the network parameters must be available to load from file (see also EvalFile),
177-
if they are not embedded in the binary.
178-
179-
* #### EvalFile
180-
The name of the file of the NNUE evaluation parameters. Depending on the GUI the
181-
filename might have to include the full path to the folder/directory that contains the file.
182-
Other locations, such as the directory that contains the binary and the working directory,
183-
are also searched.
184-
185-
* #### UCI_AnalyseMode
186-
An option handled by your GUI.
187-
188-
* #### UCI_Chess960
189-
An option handled by your GUI. If true, Stockfish will play Chess960.
190-
191-
* #### UCI_ShowWDL
192-
If enabled, show approximate WDL statistics as part of the engine output.
193-
These WDL numbers model expected game outcomes for a given evaluation and
194-
game ply for engine self-play at fishtest LTC conditions (60+0.6s per game).
195-
196-
* #### UCI_LimitStrength
197-
Enable weaker play aiming for an Elo rating as set by UCI_Elo. This option overrides Skill Level.
198-
199-
* #### UCI_Elo
200-
If enabled by UCI_LimitStrength, aim for an engine strength of the given Elo.
201-
This Elo rating has been calibrated at a time control of 60s+0.6s and anchored to CCRL 40/4.
202-
203-
* #### Skill Level
204-
Lower the Skill Level in order to make Stockfish play weaker (see also UCI_LimitStrength).
205-
Internally, MultiPV is enabled, and with a certain probability depending on the Skill Level a
206-
weaker move will be played.
207-
208-
* #### SyzygyPath
209-
Path to the folders/directories storing the Syzygy tablebase files. Multiple
210-
directories are to be separated by ";" on Windows and by ":" on Unix-based
211-
operating systems. Do not use spaces around the ";" or ":".
212-
213-
Example: `C:\tablebases\wdl345;C:\tablebases\wdl6;D:\tablebases\dtz345;D:\tablebases\dtz6`
214-
215-
It is recommended to store .rtbw files on an SSD. There is no loss in storing
216-
the .rtbz files on a regular HD. It is recommended to verify all md5 checksums
217-
of the downloaded tablebase files (`md5sum -c checksum.md5`) as corruption will
218-
lead to engine crashes.
219-
220-
* #### SyzygyProbeDepth
221-
Minimum remaining search depth for which a position is probed. Set this option
222-
to a higher value to probe less aggressively if you experience too much slowdown
223-
(in terms of nps) due to tablebase probing.
224-
225-
* #### Syzygy50MoveRule
226-
Disable to let fifty-move rule draws detected by Syzygy tablebase probes count
227-
as wins or losses. This is useful for ICCF correspondence games.
228-
229-
* #### SyzygyProbeLimit
230-
Limit Syzygy tablebase probing to positions with at most this many pieces left
231-
(including kings and pawns).
232-
233-
* #### Move Overhead
234-
Assume a time delay of x ms due to network and GUI overheads. This is useful to
235-
avoid losses on time in those cases.
236-
237-
* #### Slow Mover
238-
Lower values will make Stockfish take less time in games, higher values will
239-
make it think longer.
240-
241-
* #### nodestime
242-
Tells the engine to use nodes searched instead of wall time to account for
243-
elapsed time. Useful for engine testing.
244-
245-
* #### Debug Log File
246-
Write all communication to and from the engine into a text file.
247-
248-
For developers the following non-standard commands might be of interest, mainly useful for debugging:
249-
250-
* #### bench *ttSize threads limit fenFile limitType evalType*
251-
Performs a standard benchmark using various options. The signature of a version (standard node
252-
count) is obtained using all defaults. `bench` is currently `bench 16 1 13 default depth mixed`.
253-
254-
* #### compiler
255-
Give information about the compiler and environment used for building a binary.
256-
257-
* #### d
258-
Display the current position, with ascii art and fen.
259-
260-
* #### eval
261-
Return the evaluation of the current position.
262-
263-
* #### export_net [filename]
264-
Exports the currently loaded network to a file.
265-
If the currently loaded network is the embedded network and the filename
266-
is not specified then the network is saved to the file matching the name
267-
of the embedded network, as defined in evaluate.h.
268-
If the currently loaded network is not the embedded network (some net set
269-
through the UCI setoption) then the filename parameter is required and the
270-
network is saved into that file.
271-
272-
* #### flip
273-
Flips the side to move.
274-
275-
276-
## A note on classical evaluation versus NNUE evaluation
277-
278-
Both approaches assign a value to a position that is used in alpha-beta (PVS) search
279-
to find the best move. The classical evaluation computes this value as a function
280-
of various chess concepts, handcrafted by experts, tested and tuned using fishtest.
281-
The NNUE evaluation computes this value with a neural network based on basic
282-
inputs (e.g. piece positions only). The network is optimized and trained
283-
on the evaluations of millions of positions at moderate search depth.
284-
285-
The NNUE evaluation was first introduced in shogi, and ported to Stockfish afterward.
286-
It can be evaluated efficiently on CPUs, and exploits the fact that only parts
287-
of the neural network need to be updated after a typical chess move.
288-
[The nodchip repository](https://github.com/nodchip/Stockfish) provides additional
289-
tools to train and develop the NNUE networks. On CPUs supporting modern vector instructions
290-
(avx2 and similar), the NNUE evaluation results in much stronger playing strength, even
291-
if the nodes per second computed by the engine is somewhat lower (roughly 80% of nps
292-
is typical).
293-
294-
Notes:
295-
296-
1) the NNUE evaluation depends on the Stockfish binary and the network parameter
297-
file (see the EvalFile UCI option). Not every parameter file is compatible with a given
298-
Stockfish binary, but the default value of the EvalFile UCI option is the name of a network
299-
that is guaranteed to be compatible with that binary.
300-
301-
2) to use the NNUE evaluation, the additional data file with neural network parameters
302-
needs to be available. Normally, this file is already embedded in the binary or it
303-
can be downloaded. The filename for the default (recommended) net can be found as the default
304-
value of the `EvalFile` UCI option, with the format `nn-[SHA256 first 12 digits].nnue`
305-
(for instance, `nn-c157e0a5755b.nnue`). This file can be downloaded from
306-
```
307-
https://tests.stockfishchess.org/api/nn/[filename]
308-
```
309-
replacing `[filename]` as needed.
310-
311-
## What to expect from the Syzygy tablebases?
312-
313-
If the engine is searching a position that is not in the tablebases (e.g.
314-
a position with 8 pieces), it will access the tablebases during the search.
315-
If the engine reports a very large score (typically 153.xx), this means
316-
it has found a winning line into a tablebase position.
317-
318-
If the engine is given a position to search that is in the tablebases, it
319-
will use the tablebases at the beginning of the search to preselect all
320-
good moves, i.e. all moves that preserve the win or preserve the draw while
321-
taking into account the 50-move rule.
322-
It will then perform a search only on those moves. **The engine will not move
323-
immediately**, unless there is only a single good move. **The engine likely
324-
will not report a mate score, even if the position is known to be won.**
325-
326-
It is therefore clear that this behaviour is not identical to what one might
327-
be used to with Nalimov tablebases. There are technical reasons for this
328-
difference, the main technical reason being that Nalimov tablebases use the
329-
DTM metric (distance-to-mate), while the Syzygy tablebases use a variation of the
330-
DTZ metric (distance-to-zero, zero meaning any move that resets the 50-move
331-
counter). This special metric is one of the reasons that the Syzygy tablebases are
332-
more compact than Nalimov tablebases, while still storing all information
333-
needed for optimal play and in addition being able to take into account
334-
the 50-move rule.
335-
336-
## Large Pages
337-
338-
Stockfish supports large pages on Linux and Windows. Large pages make
339-
the hash access more efficient, improving the engine speed, especially
340-
on large hash sizes. Typical increases are 5..10% in terms of nodes per
341-
second, but speed increases up to 30% have been measured. The support is
342-
automatic. Stockfish attempts to use large pages when available and
343-
will fall back to regular memory allocation when this is not the case.
344-
345-
### Support on Linux
346-
347-
Large page support on Linux is obtained by the Linux kernel
348-
transparent huge pages functionality. Typically, transparent huge pages
349-
are already enabled, and no configuration is needed.
350-
351-
### Support on Windows
352-
353-
The use of large pages requires "Lock Pages in Memory" privilege. See
354-
[Enable the Lock Pages in Memory Option (Windows)](https://docs.microsoft.com/en-us/sql/database-engine/configure-windows/enable-the-lock-pages-in-memory-option-windows)
355-
on how to enable this privilege, then run [RAMMap](https://docs.microsoft.com/en-us/sysinternals/downloads/rammap)
356-
to double-check that large pages are used. We suggest that you reboot
357-
your computer after you have enabled large pages, because long Windows
358-
sessions suffer from memory fragmentation, which may prevent Stockfish
359-
from getting large pages: a fresh session is better in this regard.
360-
361-
## Compiling Stockfish yourself from the sources
362-
363-
Stockfish has support for 32 or 64-bit CPUs, certain hardware
364-
instructions, big-endian machines such as Power PC, and other platforms.
365-
366-
On Unix-like systems, it should be easy to compile Stockfish
367-
directly from the source code with the included Makefile in the folder
368-
`src`. In general it is recommended to run `make help` to see a list of make
369-
targets with corresponding descriptions.
370-
371-
```
372-
cd src
373-
make help
374-
make net
375-
make build ARCH=x86-64-modern
376-
```
377-
378-
When not using the Makefile to compile (for instance, with Microsoft MSVC) you
379-
need to manually set/unset some switches in the compiler command line; see
380-
file *types.h* for a quick reference.
381-
382-
When reporting an issue or a bug, please tell us which Stockfish version
383-
and which compiler you used to create your executable. This information
384-
can be found by typing the following command in a console:
385-
386-
```
387-
./stockfish compiler
388-
```
389-
390-
## Understanding the code base and participating in the project
391-
392-
Stockfish's improvement over the last decade has been a great community
393-
effort. There are a few ways to help contribute to its growth.
394-
395-
### Donating hardware
396-
397-
Improving Stockfish requires a massive amount of testing. You can donate
398-
your hardware resources by installing the [Fishtest Worker](https://github.com/glinscott/fishtest/wiki/Running-the-worker:-overview)
399-
and view the current tests on [Fishtest](https://tests.stockfishchess.org/tests).
400-
401-
### Improving the code
402-
403-
If you want to help improve the code, there are several valuable resources:
404-
405-
* [In this wiki,](https://www.chessprogramming.org) many techniques used in
406-
Stockfish are explained with a lot of background information.
407-
408-
* [The section on Stockfish](https://www.chessprogramming.org/Stockfish)
409-
describes many features and techniques used by Stockfish. However, it is
410-
generic rather than being focused on Stockfish's precise implementation.
411-
Nevertheless, a helpful resource.
412-
413-
* The latest source can always be found on [GitHub](https://github.com/official-stockfish/Stockfish).
414-
Discussions about Stockfish take place these days mainly in the [FishCooking](https://groups.google.com/forum/#!forum/fishcooking)
415-
group and on the [Stockfish Discord channel](https://discord.gg/nv8gDtt).
416-
The engine testing is done on [Fishtest](https://tests.stockfishchess.org/tests).
417-
If you want to help improve Stockfish, please read this [guideline](https://github.com/glinscott/fishtest/wiki/Creating-my-first-test)
418-
first, where the basics of Stockfish development are explained.
419-
420-
421110
## Terms of use
422111

423-
Stockfish is free, and distributed under the **GNU General Public License version 3**
112+
Fairy-Stockfish is free, and distributed under the **GNU General Public License version 3**
424113
(GPL v3). Essentially, this means you are free to do almost exactly
425114
what you want with the program, including distributing it among your
426115
friends, making it available for download from your website, selling
427116
it (either by itself or as part of some bigger software package), or
428117
using it as the starting point for a software project of your own.
429118

430-
The only real limitation is that whenever you distribute Stockfish in
119+
The only real limitation is that whenever you distribute Fairy-Stockfish in
431120
some way, you MUST always include the full source code, or a pointer
432121
to where the source code can be found, to generate the exact binary
433122
you are distributing. If you make any changes to the source code,
434123
these changes must also be made available under the GPL.
435124

436125
For full details, read the copy of the GPL v3 found in the file named
437-
[*Copying.txt*](https://github.com/official-stockfish/Stockfish/blob/master/Copying.txt).
126+
[*Copying.txt*](https://github.com/fairy-stockfish/Fairy-Stockfish/blob/master/Copying.txt).

0 commit comments

Comments
 (0)