kanzi-w64

此版本主要專注在修正Windows平台的編譯錯誤，並釋出x64執行檔
未來也會找時間把help資訊中文化，方便新手使用

kanzi為無損的檔案壓縮工具，保有高壓縮率的同時也大幅降低花費時間
下方提供各壓縮工具的實測數據，歡迎比較。

下載kanzi-w64

演算法等詳細說明請看原作者的wiki → https://github.com/flanglet/kanzi/wiki

測試數據

測試平台：

i7-6700HQ @2.60GHz, 16GB RAM, Windows 10 Pro 21H1 (x64)

測試的壓縮工具：

WinRAR v5.71.0
PACL64 v9.00b build 20190606 (pa format)
UHARC v0.6a
7-Zip v19.00
BCM v1.30
Kanzi v1.7 build 20210928c
Paq8px v178
Paq8pxd v63
cmix v18

1. enwik8 benchmark

Compressor	Encoding (sec)	Decoding (sec)	Size (Byte)	compare to zip
Original			100,000,000
winrar -afzip -m5	2	1	36,213,592	100%
rar -md64 -m5	23	1	27,644,573	76.3%
pacomp -c19	21.84	4.12	26,494,134	73.1%
uharc -mx -md32768	94	74.2	23,911,123	66.0%
7z -m0=ppmd:o32:mem1024m -mx9	41.06	42.78	21,918,001	60.5%
bcm -b100m	23.88	31.45	20,789,664	57.4%
kanzi -l 8 -b 100m	52.07	51.73	19,141,858	52.8%
paq8px -9eta	20462.48	21376.88	16,457,780	45.4%
paq8pxd -s15	?	?	15,967,201	44.1%
cmix	?	?	14,838,332	40.9%

2. Silesia corpus benchmark

Compressor	Encoding (sec)	Decoding (sec)	Size (Byte)	compare to zip
Original			211,938,580
winrar -afzip -m5	-	-	67,373,465	100%
rar -md256 -m5 -qo-	15.24	1.57	53,451,158	79.3%
pacomp -c19 -mt	43.13	8.67	48,109,141	71.4%
7z -m0=ppmd:o32:mem1024m -mx9	50.42	54.49	47,850,118	71.0%
bcm -b100m	-	-	46,506,680	69.0%
uharc -mx -md32768	124.2	109.2	45,172,307	67.0%
kanzi -l 8 -b 100m -j 6	63.25	64.81	40,473,911	60.0%
cmix	?	?	28,437,634	42.2%

備註：

所有數據皆以該壓縮工具的最佳壓縮率為主，第一筆數據例外，因Zip將作為基準點
順序以壓縮後檔案由大到小排列
"?"表示尚未以自己的電腦測試
下載enwik8測試檔
下載Silesia測試檔

使用方式

壓縮指令

   -v, --verbose=
        0=silent, 1=default, 2=display details, 3=display configuration,
        4=display block size and timings, 5=display extra information
        Verbosity is reduced to 1 when files are processed concurrently
        Verbosity is silently reduced to 0 when the output is 'stdout'
        (EX: The source is a directory and the number of jobs > 1).


   -f, --force
        overwrite the output file if it already exists


   -i, --input=
        mandatory name of the input file or directory or 'stdin'
        When the source is a directory, all files in it will be processed.
        Provide \. at the end of the directory name to avoid recursion
        (EX: myDir\. => no recursion)


   -o, --output=
        optional name of the output file or directory (defaults to
        ) or 'none' or 'stdout'. 'stdout' is not valid
        when the number of jobs is greater than 1.

   -b, --block=
        size of blocks, multiple of 16 (default 1 MB, max 1 GB, min 1 KB).


   -l, --level=
        set the compression level [0..8]
        Providing this option forces entropy and transform.
        0=None&None (store), 1=TEXT+LZ&HUFFMAN, 2=TEXT+ROLZ
        3=TEXT+ROLZX, 4=TEXT+BWT+RANK+ZRLT&ANS0, 5=TEXT+BWT+SRT+ZRLT&FPAQ
        6=LZP+TEXT+BWT&CM, 7=X86+RLT+TEXT&TPAQ, 8=X86+RLT+TEXT&TPAQX


   -e, --entropy=
        entropy codec [None|Huffman|ANS0|ANS1|Range|FPAQ|TPAQ|TPAQX|CM]
        (default is ANS0)


   -t, --transform=
        transform [None|BWT|BWTS|LZ|LZP|ROLZ|ROLZX|RLT|ZRLT]
                  [MTFT|RANK|SRT|TEXT|X86]
        EX: BWT+RANK or BWTS+MTFT (default is BWT+RANK+ZRLT)


   -x, --checksum
        enable block checksum


   -s, --skip
        copy blocks with high entropy instead of compressing them.

   -j, --jobs=
        maximum number of jobs the program may start concurrently
        (default is 1, maximum is 64).


EX. kanzi -c -i foo.txt -o none -b 4m -l 4 -v 3

EX. kanzi -c -i foo.txt -f -t BWT+MTFT+ZRLT -b 4m -e FPAQ -v 3 -j 4

EX. kanzi --compress --input=foo.txt --output=foo.knz --force
          --transform=BWT+MTFT+ZRLT --block=4m --entropy=FPAQ --verbose=3 --jobs=4

解壓縮指令

   -v, --verbose=
        0=silent, 1=default, 2=display details, 3=display configuration,
        4=display block size and timings, 5=display extra information
        Verbosity is reduced to 1 when files are processed concurrently
        Verbosity is silently reduced to 0 when the output is 'stdout'
        (EX: The source is a directory and the number of jobs > 1).


   -f, --force
        overwrite the output file if it already exists


   -i, --input=
        mandatory name of the input file or directory or 'stdin'
        When the source is a directory, all files in it will be processed.
        Provide \. at the end of the directory name to avoid recursion
        (EX: myDir\. => no recursion)


   -o, --output=
        optional name of the output file or directory (defaults to
        ) or 'none' or 'stdout'. 'stdout' is not valid
        when the number of jobs is greater than 1.

   -j, --jobs=
        maximum number of jobs the program may start concurrently
        (default is 1, maximum is 64).


EX. kanzi -d -i foo.knz -f -v 2 -j 2

EX. kanzi --decompress --input=foo.knz --force --verbose=2 --jobs=2

WhatTheBlock/kanzi-w64