1LZMA SDK 18.06 2-------------- 3 4LZMA SDK provides the documentation, samples, header files, 5libraries, and tools you need to develop applications that 6use 7z / LZMA / LZMA2 / XZ compression. 7 8LZMA is an improved version of famous LZ77 compression algorithm. 9It was improved in way of maximum increasing of compression ratio, 10keeping high decompression speed and low memory requirements for 11decompressing. 12 13LZMA2 is a LZMA based compression method. LZMA2 provides better 14multithreading support for compression than LZMA and some other improvements. 15 167z is a file format for data compression and file archiving. 177z is a main file format for 7-Zip compression program (www.7-zip.org). 187z format supports different compression methods: LZMA, LZMA2 and others. 197z also supports AES-256 based encryption. 20 21XZ is a file format for data compression that uses LZMA2 compression. 22XZ format provides additional features: SHA/CRC check, filters for 23improved compression ratio, splitting to blocks and streams, 24 25 26 27LICENSE 28------- 29 30LZMA SDK is written and placed in the public domain by Igor Pavlov. 31 32Some code in LZMA SDK is based on public domain code from another developers: 33 1) PPMd var.H (2001): Dmitry Shkarin 34 2) SHA-256: Wei Dai (Crypto++ library) 35 36Anyone is free to copy, modify, publish, use, compile, sell, or distribute the 37original LZMA SDK code, either in source code form or as a compiled binary, for 38any purpose, commercial or non-commercial, and by any means. 39 40LZMA SDK code is compatible with open source licenses, for example, you can 41include it to GNU GPL or GNU LGPL code. 42 43 44LZMA SDK Contents 45----------------- 46 47 Source code: 48 49 - C / C++ / C# / Java - LZMA compression and decompression 50 - C / C++ - LZMA2 compression and decompression 51 - C / C++ - XZ compression and decompression 52 - C - 7z decompression 53 - C++ - 7z compression and decompression 54 - C - small SFXs for installers (7z decompression) 55 - C++ - SFXs and SFXs for installers (7z decompression) 56 57 Precomiled binaries: 58 59 - console programs for lzma / 7z / xz compression and decompression 60 - SFX modules for installers. 61 62 63UNIX/Linux version 64------------------ 65To compile C++ version of file->file LZMA encoding, go to directory 66CPP/7zip/Bundles/LzmaCon 67and call make to recompile it: 68 make -f makefile.gcc clean all 69 70In some UNIX/Linux versions you must compile LZMA with static libraries. 71To compile with static libraries, you can use 72LIB = -lm -static 73 74Also you can use p7zip (port of 7-Zip for POSIX systems like Unix or Linux): 75 76 http://p7zip.sourceforge.net/ 77 78 79Files 80----- 81 82DOC/7zC.txt - 7z ANSI-C Decoder description 83DOC/7zFormat.txt - 7z Format description 84DOC/installer.txt - information about 7-Zip for installers 85DOC/lzma.txt - LZMA compression description 86DOC/lzma-sdk.txt - LZMA SDK description (this file) 87DOC/lzma-history.txt - history of LZMA SDK 88DOC/lzma-specification.txt - Specification of LZMA 89DOC/Methods.txt - Compression method IDs for .7z 90 91bin/installer/ - example script to create installer that uses SFX module, 92 93bin/7zdec.exe - simplified 7z archive decoder 94bin/7zr.exe - 7-Zip console program (reduced version) 95bin/x64/7zr.exe - 7-Zip console program (reduced version) (x64 version) 96bin/lzma.exe - file->file LZMA encoder/decoder for Windows 97bin/7zS2.sfx - small SFX module for installers (GUI version) 98bin/7zS2con.sfx - small SFX module for installers (Console version) 99bin/7zSD.sfx - SFX module for installers. 100 101 1027zDec.exe 103--------- 1047zDec.exe is simplified 7z archive decoder. 105It supports only LZMA, LZMA2, and PPMd methods. 1067zDec decodes whole solid block from 7z archive to RAM. 107The RAM consumption can be high. 108 109 110 111 112Source code structure 113--------------------- 114 115 116Asm/ - asm files (optimized code for CRC calculation and Intel-AES encryption) 117 118C/ - C files (compression / decompression and other) 119 Util/ 120 7z - 7z decoder program (decoding 7z files) 121 Lzma - LZMA program (file->file LZMA encoder/decoder). 122 LzmaLib - LZMA library (.DLL for Windows) 123 SfxSetup - small SFX module for installers 124 125CPP/ -- CPP files 126 127 Common - common files for C++ projects 128 Windows - common files for Windows related code 129 130 7zip - files related to 7-Zip 131 132 Archive - files related to archiving 133 134 Common - common files for archive handling 135 7z - 7z C++ Encoder/Decoder 136 137 Bundles - Modules that are bundles of other modules (files) 138 139 Alone7z - 7zr.exe: Standalone 7-Zip console program (reduced version) 140 Format7zExtractR - 7zxr.dll: Reduced version of 7z DLL: extracting from 7z/LZMA/BCJ/BCJ2. 141 Format7zR - 7zr.dll: Reduced version of 7z DLL: extracting/compressing to 7z/LZMA/BCJ/BCJ2 142 LzmaCon - lzma.exe: LZMA compression/decompression 143 LzmaSpec - example code for LZMA Specification 144 SFXCon - 7zCon.sfx: Console 7z SFX module 145 SFXSetup - 7zS.sfx: 7z SFX module for installers 146 SFXWin - 7z.sfx: GUI 7z SFX module 147 148 Common - common files for 7-Zip 149 150 Compress - files for compression/decompression 151 152 Crypto - files for encryption / decompression 153 154 UI - User Interface files 155 156 Client7z - Test application for 7za.dll, 7zr.dll, 7zxr.dll 157 Common - Common UI files 158 Console - Code for console program (7z.exe) 159 Explorer - Some code from 7-Zip Shell extension 160 FileManager - Some GUI code from 7-Zip File Manager 161 GUI - Some GUI code from 7-Zip 162 163 164CS/ - C# files 165 7zip 166 Common - some common files for 7-Zip 167 Compress - files related to compression/decompression 168 LZ - files related to LZ (Lempel-Ziv) compression algorithm 169 LZMA - LZMA compression/decompression 170 LzmaAlone - file->file LZMA compression/decompression 171 RangeCoder - Range Coder (special code of compression/decompression) 172 173Java/ - Java files 174 SevenZip 175 Compression - files related to compression/decompression 176 LZ - files related to LZ (Lempel-Ziv) compression algorithm 177 LZMA - LZMA compression/decompression 178 RangeCoder - Range Coder (special code of compression/decompression) 179 180 181Note: 182 Asm / C / C++ source code of LZMA SDK is part of 7-Zip's source code. 183 7-Zip's source code can be downloaded from 7-Zip's SourceForge page: 184 185 http://sourceforge.net/projects/sevenzip/ 186 187 188 189LZMA features 190------------- 191 - Variable dictionary size (up to 1 GB) 192 - Estimated compressing speed: about 2 MB/s on 2 GHz CPU 193 - Estimated decompressing speed: 194 - 20-30 MB/s on modern 2 GHz cpu 195 - 1-2 MB/s on 200 MHz simple RISC cpu: (ARM, MIPS, PowerPC) 196 - Small memory requirements for decompressing (16 KB + DictionarySize) 197 - Small code size for decompressing: 5-8 KB 198 199LZMA decoder uses only integer operations and can be 200implemented in any modern 32-bit CPU (or on 16-bit CPU with some conditions). 201 202Some critical operations that affect the speed of LZMA decompression: 203 1) 32*16 bit integer multiply 204 2) Mispredicted branches (penalty mostly depends from pipeline length) 205 3) 32-bit shift and arithmetic operations 206 207The speed of LZMA decompressing mostly depends from CPU speed. 208Memory speed has no big meaning. But if your CPU has small data cache, 209overall weight of memory speed will slightly increase. 210 211 212How To Use 213---------- 214 215Using LZMA encoder/decoder executable 216-------------------------------------- 217 218Usage: LZMA <e|d> inputFile outputFile [<switches>...] 219 220 e: encode file 221 222 d: decode file 223 224 b: Benchmark. There are two tests: compressing and decompressing 225 with LZMA method. Benchmark shows rating in MIPS (million 226 instructions per second). Rating value is calculated from 227 measured speed and it is normalized with Intel's Core 2 results. 228 Also Benchmark checks possible hardware errors (RAM 229 errors in most cases). Benchmark uses these settings: 230 (-a1, -d21, -fb32, -mfbt4). You can change only -d parameter. 231 Also you can change the number of iterations. Example for 30 iterations: 232 LZMA b 30 233 Default number of iterations is 10. 234 235<Switches> 236 237 238 -a{N}: set compression mode 0 = fast, 1 = normal 239 default: 1 (normal) 240 241 d{N}: Sets Dictionary size - [0, 30], default: 23 (8MB) 242 The maximum value for dictionary size is 1 GB = 2^30 bytes. 243 Dictionary size is calculated as DictionarySize = 2^N bytes. 244 For decompressing file compressed by LZMA method with dictionary 245 size D = 2^N you need about D bytes of memory (RAM). 246 247 -fb{N}: set number of fast bytes - [5, 273], default: 128 248 Usually big number gives a little bit better compression ratio 249 and slower compression process. 250 251 -lc{N}: set number of literal context bits - [0, 8], default: 3 252 Sometimes lc=4 gives gain for big files. 253 254 -lp{N}: set number of literal pos bits - [0, 4], default: 0 255 lp switch is intended for periodical data when period is 256 equal 2^N. For example, for 32-bit (4 bytes) 257 periodical data you can use lp=2. Often it's better to set lc0, 258 if you change lp switch. 259 260 -pb{N}: set number of pos bits - [0, 4], default: 2 261 pb switch is intended for periodical data 262 when period is equal 2^N. 263 264 -mf{MF_ID}: set Match Finder. Default: bt4. 265 Algorithms from hc* group doesn't provide good compression 266 ratio, but they often works pretty fast in combination with 267 fast mode (-a0). 268 269 Memory requirements depend from dictionary size 270 (parameter "d" in table below). 271 272 MF_ID Memory Description 273 274 bt2 d * 9.5 + 4MB Binary Tree with 2 bytes hashing. 275 bt3 d * 11.5 + 4MB Binary Tree with 3 bytes hashing. 276 bt4 d * 11.5 + 4MB Binary Tree with 4 bytes hashing. 277 hc4 d * 7.5 + 4MB Hash Chain with 4 bytes hashing. 278 279 -eos: write End Of Stream marker. By default LZMA doesn't write 280 eos marker, since LZMA decoder knows uncompressed size 281 stored in .lzma file header. 282 283 -si: Read data from stdin (it will write End Of Stream marker). 284 -so: Write data to stdout 285 286 287Examples: 288 2891) LZMA e file.bin file.lzma -d16 -lc0 290 291compresses file.bin to file.lzma with 64 KB dictionary (2^16=64K) 292and 0 literal context bits. -lc0 allows to reduce memory requirements 293for decompression. 294 295 2962) LZMA e file.bin file.lzma -lc0 -lp2 297 298compresses file.bin to file.lzma with settings suitable 299for 32-bit periodical data (for example, ARM or MIPS code). 300 3013) LZMA d file.lzma file.bin 302 303decompresses file.lzma to file.bin. 304 305 306Compression ratio hints 307----------------------- 308 309Recommendations 310--------------- 311 312To increase the compression ratio for LZMA compressing it's desirable 313to have aligned data (if it's possible) and also it's desirable to locate 314data in such order, where code is grouped in one place and data is 315grouped in other place (it's better than such mixing: code, data, code, 316data, ...). 317 318 319Filters 320------- 321You can increase the compression ratio for some data types, using 322special filters before compressing. For example, it's possible to 323increase the compression ratio on 5-10% for code for those CPU ISAs: 324x86, IA-64, ARM, ARM-Thumb, PowerPC, SPARC. 325 326You can find C source code of such filters in C/Bra*.* files 327 328You can check the compression ratio gain of these filters with such 3297-Zip commands (example for ARM code): 330No filter: 331 7z a a1.7z a.bin -m0=lzma 332 333With filter for little-endian ARM code: 334 7z a a2.7z a.bin -m0=arm -m1=lzma 335 336It works in such manner: 337Compressing = Filter_encoding + LZMA_encoding 338Decompressing = LZMA_decoding + Filter_decoding 339 340Compressing and decompressing speed of such filters is very high, 341so it will not increase decompressing time too much. 342Moreover, it reduces decompression time for LZMA_decoding, 343since compression ratio with filtering is higher. 344 345These filters convert CALL (calling procedure) instructions 346from relative offsets to absolute addresses, so such data becomes more 347compressible. 348 349For some ISAs (for example, for MIPS) it's impossible to get gain from such filter. 350 351 352 353--- 354 355http://www.7-zip.org 356http://www.7-zip.org/sdk.html 357http://www.7-zip.org/support.html 358