Date        Version  New features compared to previous version

2007-04-16  0.1      Bytecoder, bitcoder and 2^n semi-dynamic RangeCoder
                     Greedy parsing
                     Only 4+ byte matches
                     Hashtable matchfinders: simple and caching
                     Blockwise compression

2007-05-11  0.2      Huffman coder
                     Lazy parsing
                     3-byte matches
                     Dropping too far matches
                     Sliding window
                     Gzip-like cmdline interface

2007-07-25  0.25     Repdist, repboth and repchar codes
                     2-byte matches
                     Improved lz parsing
                     Delta filter

2008-03-13  0.3      Huffman decoder with variable block size
                     Tested on 100 gb of data and fixed all bugs found

2008-04-08  0.4      Cycling matchfinder
                     Large, updated at each step hashes for 2/3-byte strings
                     More interactive user interface

2008-06-02  0.4a     Fixed bug in Huffman coder: unstable sorting
                     Win32 & Linux versions

2008-10-28  0.4b     Added -11:c1 .. -11:c3 modes

2014-01-15  0.5      Removed modes -6 -8 -10 -12, renumerated -7/9/11 to -6/7/8
                     -7/-8: 1% better compression and 5-20% better speed due to use of ht5
                     New -9..-11 modes: 2gb hash and ht5..ht7 match finders
                     In small tornado.exe -c1..-c3 available for -7..-11 modes
                     Support for -l not power of 2   (in all matchfinders?)
                     Options -ah, -al, -t#, -cpu, -q, -delete, stats->stderr
                     Checks at decoding in order to prevent segfaults
                     MSVC/ICL compatibility, use of wall-clock times
                     Linux: 64-bit file offsets!
                     PowerPC (big-endian cpus) compatibility, thanks to Joachim Henke
                     64-bit compilers support!
                     -3 mode compression improved by 2-3%
                     2mb hash in -4/-5 modes improved speed by 20-30%
                     Fixed long-standing decompression bug (table split across 2 write chunks)
                     Removed CHECK_FOR_DATA_TABLE(2) in -3/-4 modes


2014-03-08  0.6      http://freearc.org/download/research/tornado06.zip

Compression:
* new -11..-16 modes employing the new optimal parser (-p4)
* -fb (FAST BYTES), new parameter used only by the optimal parser
* old -10 mode was removed, old -11 mode was better tuned and renamed to -10
* -s1 option meaning changed to "64kb hash3 + 4kb hash2", slightly improving compression in -5/-6 modes
* new chash5..7 (-x15..17) and bt4..7 (-x24..27) match finders
* all chash/cchash/bt matchfinders support -l values that is not power of 2
* new manual.txt explaining all compression parameters

Command line and progress indicator:
* -slp/-rem options, large pages are allocated by default
* Win7 taskbar progress indication (green bar) plus info in the console window title
* returns Errcode 2 on any error, clears the window title on ^Break
* prints ranges and default values for each option, checks option correctness
* on Windows, prohibits simultaneous writing by several compression processes to the same output file
* prints byte-exact filesizes; speeds are measured in MiB/s
* progress indicator is more accurate and updated only once per 0.2 seconds

Compilation:
* on Windows, compile.cmd supports many GCC/MSVC/ICL versions
* on Linux, compile.sh now produces executables that can process files larger than 2 GB

