Moinak Ghosh
e788eb43b8
Implement Delta Encoding based on modified bsdiff.
...
Change to more accurate Sketch value computation approach.
2012-07-19 21:41:07 +05:30
Moinak Ghosh
1da2c40888
Use a rolling checksum based sketch value for a rabin chunk instead of a CRC64 checksum.
...
Avoids additional table-lookup memory access.
Reduce Rabin window size to avoid overflows in sketch value.
No need to maintain rolling checksum in Rabin context.
A few comment cleanups.
2012-07-13 22:06:55 +05:30
Moinak Ghosh
a873f92e41
Fix crash when decompressing deduped archive.
...
Ensure correct level is passed to lzma.
Avoid branch when wrapping rabin window position and check for rabin window size to be power of 2.
Update rabin parameters check for adaptive modes.
Add detection of 7-bit text/8-bit binary data for later use.
2012-07-10 20:14:23 +05:30
Moinak Ghosh
010f49f412
Implement ability to partition chunks at the last rabin boundary instead of fixed size.
2012-07-08 21:44:08 +05:30
Moinak Ghosh
d3f5287ee5
Update License info to LGPLv3.
2012-07-07 22:18:29 +05:30
Moinak Ghosh
172432698e
Adjust Rabin parameters for chunksize > LZMA window size.
2012-07-07 00:01:13 +05:30
Moinak Ghosh
ea923b84f0
Use different min block size and Rabin break pattern depending on compression algo.
...
Cleanup some cruft.
2012-07-06 23:24:12 +05:30
Moinak Ghosh
f5ce45b16e
Techniques to better reduce Rabin Metadata.
...
Fix wrong chunk sizing with dedup enabled.
2012-07-06 00:16:02 +05:30
Moinak Ghosh
3f0e1952ef
Fix compile warnings.
2012-07-04 23:20:10 +05:30
Moinak Ghosh
1eee08040f
Use diferent average Rabin block sizes depending on compression algorithm.
...
Misc cleanups.
2012-07-03 22:47:24 +05:30
Moinak Ghosh
a13c61e926
Change rabin index encoding scheme for better metadata compression.
2012-07-02 22:08:03 +05:30
Moinak Ghosh
a1825a2305
Implement Parallel deduplication support.
...
Restructure compression functions to take chunk flag as argument.
Add missing error flag printing in LZMA.
Only create enough threads as needed by chunk size and file size.
Minor cleanups and variable name changes.
2012-07-01 21:44:02 +05:30
Moinak Ghosh
f9c3644459
Updates to Rabin based Dedup.
...
Change command line option.
2012-06-29 23:45:06 +05:30
Moinak Ghosh
cbf9728278
Implement Deduplication based on Rabin Fingerprinting: work in progress.
...
Fix bug that prevented pipe mode from being used.
Allow building without specialized allocator.
Use basic optimize flag in debuig build.
2012-06-29 18:23:55 +05:30
Moinak Ghosh
8f5f531967
Add license and other minor fixes.
2012-06-21 20:40:43 +05:30
Moinak Ghosh
733923cbf2
Add ability to adjust chunk boundary based on Rabin Fingerprinting to improve compression.
...
Remove unnecessary checks in compression loop.
2012-06-21 20:27:05 +05:30