94563a7ecd
Reduce memory usage of Rabin block array. Add an SSE optimization for bsdiff. Move integer hashing function to utils file. More updates to README. |
||
|---|---|---|
| bsdiff | ||
| lz4 | ||
| lzfx | ||
| lzma | ||
| rabin | ||
| adaptive_compress.c | ||
| allocator.c | ||
| allocator.h | ||
| bzip2_compress.c | ||
| COPYING | ||
| COPYING.LESSER | ||
| lz4_compress.c | ||
| lzfx_compress.c | ||
| lzma_compress.c | ||
| main.c | ||
| Makefile | ||
| pcompress.h | ||
| ppmd_compress.c | ||
| README.md | ||
| utils.c | ||
| utils.h | ||
| zlib_compress.c | ||
Pcompress
Copyright (C) 2012 Moinak Ghosh. All rights reserved. Use is subject to license terms.
Pcompress is a utility to do compression and decompression in parallel by splitting input data into chunks. It has a modular structure and includes support for multiple algorithms like LZMA, Bzip2, PPMD, etc., with CRC64 chunk checksums. SSE optimizations for the bundled LZMA are included. It also implements chunk-level Content-Aware Deduplication and Delta Compression features based on a Semi-Rabin Fingerprinting scheme. Delta Compression is implemented via the widely popular bsdiff algorithm. Similarity is detected using a custom hashing of maximal features of a block. When doing chunk-level dedupe it attempts to merge adjacent non-duplicate blocks index entries into a single larger entry to reduce metadata. In addition to all these it can internally split chunks at rabin boundaries to help dedupe and compression.
It has low metadata overhead and overlaps I/O and compression to achieve maximum parallelism. It also bundles a simple slab allocator to speed repeated allocation of similar chunks. It can work in pipe mode, reading from stdin and writing to stdout. It also provides some adaptive compression modes in which multiple algorithms are tried per chunk to determine the best one for the given chunk. Finally it supports 14 compression levels to allow for ultra compression modes in some algorithms.
Usage
To compress a file:
pcompress -c <algorithm> [-l <compress level>] [-s <chunk size>] <file>
Where <algorithm> can be the folowing:
lzfx - Very fast and small algorithm based on LZF.
lz4 - Ultra fast, high-throughput algorithm reaching RAM B/W at level1.
zlib - The base Zlib format compression (not Gzip).
lzma - The LZMA (Lempel-Ziv Markov) algorithm from 7Zip.
bzip2 - Bzip2 Algorithm from libbzip2.
ppmd - The PPMd algorithm excellent for textual data. PPMd requires
at least 64MB X CPUs more memory than the other modes.
adapt - Adaptive mode where ppmd or bzip2 will be used per chunk,
depending on which one produces better compression. This mode
is obviously fairly slow and requires lots of memory.
adapt2 - Adaptive mode which includes ppmd and lzma. This requires
more memory than adapt mode, is slower and potentially gives
the best compression.
<chunk_size> - This can be in bytes or can use the following suffixes:
g - Gigabyte, m - Megabyte, k - Kilobyte.
Larger chunks produce better compression at the cost of memory.
<compress_level> - Can be a number from 0 meaning minimum and 14 meaning
maximum compression.
To decompress a file compressed using above command:
pcompress -d <compressed file> <target file>
To operate as a pipe, read from stdin and write to stdout:
pcompress -p ...
Attempt Rabin fingerprinting based deduplication on chunks:
pcompress -D ...
pcompress -D -r ... - Do NOT split chunks at a rabin boundary. Default
is to split.
Perform Delta Encoding in addition to Exact Dedup:
pcompress -E ... - This also implies '-D'.
Number of threads can optionally be specified: -t <1 - 256 count>
Pass '-M' to display memory allocator statistics
Pass '-C' to display compression statistics
Environment Variables
Set ALLOCATOR_BYPASS=1 in the environment to avoid using the the built-in allocator. Due to the the way it rounds up an allocation request to the nearest slab the built-in allocator can allocate extra unused memory.
Examples
Compress "file.tar" using bzip2 level 6, 64MB chunk size and use 4 threads. In addition perform exact deduplication and delta compression prior to compression.
pcompress -D -E -c bzip2 -l6 -s64m -t4 file.tar
Compress "file.tar" using extreme compression mode of LZMA and a chunk size of of 1GB. Allow pcompress to detect the number of CPU cores and use as many threads.
pcompress -c lzma -l14 -s1g file.tar
Compression Algorithms
LZFX - Ultra Fast, average compression. This algorithm is the fastest overall. Levels: 1 - 5 LZ4 - Very Fast, better compression than LZFX. Levels: 1 - 3 Zlib - Fast, better compression. Levels: 1 - 9 Bzip2 - Slow, much better compression than Zlib. Levels: 1 - 9 LZMA - Very slow. Extreme compression. Levels: 1 - 14 PPMD - Slow. Extreme compression for Text, average compression for binary. Levels: 1 - 14.
Adapt - Very slow synthetic mode. Both Bzip2 and PPMD are tried per chunk and better result selected. Levels: 1 - 14 Adapt2 - Ultra slow synthetic mode. Both LZMA and PPMD are tried per chunk and better result selected. Can give best compression ration when splitting file into multiple chunks. Levels: 1 - 14
It is possible for a single chunk to span the entire file if enough RAM is available. However for adaptive modes to be effective for large files, especially multi-file archives splitting into chunks is required so that best compression algorithm can be selected for textual and binary portions.
Caveats
This utility can gobble up RAM depending on compression algorithm, compression level, and dedupe being enabled. Larger chunk sizes can give better compression ratio but at the same time use more RAM.
In some cases for files less than a gigabyte. Using Delta Compression in addition to exact Dedupe can have a slight negative impact on LZMA compression ratio especially when using the large-window ultra compression levels above 12.