Gregory Maxwell wrote: > Please no bzip2. > The problem with bzip2 is that it is every bit as slow to compress as > it is to decompress.
Ok, lets do a quick comparison for x86 (Run on a 2.8Ghz Xeon): === .txt file (Apache Log) 1352430763 bytes. ... gzip: gzip -9: real 1m33.988s user 1m13.559s sys 0m3.280s log.txt.gz: 100924724 bytes (92.6% compression) gunzip: real 0m37.030s (32% of compression time) user 0m7.998s sys 0m6.437s ... bzip2: bzip2 -9: real 23m21.282s (15x longer than gzip) user 21m45.897s sys 0m4.417s log.txt.bz2: 61218312 bytes (95.5% compression) bunzip2: real 3m46.324s (16.1% of compression time) user 3m20.079s sys 0m13.365s ... lzma: 7zr a -t7z -m0=lzma -mx=9 -mfb=64 -md=32m -ms=on: real 26m26.715s (17x longer than gzip) user 30m40.920s sys 0m23.894s log.txt.7z: 64230100 bytes (95.3% compression) 7zr e: real 0m48.408s (3% of compression time) user 0m23.985s sys 0m7.324s === .mpg file 115996672 bytes ... gzip: gzip -9: real 0m13.458s user 0m12.402s sys 0m0.641s test.mpg.gz: 88087369 bytes (24% compression) gunzip: real 0m2.751s (23% of compression time) user 0m1.982s sys 0m0.656s ... bzip2: bzip2 -0: real 1m7.146s (5.2x longer than gzip) user 1m3.320s sys 0m0.590s test.mpg.bz2: 86543950 bytes (25.4% compression) bunzip2: real 0m33.261s (49.3% of compression time) user 0m30.615s sys 0m1.497s ... lzma: 7zr a -t7z -m0=lzma -mx=9 -mfb=64 -md=32m -ms=on real 2m16.828s (10.5x longer than gzip) user 3m7.780s sys 0m2.843s test.mpg.7z: 84236048 bytes (27.4% compression) 7zr e: real 0m19.416s (14% of compression time) user 0m17.697s sys 0m0.741s === Run on an otherwise inactive box. -- Ken Snider
