Hi,
Out of this thread Bill Freeman and I got onto a side discussion of the
compression efficiency of gzip(1) vs blocksize. This led to an
experiment to take the first 2MB of /usr/doc/HOWTO/* and /bin/* and see
how well that blob compresses when broken up into blocks where each block
is gzip compressed separately. (for blocksizes 128, 256, 512, ... to 2MB.)
The results are below, in the form:
blocksize, compression-ratio, compressed-size/uncompressed-size
This could be useful background info to keep in the back of one's mind...
Even small blocks of 1-2K get a reasonable fraction of the full compression.
/usr/doc/HOWTO/*
total size: 2000000
128 0.9773 1954520/2000000
256 0.7569 1513703/2000000
512 0.6234 1246810/2000000
1024 0.5348 1069536/2000000
2048 0.4703 940685/2000000
4096 0.4240 847974/2000000
8192 0.3896 779129/2000000
16384 0.3637 727375/2000000
32768 0.3427 685395/2000000
65536 0.3283 656661/2000000
131072 0.3206 641290/2000000
262144 0.3166 633217/2000000
524288 0.3146 629162/2000000
1048576 0.3136 627276/2000000
2000000 0.3134 626753/2000000
/bin/*
total size: 2000000
128 1.0111 2022191/2000000
256 0.8391 1678176/2000000
512 0.7242 1448394/2000000
1024 0.6346 1269181/2000000
2048 0.5748 1149687/2000000
4096 0.5370 1073970/2000000
8192 0.5130 1026006/2000000
16384 0.4961 992103/2000000
32768 0.4837 967453/2000000
65536 0.4758 951601/2000000
131072 0.4730 946040/2000000
262144 0.4710 942082/2000000
524288 0.4704 940761/2000000
1048576 0.4692 938445/2000000
2000000 0.4693 938545/2000000
*****************************************************************
To unsubscribe from this list, send mail to [EMAIL PROTECTED]
with the text 'unsubscribe gnhlug' in the message body.
*****************************************************************
