Re: [Lzip-bug] lzip -1 versus gzip -1 versus xz -1

[Tino Lange]

Hi!

Just some example for my question:

> $ ll 125.db; time lzip 125.db; ll 125.db.lz; lzip -d 125.db.lz
> -rw-rw---- 1 tlange tlange 13574639 Mar 12 07:18 125.db
> real    0m13.661s
> user    0m13.529s
> sys     0m0.131s
> -rw-rw---- 1 tlange tlange 11515382 Mar 12 07:18 125.db.lz

> $ ll 125.db; time lzip -1 125.db; ll 125.db.lz; lzip -d 125.db.lz
> -rw-rw---- 1 tlange tlange 13574639 Mar 12 07:18 125.db
> real    0m8.585s
> user    0m8.522s
> sys     0m0.062s
> -rw-rw---- 1 tlange tlange 11672937 Mar 12 07:18 125.db.lz

> $ ll 125.db; time xz 125.db; ll 125.db.xz; xz -d 125.db.xz
> -rw-rw---- 1 tlange tlange 13574639 Mar 12 07:18 125.db
> real    0m8.046s
> user    0m7.873s
> sys     0m0.136s
> -rw-rw---- 1 tlange tlange 11514524 Mar 12 07:18 125.db.xz

> $ ll 125.db; time xz -1 125.db; ll 125.db.xz; xz -d 125.db.xz
> -rw-rw---- 1 tlange tlange 13574639 Mar 12 07:18 125.db
> real    0m3.028s
> user    0m2.996s
> sys     0m0.034s
> -rw-rw---- 1 tlange tlange 11869128 Mar 12 07:18 125.db.xz

> $ ll 125.db; time gzip 125.db; ll 125.db.gz; gzip -d 125.db.gz
> -rw-rw---- 1 tlange tlange 13574639 Mar 12 07:18 125.db
> real    0m0.871s
> user    0m0.833s
> sys     0m0.038s
> -rw-rw---- 1 tlange tlange 12105904 Mar 12 07:18 125.db.gz

> $ ll 125.db; time gzip -1 125.db; ll 125.db.gz; gzip -d 125.db.gz
> -rw-rw---- 1 tlange tlange 13574639 Mar 12 07:18 125.db
> real    0m0.802s
> user    0m0.786s
> sys     0m0.016s
> -rw-rw---- 1 tlange tlange 12135722 Mar 12 07:18 125.db.gz

It seems that while xz and lzip are comparable with the default setting, 
they differ much in using -1. xz becomes quite fast while still having a 
quite good compression, whereas lzip becomes faster but the time_needed 
per compression_result ratio is much worse?

Cheers,

Tino