1 |
÷ Mon, 12 May 2014 14:17:11 +0200 |
2 |
Tom Wijsman <TomWij@g.o> ÐÉÛÅÔ: |
3 |
|
4 |
> On Mon, 12 May 2014 14:47:36 +0400 |
5 |
> Alexander Tsoy <alexander@××××.me> wrote: |
6 |
> |
7 |
> > Here is my test results. xz options: "--lzma2=preset=6e,dict=4MiB". |
8 |
> > Larger dictionary size does not improve compression ratio, I get |
9 |
> > even worse results with just "-6e" or "-9e". man-bz2 is a full copy of |
10 |
> > my /usr/share/man, man-xz is a recompressed one. |
11 |
> |
12 |
> Picking a random post to reply; if you don't already, please consider |
13 |
> to do these tests in tmpfs to cancel out any fs / storage differences. |
14 |
> |
15 |
|
16 |
The same test in tmpfs. |
17 |
|
18 |
$ time find man-bz2/ -type f -name "*.bz2" -exec bzcat '{}' > /dev/null \; |
19 |
|
20 |
real 0m35.895s |
21 |
user 0m14.232s |
22 |
sys 0m14.121s |
23 |
$ time find man-xz/ -type f -name "*.xz" -exec xzcat '{}' > /dev/null \; |
24 |
|
25 |
real 0m44.342s |
26 |
user 0m16.842s |
27 |
sys 0m21.459s |
28 |
|
29 |
|
30 |
And here is additional test. It shows where is actually a bottleneck. |
31 |
xz is faster in decompression, but looks like it just has a slower |
32 |
process initialization speed. So it's slower in decompressing of a single |
33 |
little file. |
34 |
|
35 |
$ time find man-bz2/ -type f -name "*.bz2" -exec bzcat '{}' > /dev/null \+ |
36 |
|
37 |
real 0m10.096s |
38 |
user 0m9.000s |
39 |
sys 0m0.787s |
40 |
$ time find man-xz/ -type f -name "*.xz" -exec xzcat '{}' > /dev/null \+ |
41 |
|
42 |
real 0m7.846s |
43 |
user 0m7.108s |
44 |
sys 0m0.487s |
45 |
|
46 |
-- |
47 |
Alexander Tsoy |