[H-GEN] Compressing ISO images (Was: Darwin ?)

[Greg Black]

[ Humbug *General* list - semi-serious discussions about Humbug and     ]
[ Unix-related topics. Posts from non-subscribed addresses will vanish. ]

Jason Henry Parker wrote:

| Sandra Milne <silne at optushome.com.au> writes:
| 
| > sorry to change tack, but does gzipping an iso make much difference to
| > the size?
| 
| It would appear the answer is "yes, perhaps":

With compression, the answer really does depend on the data and
it's frequently difficult to estimate the results without really
doing the testing.

I have no reason to download the ISO that was mentioned, but
I've done some tests on a couple of handy ISO images, using both
gzip and bzip2:

    $ ls -l
    total 3410120
    -rw-r--r--  1 gjb  gjb  667224064 May 26 10:20 cd1.iso
    -rw-r--r--  1 gjb  gjb  637675791 May 26 12:31 cd1.iso.bz2
    -rw-r--r--  1 gjb  gjb  633177816 May 26 11:14 cd1.iso.gz
    -rw-r--r--  1 gjb  gjb  627050496 May 26 10:26 cd2.iso
    -rw-r--r--  1 gjb  gjb  457136318 May 26 15:06 cd2.iso.bz2
    -rw-r--r--  1 gjb  gjb  467892293 May 26 11:46 cd2.iso.gz

In one case, the compression was hardly worth the effort, but in
the second case it made a significant difference.  Also, it's
worth noting that bzip2, despite being /much/ slower than gzip,
sometimes does worse than gzip and sometimes does marginally
better.  The times for the various compressions were also quite
interesting:

    $ time gzip -9 < cd1.iso > cd1.iso.gz
    541.596s real   496.074s user   25.933s system
    $ time gzip -9 < cd2.iso > cd2.iso.gz
    855.109s real   799.453s user   22.575s system
    $ time bzip2 -9 < cd1.iso > cd1.iso.bz2
    2240.287s real  2107.428s user  22.281s system
    $ time bzip2 -9 < cd2.iso > cd2.iso.bz2
    1948.464s real  1837.384s user  17.390s system

We can see that gzip managed both jobs in between a half and a
quarter the time it took bzip2; but gzip wasted far less time
than bzip2 on the largely incompressible 1st ISO (9m versus
37m); and gzip managed to shave 160 Mbytes off the second file
in 14m compared with bzip2's 170 Mbytes in 32m.

Anyway, as they say, YMMV.

Greg

--
* This is list (humbug) general handled by majordomo at lists.humbug.org.au .
* Postings to this list are only accepted from subscribed addresses of
* lists 'general' or 'general-post'.  See http://www.humbug.org.au/