Date: Fri, 13 Oct 2006 02:55:01 +0200 From: Pieter de Goeje <pieter@degoeje.nl> To: Oliver Fromme <olli@lurza.secnetix.de> Cc: freebsd-hackers@freebsd.org, kientzle@freebsd.org Subject: Re: "tar -c|gzip" faster than "tar -cz"?!? Message-ID: <200610130255.01566.pieter@degoeje.nl> In-Reply-To: <200610121524.k9CFOOmS069191@lurza.secnetix.de> References: <200610121524.k9CFOOmS069191@lurza.secnetix.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thursday 12 October 2006 17:24, Oliver Fromme wrote:
> Interestingly, in both tests the compressed size of the
> "gzip" case was slightly larger than the "tar cz" case.
> That's the opposite of what I got in my very first test
> (when archiving the root file system).
I also found that the difference in size is insignificant.
> I'm not concerned about the difference in compression
> sizes, because it's in the sub-percent range. But I'm
> more concerned about the CPU times ("user" times).
> It makes quite a clear difference in all of my tests.
I can confirm, however in my opinion the difference isn't really significant.
Maybe a different compiler or gcc with better optimization settings could
produce executables that are equally fast.
First test on an AMD Athlon64 with 1GB memory running -STABLE and
~3GB /usr/ports dir.
Warmup caches with tar -cf /dev/null /usr/ports.
tar -czf /dev/null /usr/ports
622.64 real 244.37 user 9.36 sys
tar -cf - /usr/ports | gzip > /dev/null
565.15 real 195.65 user 12.38 sys
The tar|gzip command uses 18% less CPU and is 10% faster. It is clear the HDD
is the bottleneck.
Second test on an AMD Sempron with 512MB memory and 240MB directory also
running -STABLE.
Warmup caches with tar -cf /dev/null /usr/ports/distfiles/gnome2
tar -czf - /usr/ports/distfiles/gnome2 > /dev/null
33.71 real 31.56 user 1.04 sys
32.98 real 31.00 user 0.96 sys
tar -cf - /usr/ports/distfiles/gnome2 | gzip > /dev/null
29.09 real 26.65 user 1.52 sys
29.18 real 26.62 user 1.62 sys
The tar|gzip command uses 15% less CPU and is 12% faster. Very little disk I/O
occured during this test.
-- Pieter
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200610130255.01566.pieter>