[FFmpeg-devel] ffmpegs future version control system

[Måns Rullgård]

Reimar D?ffinger <Reimar.Doeffinger at stud.uni-karlsruhe.de> writes:

> Hello,
> On Thu, May 10, 2007 at 08:12:18PM +0300, Uoti Urpala wrote:
>> On Thu, 2007-05-10 at 14:08 +0100, Robert Swain wrote:
>> > On 10 May 2007, at 12:13, M?ns Rullg?rd wrote:
>> > > There is now a git repo at git://git.mplayerhq.hu/ffmpeg tracking
>> > > FFmpeg svn.  A gitweb interface is at http://git.mplayerhq.hu/.
>> > > I will try to get git-http and cvs access working tonight.
>> > 
>> > $ time git clone git://git.mplayerhq.hu/ffmpeg
>> 
>> > real    3m44.084s
>> > user    0m11.778s
>> > sys     0m54.586s
>> > 
>> > During the Indexing phase I was getting about 120-130kB/s downstream  
>> > traffic and the Resolving phase ate one core of my Core Duo 1.83GHz  
>> > (with 2GB ram, if that matters) for a minute or thereabouts. By  
>> > comparison:
>> 
>> On a significantly slower machine I got:
>> real    0m46.690s
>> user    0m21.892s
>> sys     0m0.668s
>
> My values (on an Athlon64 3200+ running at 800 MHz instead of 2 GHz
> though, 1.25 GB RAM):
> real    3m21.166s
> user    0m14.389s
> sys     0m0.860s
>
> The "real" values are obviously due to network speed, but the 54 seconds
> for sys in the first mail look like the system has some serious problems
> with git... Or maybe something else was going on that thrashed the disk
> cache? git is very sensitive to that I think.

Core2 Quad 2.66GHz, 4GB, 4Mbps ADSL:
real    2m15.700s
user    0m4.000s
sys     0m0.213s

-- 
M?ns Rullg?rd
mans at mansr.com