Re: Receiving objects: 44% (121381/272167), 124.05 MiB | 78 KiB/s



From University of Toronto network, git.gnome.org and svn.gnome.org look identical, with git.kernel.org being 60% faster. Next question is, whether / how much the ping time is relevant to the throughput.

behdad

behdad:pango 0 (harfbuzz-ng*)$ ping git.gnome.org
PING git.gnome.org (209.132.176.202) 56(84) bytes of data.
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=1 ttl=49 time=131 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=2 ttl=49 time=110 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=3 ttl=49 time=119 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=4 ttl=49 time=119 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=5 ttl=49 time=142 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=6 ttl=49 time=126 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=7 ttl=49 time=113 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=8 ttl=49 time=110 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=9 ttl=49 time=117 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=10 ttl=49 time=300 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=11 ttl=49 time=110 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=12 ttl=49 time=121 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=13 ttl=49 time=131 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=14 ttl=49 time=111 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=15 ttl=49 time=133 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=16 ttl=49 time=121 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=17 ttl=49 time=240 ms
64 bytes from git.gnome.org (209.132.176.202): icmp_seq=18 ttl=49 time=144 ms
^C
--- git.gnome.org ping statistics ---
18 packets transmitted, 18 received, 0% packet loss, time 17758ms
rtt min/avg/max/mdev = 110.370/139.258/300.620/48.509 ms
behdad:pango 0 (harfbuzz-ng*)$ ping svn.gnome.org
PING svn.gnome.org (91.189.93.3) 56(84) bytes of data.
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=1 ttl=51 time=127 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=2 ttl=51 time=115 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=3 ttl=51 time=133 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=4 ttl=51 time=123 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=5 ttl=51 time=153 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=6 ttl=51 time=114 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=7 ttl=51 time=110 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=8 ttl=51 time=134 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=9 ttl=51 time=270 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=10 ttl=51 time=316 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=11 ttl=51 time=349 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=12 ttl=51 time=111 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=13 ttl=51 time=137 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=14 ttl=51 time=129 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=15 ttl=51 time=140 ms
64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=16 ttl=51 time=134 ms
^C
--- svn.gnome.org ping statistics ---
16 packets transmitted, 16 received, 0% packet loss, time 15529ms
rtt min/avg/max/mdev = 110.175/162.690/349.575/74.103 ms
behdad:pango 0 (harfbuzz-ng*)$ ping git.kernel.org
PING git.us.kernel.org (204.152.191.40) 56(84) bytes of data.
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=1 ttl=53 time=72.0 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=2 ttl=53 time=83.3 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=3 ttl=53 time=70.7 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=4 ttl=53 time=313 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=5 ttl=53 time=79.1 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=6 ttl=53 time=103 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=7 ttl=53 time=85.8 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=8 ttl=53 time=98.9 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=9 ttl=53 time=67.5 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=10 ttl=53 time=87.9 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=11 ttl=53 time=82.9 ms
64 bytes from git2.kernel.org (204.152.191.40): icmp_seq=12 ttl=53 time=87.1 ms
^C
--- git.us.kernel.org ping statistics ---
12 packets transmitted, 12 received, 0% packet loss, time 11491ms
rtt min/avg/max/mdev = 67.544/102.744/313.462/64.357 ms



On 05/05/2009 05:33 PM, Joe Shaw wrote:
Hi,

On Tue, May 5, 2009 at 5:26 PM, Simos Xenitellis
<simos lists googlemail com>  wrote:
The comparison between svn.gnome.org and git.gnome.org is two-fold,

1. It shows that with svn.gnome.org, developers from Europe have been
spoiled with very fast checkouts,
thus they are likely to consider git clones to be slow in any case.

Interestingly, I am on Comcast in Cambridge, MA, USA and git.gnome.org
has twice the ping time as well:

PING svn.gnome.org (91.189.93.3): 56 data bytes
64 bytes from 91.189.93.3: icmp_seq=0 ttl=48 time=86.194 ms
64 bytes from 91.189.93.3: icmp_seq=1 ttl=48 time=83.672 ms
64 bytes from 91.189.93.3: icmp_seq=2 ttl=48 time=94.215 ms
64 bytes from 91.189.93.3: icmp_seq=3 ttl=48 time=86.731 ms
64 bytes from 91.189.93.3: icmp_seq=4 ttl=48 time=86.210 ms

PING git.gnome.org (209.132.176.202): 56 data bytes
64 bytes from 209.132.176.202: icmp_seq=0 ttl=48 time=165.540 ms
64 bytes from 209.132.176.202: icmp_seq=1 ttl=48 time=165.327 ms
64 bytes from 209.132.176.202: icmp_seq=2 ttl=48 time=165.998 ms
64 bytes from 209.132.176.202: icmp_seq=3 ttl=48 time=163.998 ms
64 bytes from 209.132.176.202: icmp_seq=4 ttl=48 time=165.434 ms

(This is a fraction of a longer run; the times are pretty steady.)

More samples from contributors would be necessary to say definitively,
but it would seem as though svn.g.o has a "better" net connection than
git.g.o.  If the size of git downloads is pretty uniformly larger than
svn's (we know it is for initial checkouts, I'm not sure about
updates) then we're taking a hit from both sides and the pain is
compounded.

Joe
_______________________________________________
gnome-infrastructure mailing list
gnome-infrastructure gnome org
http://mail.gnome.org/mailman/listinfo/gnome-infrastructure



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]