[petsc-dev] [mpich-discuss] MPICH migration to git

Jed Brown jedbrown at mcs.anl.gov
Mon Mar 11 19:47:04 CDT 2013


On Mon, Mar 11, 2013 at 3:51 PM, Sean Farley <sean at mcs.anl.gov> wrote:

> I think you're conflating the timings here. Let's first isolate network
> speed:
>
> $ time git clone -n https://bitbucket.org/jedbrown/petsc-git-leanpetsc-git
> Cloning into 'petsc-git'...
> remote: Counting objects: 297100, done.
> remote: Compressing objects: 100% (67974/67974), done.
> remote: Total 297100 (delta 228357), reused 297100 (delta 228357)
> Receiving objects: 100% (297100/297100), 41.22 MiB | 873 KiB/s, done.
> Resolving deltas: 100% (228357/228357), done.
>
> real 1m6.176s
> user 0m7.732s
> sys 0m2.283s
>

> $ time hg clone --uncompressed -U
> https://bitbucket.org/petsc/petsc-dev petsc-dev
> streaming all changes
> 10150 files to transfer, 150 MB of data
> transferred 150 MB in 181.4 seconds (847 KB/sec)
>
> real 3m11.839s
> user 0m5.111s
> sys 0m5.308s
>

If your network is slow enough, any system will saturate the network
bandwidth. But if you have a fast network, there can be a big difference.

jedbrown at cg:/sandbox/jedbrown$ time git clone -n
https://bitbucket.org/petsc/petsc
Cloning into 'petsc'...
remote: Counting objects: 490113, done.
remote: Compressing objects: 100% (111365/111365), done.
remote: Total 490113 (delta 376104), reused 490113 (delta 376104)
Receiving objects: 100% (490113/490113), 66.39 MiB | 20.24 MiB/s, done.
Resolving deltas: 100% (376104/376104), done.

*real 0m16.811s*
user 0m15.493s
sys 0m4.100s

jedbrown at cg:/sandbox/jedbrown$ time hg clone --uncompressed -U
https://bitbucket.org/petsc/petsc-dev petsc-hg
streaming all changes
10497 files to transfer, 158 MB of data
transferred 158 MB in 74.6 seconds (2.11 MB/sec)

*real 1m47.693s*
user 0m4.092s
sys 0m1.912s

If we just want 'master' (e.g., building for an app on a production
machine), we can make it even faster.

jedbrown at cg:/sandbox/jedbrown$ time git clone -n --depth 1 --single-branch
https://bitbucket.org/petsc/petsc petsc-master
Cloning into 'petsc-master'...
remote: Counting objects: 5053, done.
remote: Compressing objects: 100% (4609/4609), done.
remote: Total 5053 (delta 918), reused 2195 (delta 361)
Receiving objects: 100% (5053/5053), 9.12 MiB | 5.67 MiB/s, done.
Resolving deltas: 100% (918/918), done.

*real 0m3.220s*
user 0m0.508s
sys 0m0.104s
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20130311/03ac853e/attachment.html>


More information about the petsc-dev mailing list