Hi,
A quick note that following Kostas' advice we tried a very small buffer
size, 4kB. This reduced bandwidth consumption even further but cpu
efficiency was still relatively poor (hitting network latency rather
than local RAM). Our advice is still to try the big buffer (128MB)
unless your network is significantly bottle necking.
John
Graeme Stewart wrote:
> Hi All
>
> There will be another atlas analysis test tomorrow, with 300 jobs per site:
>
> http://gangarobot.cern.ch/st/test_63/
>
> Most sites have, I think, followed Liverpool's advice about setting a
> large read ahead on the rfio client (we messed it up last week at
> Glasgow, but it's correct now).
>
> It seems to me that load on the DPM headnode is becoming a limiting
> factor now. If we can demonstrate this as most sites tomorrow then we
> can ask the DPM developers for their comments and hopfully get some
> tuning advice.
>
> Cheers
>
> Graeme
>
--
Dr John Bland, Systems Administrator
Room 220, Oliver Lodge
Particle Physics Group, University of Liverpool
Mail: [log in to unmask]
Tel : 0151 794 2911
"I canna change the laws of physics, Captain!"
|