Dear David,
I was working at Imperial with Janusz and I know his work.
The problem about develop software for BaBar is we have to provide at least
the same level of reliability, flexibility and functionality of SLAC. Years
ago users tried to use Alibabar and Ganga, it does not work and users went
back to slac.
The system must be in command line mode to allow integration with user's
scripts, it must be based in sequence number from bookkeeping metadatada
(never handlers), and it must provide a set of functions independently and
not be heavy.
In my point of view, users are quite good programming, and the approach
adopted in EasyGrid using a framework (like the experiment software) is the
best approach to achieve our requirements.
James
>From: David Britton <[log in to unmask]>
>To: Karl Harrison <[log in to unmask]>
>CC: [log in to unmask], [log in to unmask]
>Subject: Re: Provoking?
>Date: Fri, 02 Sep 2005 09:49:21 -0400
>
>Dear Karl, James,
>
>A couple of years ago a Ganga front end to BaBar was provided by Janusz
>Martyniak at Imperial ([log in to unmask]). Janusz is still
>around (working on GridCC) so you could contact him for any insight. I
>don't know why this wasn't taken further but it is possible that it's time
>has now come (Ganga has come a long way design-wise; and BaBar have make
>great progress too)? Notionally, this would be a great way to proceed and
>would certainly qualify as a highlight success.
>
>Regards, Dave.
>
>Karl Harrison wrote:
>
>>Dear James,
>>
>> I found your web pages interesting, although it looks as if EasyGrid
>>is trying to the same as Ganga, but limited to BaBar. Ganga provides
>>a generic job definition/submission framework, into which
>>extensions specific to a particular experiment/user are easily added.
>>The EasyGrid instructions contain a lot of "in file x change lines y to
>>z", which wouldn't be needed in Ganga, where configurable parameters are
>>placed in evidence in application/backend schemas.
>>
>> Is there any chance of you having a look at Ganga
>>(http://ganga.web.cern.ch/ganga), to see if it's potentially useful
>>for BaBar? The most recent paper is the one submitted to AHM05:
>> http://ganga.web.cern.ch/ganga/documents/pdf/ganga_ahm05.pdf
>>or to try it out, you could have a look at:
>> http://ganga.web.cern.ch/ganga/user/v4/CLIP/
>>(I'm afraid that this only describes version 4.0.0-beta1, and for
>>LCG submission you'll need 4.0.0-beta3, to be released next week.)
>>
>>Best wishes,
>>
>>Karl.
>>
>>===========================================================================
>>
>>On Fri, 2 Sep 2005, James Werner wrote:
>>
>>
>>
>>>Dear David,
>>>I would appreciate very much if you write in "Aplications/ 10 Integration
>>>with
>>>running experiments" about my web page:
>>>
>>>HEP analysis, Grid and EasyGrid Job Submission Prototype: Babar/CM2
>>>showcase
>>>http://www.hep.man.ac.uk/u/jamwer/
>>>
>>>This page contains the development of EasyGrid (a job submission system
>>>for
>>>LCG) that has all necessary modules to do HEP analysis using grid. A
>>>babar
>>>software tutorial is also provided.
>>>EasyGrid especification is available at
>>>http://www.hep.man.ac.uk/u/jamwer/#sec9 and user manual at
>>>http://www.hep.man.ac.uk/u/jamwer/userman.html
>>>
>>>There are 3 benchmarks:
>>>1. Discrimination of Pions and Kaons in the electromagnetic calorimeter:
>>>18
>>>million events in 26 jobs, 10 jobs in parallel.
>>>2. Pi 0 reconstruction: see
>>>http://www.hep.man.ac.uk/u/jamwer/pi0alg5.html
>>>for a very interesting analysis. 500 million events running in 200 jobs
>>>in
>>>several sites (Manchester testbed, Manchester production, and RAL). I
>>>generated 6 million monte carlo eventsat Manchester production. The
>>>parametric
>>>model was Breit-Wigner invariant mass model.
>>>3. Searching for deuterons: this project was a stress test for grid. see
>>>http://www.hep.man.ac.uk/u/jamwer/deutdesc.html
>>>The first run was 80 jobs with 20 million events each. The second was 800
>>>jobs
>>>with 2 million events each.
>>>
>>>Grid evaluation can be found at http://www.hep.man.ac.uk/u/jamwer/#sec15
>>>
>>>Best regards,
>>>James
>>>
>>>
>>>
>>>>From: David Britton <[log in to unmask]>
>>>>Reply-To: David Britton <[log in to unmask]>
>>>>To: [log in to unmask]
>>>>Subject: Provoking?
>>>>Date: Thu, 1 Sep 2005 14:45:33 -0400
>>>>
>>>>Dear Colleagues,
>>>>
>>>>At the last Oversight Committee meeting, GridPP was asked to document
>>>>the
>>>>"Value added or Delivered by the Project". Although this seems like
>>>>another
>>>>piece of make-work, presumably its value is that it can be passed up
>>>>the
>>>>political food-chain and is an opportunity for us to define our
>>>>successes.
>>>>Of course, being at the bottom of the food-chain, it has landed on my
>>>>desk
>>>>or, more accurately, it appeared on the GridPP14 Programme
>>>>( http://www.gridpp.ac.uk/gridpp14/programme.html ) with my name
>>>>attached as
>>>>"Provocateur". So, in order to provoke, I have pre-prepared (off the top
>>>>of
>>>>my head so to speak) a list of 22 possible key things that we might
>>>>document. They are available on the web linked from the relevant item om
>>>>the
>>>>GridPP14 programme, or at:
>>>>http://www.gridpp.ac.uk/pmb/ProjectManagement/GridPPAchievements.htm
>>>>
>>>>I think most members of GridPP will have an opinion about at least a few
>>>>items that they would like to appear in such a list. Please feel free to
>>>>contact me or to contribute during the GridPP14 session, if you feel
>>>>some
>>>>key item is missing. For items that are there, the next step is the
>>>>wording
>>>>(in most cases I have made a first attempt which you can see by clicking
>>>>on
>>>>the relevant item in the list). Again, I will be happy to receive
>>>>comments
>>>>and suggestions on the wording in advance or during the session. The
>>>>final
>>>>step will be to add tertiary information that backs up the higher level
>>>>statements by referencing more specific items (such as code delivered or
>>>>roles held etc).
>>>>
>>>>Regards, Dave.
>>>>
>>>>
>
|