Transcript Gian Luigi Granato

GALSYNTH
a WEB interface for GRASIL
by
Lorenzo Paoletti
Amedeo Petrella
Danilo Selvestrel
(INAF-PD)
Speaker Gian Luigi Granato
(INAF-PD)
GRA(phite)SIL(icate) in a nutshell
(Silva et al 98, Silva 99, Granato et al 00, Bressan et al 02, Panuzzo et al
03, Vega et al 05)
3 dusty environments: dense phase ISM (star forming Molecular Clouds), diffuse
phase (cirrus)  clumping of stars and dust, dusty envelopes of AGB stars
Stars are born within MCs and gradually escape as a function of their age  agedependent extinction
Stars and dust are distributed in a bulge (King profile) + disk (double
exponential profile) geometry
Radiative transfer exactly solved in
MCs, with approximation in the cirrus
Dust includes big grains, very small
grains and PAHs. The emission is
appropriately computed for each
component
UV-to radio SEDs (continuum, nebular
lines, spectral indexes)
Motivations to run GRASIL
GRASIL is a complex model with several
physical and algorithmic parameters, thus
difficult to use quickly, and time consuming.
Much handier to use rough prescriptions.
BUT are they sufficient in general?
Why a good model is compulsory?
just an example with GALFORM SAM (Durham):
850 m counts predicted by
GALFORM with GRASIL and a with
a relatively simple prescription
calibrated to match GRASIL
Same at 25!
GALFORM+GRASIL vs SWIRE LFs
Babbedge et al in prep.
Simple prescription may be calibrated (with
observations or good models) to work under
certain conditions, but not in general!
Reliability is unpredictable.
vital to have a WEB interface for GRASIL to
•guide the user
•allow use without compilation/installation
•provide computing power
GALSYNTH
(one of our task in MAGPOP proposal..)
GALSYNTH
the web interface
• After login, users access their own workspaces
• Interactive editing of parameters, with explanations
and bound checking (3 levels)
• Sets of parameters can be saved, modified, reused.
Sets include the possibility of iteration over
parameters (many models)
• The user defines jobs that will be executed on a pool
of application servers (at present 3 PCs, easy to add
more)
• When job completed, user is notified and can
download the results of all models packed in one
archive (which remains also on the WEB server)
Registration and login
Users access their own workspaces
SETs are sets of input parameters for the codes (CHE sets for
CHE_EVO or PAR sets for GRASIL), including definitions of
possible iterations on parameters
Jobs are batch executions of CHE_EVO+GRASIL sequence
according to selected SETs
SET definition I
Three Levels, the higher the more parameter to set and the
greater the allowed range
SET definition II
SET defined
JOB creation
Job Status
Jobs list and results donwnload
Naming of files
The .tar.gz output files of a job consisting of iterations on n
parameters, contains a set of files for each specific GRASIL
model, with names in the format
<par name 1>@<par value 1>_…. <par name n>@<par value>_.<extension>
Example (job with iteration on etastart, mmolfraz and tgal) :
[email protected][email protected]_tgal@7_.spe
Results I…
Result II…
A completely imaginary galaxy SED at 1,3,5,7,9,11 Gyr
NB: GALSYNTH (by now) exploits only part of the
possibilities of GRASIL…
GALSYNTH
a WEB interface to run immediately GRASIL
http://web.pd.astro.it/galsynth
• Anyone can register and be a
user
• Users have their workspaces
• Interactive editing of parameters,
with explanations and bound
checking, including iterations
• Sets of parameters can be saved,
modified, reused.
• Batch jobs are executed on a pool
of our computers
• When job completed, user is
notified and downloads models