Core Minutes 7/21/2015
Core Minutes 7/21/2015
ScienceTools: (Jim) met with the Data Relase Sterring committe yesterday. Possible post-Pass 8 FSSC releases were proposed: one a bug fix release, another to incorporate a later version of ROOT. The latter is needed in order to properly support plotting on the Mac. [See more about externals under Core Talk: Upgrading externals.]
New tags of Likelihood and pyLikelihood include Eric Charles' gttscube application. In the next couple months Jim will work on refactoring Likelihood and optimizers code to include, among other things, Eric's healpix-based analysis and modifications to the Likelihood calculation requested by Jean.
FSSC: (Joe) No news beyond what was mentioned in Jim's report. He would like to know how to proceed with tip. (Jim) It runs as is, but he would like to get rid of work-around code which used to be required because of defects in the old cfitsio external.
Reprocessing: (Tom) Reprocessing for Pass8, including catch-up runs, is complete.
L1 has seen occasional set faults in Pass 8 code.
Concerning the request to reinstate diffuse response columns—there are two options:
Hardware: (Tom) The user disk moved to gpfs a couple weeks ago. There was a hiccup recently when a battery died (since repaired) but no complaints from users. The next step will be to move the contents of wain32 (MC/RSP/Reprocessing) and wain26 (RM executables) to gpfs.
There is not much monitoring of the new disk compared to the old. We're informed of catastrophic failures, but don't get enough information to see problems develop before they become critical. This is topic is under discussion with Yemi.
(Richard) The Computer Center would like to get 8 new tape drives and they want Fermi to pay for two of them (at $16K apiece). We had not budgeted for such a purchase; we will need to get creative, possibly by deferring purchase of some disk storage.
Core talk:
Upgrading externals (Heather) would like to upgrade all 4 outstanding externals (ROOT, Python, ape, cfitsio) in one go if possible, or at the very least do ROOT and Python together, since building the new Python will force a rebuild of ROOT. (Jim) ape is ready to go. He suggests making cfitsio available in LATEST so that remaining issues there will get addressed. See GRINF-67.
Using FSSC VMs (Jim) discussed with Jeremy the possibility of getting FSSC-prepared Virtual Box VMs for ubuntu and other platforms so that he can test and debug code for those OSes. It would be even better to integrate the VMs with RM so that regular auomatic testing on those platforms could be done.
FSSC Test Suite (Jim) To get this going at SLAC, he needs a workaround from Joe to avoid the requirement to be root during the process.
previous
|
minutes index
|
next
 |
J. Bogart, Last Modified: