Minutes of the Physics Departmental Computing Meeting: 2/11/2009
Present: -------- David Colling, Raymond Beuselinck (HEP), Robert Kingham (Plasma), Stefan Scheel (QOLS), James Spencer (CMTH), Carl Paterson (Photonics), Elizabeth Lucek (2nd Year Computing), John Conway (Learning Technology), Stephen Quirke, Adrian Wong, Adrian Mannall (ICT) General summary: ---------------- In the absence of a specific agenda representatives of each group present gave a brief summary of activity and computing requirements within their section with a view to identifying common areas of interest. The main issue to emerge concerned provision of software licences in a cost effective manner, in particular for Mathematica. There were three main action items to emerge from the meeting for consideration: 1. Mathematica licensing 2. Undergraduate computing project servers 3. Distribution list for the Daily ICT bulletin For details on the action items see the earlier message from David Colling. Adrian Mannall gave a report of news from ICT. This has already been forwarded to this mailing list. It was agreed to aim for future meetings roughly one per term. Reports from the groups: ------------------------ Plasma: The group are heavy users of HPC resources and have contributed 500 cores to the system in exchange for a dedicated queue. It is felt that the documentation of the HPC system is inadequate. There is also a small cluster of machines in the level 1 Huxley computer room. The group runs a web server for data hosting. There are 14 Matlab licences used in the plasma group. QOLS: There are four sub-groups within QOLS, each with their own systems. They are not using HPC yet. The CCM group are collecting huge amounts of data (Terabyte datasets) and are using a commercial backup service in Germany which costs 49 Euros per Tb per month; this is cheaper than ICT can provide. CMTH: The group has a half-time sysadmin. There are about 40 Linux workstations and a couple of very old clusters in the group. CMTH are heavy HPC users. They have licences for Maple, Mathematica (~5) and Intel compilers. They run their own web server and NFS server. They currently do their own tape backups. Photonics: The group is running a mixture of operating systems. They are not using HPC very much. There is an ageing cluster of 36 machines. There are a number of computers running equipment which results in old operating systems tied to old hardware. There is an NFS3 server and lots of ad hoc systems behind a local firewall. There are number of web servers and blog servers running. There are some Matlab users and some use of Mathematica. Due to the high costs of licensing for large clusters there is increasing use of Python for large computations. The group has an increasing need for data storage. HEP: The group has a mixture of Linux and Windows desktops. There is general purpose Linux cluster for group use, in addition to the Grid resources which form a large part of the London Tier 2 located in our computer room in level 1 Blackett. The group has a full time system manager and is in the process of hiring an additional system manager to support the Grid resources. The group runs its own web server and mail server. There are two backup servers for local accounts and essential systems. On the Grid side we have about 700Tbytes of storage locally, but none of it is unique data. The group's most notable requirement is for high network throughput to external sites for Grid traffic. The network link to the computer room has been upgraded to 10Gb. Teaching: There was some discussion of what we should be teaching the undergraduates. We currently teach C++ and this remains popular with the students as it is perceived to be a valuable addition to their CV. Whether we should be using a Windows environment for this is less obvious. Most research computation is done on Unix based systems. The possibility of running Linux in a virtual machine on the teaching PCs was suggested. Provision of dedicated servers for long running undergraduate projects was also discussed; see the earlier summary of actions circulated by David Colling. John Conway distributed copies of information on the various services available to the department in support of teaching activities. In particular the availability of scripting tools to automate tasks such as collating course marks and generating a variety of reports and statistics was discussed.
Quick question... I was unable to attend this meeting as I was already committed to something else. I note that there are various reports from research groups, but not all groups are represented. Am I meant to report from astro as well as first year lab? If so, I can supply a report at some future date. But what about other groups not represented - SPAT, Theory etc.? I guess I'm not yet sure what this committee is meant to be doing... Cheers! Dave
Not all groups were able to attend the meeting. The comments in the minutes were from those that did attend; any inaccuracies are my own. The main intention was to see whether any common interests would emerge, such as trying to obtain cost effective licensing deals on a departmental basis for Mathematica. You final comment pretty much sums up the intent of the first meeting...to decide what we are meant to be doing. Ray On Sat, 14 Nov 2009, Dave Clements wrote:
Quick question...
I was unable to attend this meeting as I was already committed to something else. I note that there are various reports from research groups, but not all groups are represented. Am I meant to report from astro as well as first year lab? If so, I can supply a report at some future date. But what about other groups not represented - SPAT, Theory etc.?
I guess I'm not yet sure what this committee is meant to be doing...
Cheers!
Dave
Hi dave, The full list of people who are on the committee and so should also be on the mailing list are (in no particular order): Adrian Mannall ICT Steve Rochford ICT Carl Paterson Photonics David Clements 1st year computing David Colling HEP (Chair) Elizabeth Lucek 2nd year computing James Spencer Condensed Matter Theory Jing Zhang EXSS Section Ray Beuselinck /Kostas Georgiou HEP Peter Torok Photonics Robert Kingham Plasma Physics Stefan Scheel Quantum Optics and laser science Simon Bland Plasma Carlo Contaldi/Tim Evans Theory John Tovey Astro Stephen Quirke/Adrian Wong ICT Harry Vine Departmental Services Manager Simon Bland was unwell and sent in a mail and you will notice that we seriously discussed your thoughts on moving away from MS C++. This committee is still in the process of revival, but hopefully will grow back into being a useful body. All the best, david Raymond Beuselinck wrote:
Not all groups were able to attend the meeting. The comments in the minutes were from those that did attend; any inaccuracies are my own. The main intention was to see whether any common interests would emerge, such as trying to obtain cost effective licensing deals on a departmental basis for Mathematica.
You final comment pretty much sums up the intent of the first meeting...to decide what we are meant to be doing.
Ray
On Sat, 14 Nov 2009, Dave Clements wrote:
Quick question...
I was unable to attend this meeting as I was already committed to something else. I note that there are various reports from research groups, but not all groups are represented. Am I meant to report from astro as well as first year lab? If so, I can supply a report at some future date. But what about other groups not represented - SPAT, Theory etc.?
I guess I'm not yet sure what this committee is meant to be doing...
Cheers!
Dave
_______________________________________________ Physics-Departmental-Computing mailing list Physics-Departmental-Computing@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/physics-departmental-computing
And Richard Bantges for SPAT... -----Original Message----- From: physics-departmental-computing-bounces@imperial.ac.uk [mailto:physics-departmental-computing-bounces@imperial.ac.uk] On Behalf Of David Colling Sent: 14 November 2009 16:58 To: Beuselinck, Raymond Cc: physics-departmental-computing Subject: Re: [Physics-Departmental-Computing] Minutes of the Physics Departmental Computing Meeting: 2/11/2009 Hi dave, The full list of people who are on the committee and so should also be on the mailing list are (in no particular order): Adrian Mannall ICT Steve Rochford ICT Carl Paterson Photonics David Clements 1st year computing David Colling HEP (Chair) Elizabeth Lucek 2nd year computing James Spencer Condensed Matter Theory Jing Zhang EXSS Section Ray Beuselinck /Kostas Georgiou HEP Peter Torok Photonics Robert Kingham Plasma Physics Stefan Scheel Quantum Optics and laser science Simon Bland Plasma Carlo Contaldi/Tim Evans Theory John Tovey Astro Stephen Quirke/Adrian Wong ICT Harry Vine Departmental Services Manager Simon Bland was unwell and sent in a mail and you will notice that we seriously discussed your thoughts on moving away from MS C++. This committee is still in the process of revival, but hopefully will grow back into being a useful body. All the best, david Raymond Beuselinck wrote:
Not all groups were able to attend the meeting. The comments in the minutes were from those that did attend; any inaccuracies are my own. The main intention was to see whether any common interests would emerge, such as trying to obtain cost effective licensing deals on a departmental basis for Mathematica.
You final comment pretty much sums up the intent of the first meeting...to decide what we are meant to be doing.
Ray
On Sat, 14 Nov 2009, Dave Clements wrote:
Quick question...
I was unable to attend this meeting as I was already committed to something else. I note that there are various reports from research groups, but not all groups are represented. Am I meant to report from astro as well as first year lab? If so, I can supply a report at some future date. But what about other groups not represented - SPAT, Theory etc.?
I guess I'm not yet sure what this committee is meant to be doing...
Cheers!
Dave
_______________________________________________ Physics-Departmental-Computing mailing list Physics-Departmental-Computing@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/physics-departmental-computing
_______________________________________________ Physics-Departmental-Computing mailing list Physics-Departmental-Computing@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/physics-departmental-computing
My apologies Richard, I had added you to the list but not the document. Now corrected. Bantges, Richard J wrote:
And Richard Bantges for SPAT...
-----Original Message----- From: physics-departmental-computing-bounces@imperial.ac.uk [mailto:physics-departmental-computing-bounces@imperial.ac.uk] On Behalf Of David Colling Sent: 14 November 2009 16:58 To: Beuselinck, Raymond Cc: physics-departmental-computing Subject: Re: [Physics-Departmental-Computing] Minutes of the Physics Departmental Computing Meeting: 2/11/2009
Hi dave,
The full list of people who are on the committee and so should also be on the mailing list are (in no particular order):
Adrian Mannall ICT Steve Rochford ICT Carl Paterson Photonics David Clements 1st year computing David Colling HEP (Chair) Elizabeth Lucek 2nd year computing James Spencer Condensed Matter Theory Jing Zhang EXSS Section Ray Beuselinck /Kostas Georgiou HEP Peter Torok Photonics Robert Kingham Plasma Physics Stefan Scheel Quantum Optics and laser science Simon Bland Plasma Carlo Contaldi/Tim Evans Theory John Tovey Astro Stephen Quirke/Adrian Wong ICT Harry Vine Departmental Services Manager
Simon Bland was unwell and sent in a mail and you will notice that we seriously discussed your thoughts on moving away from MS C++.
This committee is still in the process of revival, but hopefully will grow back into being a useful body.
All the best, david
Raymond Beuselinck wrote:
Not all groups were able to attend the meeting. The comments in the minutes were from those that did attend; any inaccuracies are my own. The main intention was to see whether any common interests would emerge, such as trying to obtain cost effective licensing deals on a departmental basis for Mathematica.
You final comment pretty much sums up the intent of the first meeting...to decide what we are meant to be doing.
Ray
On Sat, 14 Nov 2009, Dave Clements wrote:
Quick question...
I was unable to attend this meeting as I was already committed to something else. I note that there are various reports from research groups, but not all groups are represented. Am I meant to report from astro as well as first year lab? If so, I can supply a report at some future date. But what about other groups not represented - SPAT, Theory etc.?
I guess I'm not yet sure what this committee is meant to be doing...
Cheers!
Dave
_______________________________________________ Physics-Departmental-Computing mailing list Physics-Departmental-Computing@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/physics-departmental-computing
_______________________________________________ Physics-Departmental-Computing mailing list Physics-Departmental-Computing@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/physics-departmental-computing
Hi, Actually, it was Richard I meant in my earlier message not Simon at all, I have no idea why I said Simon. Working on a Saturday evening is clearly error prone! All the best, david
Bantges, Richard J wrote:
And Richard Bantges for SPAT...
-----Original Message----- From: physics-departmental-computing-bounces@imperial.ac.uk [mailto:physics-departmental-computing-bounces@imperial.ac.uk] On Behalf Of David Colling Sent: 14 November 2009 16:58 To: Beuselinck, Raymond Cc: physics-departmental-computing Subject: Re: [Physics-Departmental-Computing] Minutes of the Physics Departmental Computing Meeting: 2/11/2009
Hi dave,
The full list of people who are on the committee and so should also be on the mailing list are (in no particular order):
Adrian Mannall ICT Steve Rochford ICT Carl Paterson Photonics David Clements 1st year computing David Colling HEP (Chair) Elizabeth Lucek 2nd year computing James Spencer Condensed Matter Theory Jing Zhang EXSS Section Ray Beuselinck /Kostas Georgiou HEP Peter Torok Photonics Robert Kingham Plasma Physics Stefan Scheel Quantum Optics and laser science Simon Bland Plasma Carlo Contaldi/Tim Evans Theory John Tovey Astro Stephen Quirke/Adrian Wong ICT Harry Vine Departmental Services Manager
Simon Bland was unwell and sent in a mail and you will notice that we seriously discussed your thoughts on moving away from MS C++.
This committee is still in the process of revival, but hopefully will grow back into being a useful body.
All the best, david
Raymond Beuselinck wrote:
Not all groups were able to attend the meeting. The comments in the minutes were from those that did attend; any inaccuracies are my own. The main intention was to see whether any common interests would emerge, such as trying to obtain cost effective licensing deals on a departmental basis for Mathematica.
You final comment pretty much sums up the intent of the first meeting...to decide what we are meant to be doing.
Ray
On Sat, 14 Nov 2009, Dave Clements wrote:
Quick question...
I was unable to attend this meeting as I was already committed to something else. I note that there are various reports from research groups, but not all groups are represented. Am I meant to report from astro as well as first year lab? If so, I can supply a report at some future date. But what about other groups not represented - SPAT, Theory etc.?
I guess I'm not yet sure what this committee is meant to be doing...
Cheers!
Dave
_______________________________________________ Physics-Departmental-Computing mailing list Physics-Departmental-Computing@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/physics-departmental-computing
Physics-Departmental-Computing mailing list Physics-Departmental-Computing@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/physics-departmental-computing
_______________________________________________ Physics-Departmental-Computing mailing list Physics-Departmental-Computing@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/physics-departmental-computing
Hi Ray, I was also unable to attend this meeting, but I did send David Colling a summary of SPAT's current computing prior to the meeting, which I understand was not shared with the rest of the people at the meeting. I've attached this below, in case anyone is interested: Hi David, Apologies, but I won't be able to attend this first meeting as I'm recovering from the flu (I've cc'ed Ralf Toumi just in case he would like to come along). I would be very interested to hear about what is discussed at the meeting, so if I could arrange to come and have a chat with you later in the week, or if a summary is provided to the list after the meeting, then that would be appreciated. I have a couple of issues / questions that I would like to voice: 1. SPAT has an excess of ITTVIS IDL licences (we have 250 for SPAT, which we realistically only use about 100 of at any one time). Would other groups in the Department be interested in getting access to our IDL licences (for a small price)? 2. Can we persuade ICT to allow the reps access to the PXE network installation images for Linux and Windows? 3. What are the future plans for network infrastructure improval - moving to 1Gb/s throughout Huxley and Blackett. 4. Plans for the Departmental Server Room (H147) - is this facility to remain for the foreseeable future? Are there any plans for this room that we should know about? I'll try and summarise SPAT's computing: - Series of Suns (running SOLARIS 10 SPARC), Workstations running the College Redhat Linux ES4 (64bit & 32bit), and various Windows PC and Macs. - Main use of computing facilities [1] archiving and analysis of satellite data sets [~12TB storage in H147 - NAS devices] [2] numerical modelling of the Earth's atmosphere / ocean / climate / solar wind / planetary magnetic fields [3] space craft operations (Cassini, Cluster, GERB) - Additional computing usage: College HPC (climate models, coupled ocean-atmosphere models, etc.) - Sun computers have about 4TB of high performance disks - backed up offsite (in Mech Eng using our own backup device) There are probably lots of other things that I've missed. I'm very sorry to miss this first meeting. Rich -----Original Message----- From: physics-departmental-computing-bounces@imperial.ac.uk [mailto:physics-departmental-computing-bounces@imperial.ac.uk] On Behalf Of Raymond Beuselinck Sent: 14 November 2009 15:18 To: Clements, David L Cc: physics-departmental-computing Subject: Re: [Physics-Departmental-Computing] Minutes of the Physics Departmental Computing Meeting: 2/11/2009 Not all groups were able to attend the meeting. The comments in the minutes were from those that did attend; any inaccuracies are my own. The main intention was to see whether any common interests would emerge, such as trying to obtain cost effective licensing deals on a departmental basis for Mathematica. You final comment pretty much sums up the intent of the first meeting...to decide what we are meant to be doing. Ray On Sat, 14 Nov 2009, Dave Clements wrote:
Quick question...
I was unable to attend this meeting as I was already committed to something else. I note that there are various reports from research groups, but not all groups are represented. Am I meant to report from astro as well as first year lab? If so, I can supply a report at some future date. But what about other groups not represented - SPAT, Theory etc.?
I guess I'm not yet sure what this committee is meant to be doing...
Cheers!
Dave
_______________________________________________ Physics-Departmental-Computing mailing list Physics-Departmental-Computing@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/physics-departmental-computing
participants (4)
-
Bantges, Richard J
-
Dave Clements
-
David Colling
-
Raymond Beuselinck