Changes between Version 1 and Version 2 of WorkingGroups/HPC/Mins_2015_12_04

Timestamp:

2016-03-08T18:44:36+01:00 (8 years ago)

Author:

mikebell

Comment:

--

WorkingGroups/HPC/Mins_2015_12_04

@@ -1 @@
-= Minutes =
+= Minutes Draft minutes NEMO HPC working group: 14:30 - 16:00 CET Friday 4th December =
+
+Present: Lucian Anton, Mike Bell, Tom Bradley, Miguel Castrillo, Marie-Alice Foujols, Matt Glover, Tim Graham, Jason Holt, Dmitry Kuts, Claire Levy, Silvia Mocavero, Oriol Mula-Valls, Andy Porter, Martin Schreiber, Kim Serradell 
+
+Apologies: Monder Chekki, Julien le Sommer, Cyril Mazauric
+
+
+
+1.      Welcome of new members: Lucian Anton, Tom Bradley, Matt Glover, Dmitry Kuts and Andy Porter introduced themselves.  Cyril Mazauric has also just joined the group. 
+2.      Actions from previous meeting not covered by agenda below (10 mins) 
+All done except action 3 which has started, is important, and is ongoing
+
+Action 1: Tim and Sylvia to discuss what specification of this [roofline diagnostics] would be useful and then iterate with Martin, Oriol and Miguel.
+
+3.      Check ownership of tasks on Sebastien’s list (see attachment and list of volunteers below) (5 mins) 
+Ownership of tasks on Sebastien’s list is agreed. The 5th task and Gurvan’s additional task do not have owners. 
+
+4.      Presentation by on Psyclone (Andy Porter)
+Andy gave a very clear presentation of the motivation and concepts involved in Psyclone. Various questions were discussed such as the risks of either adopting Psyclone or not. In principle, Psyclone could help the very diverse range of users of NEMO to obtain reasonable optimisation on a common code. 
+
+5.      Discuss NEMO HPC working document: https://docs.google.com/document/d/1D8LdTPYMmzNK3jW-UOA0VQYJauftbRQiXEqlejXlv0A/edit?usp=sharing 
+The aim of the document is to develop a coherent overview of the objectives and work of the group. It was agreed that it has an important role in helping to define and progress a common strategy based on assessments of the performance of common configurations. As the document grows it may be necessary to introduce more structure through links to other documents. It will need inputs from a number of groups and will help us all to understand the different aspects of the problem.  The document will take some effort to develop and maintain but it is important and will be worth it. 
+
+6.      Update on Copernicus call for H2020 
+The NEMO steering committee has asked a writing committee to propose a set of tasks to include in a NEMO bid before Christmas. Claire Levy is chairing this group. Its members include Mike Bell,
+Jason Holt and Julien le Sommer from this group. The tasks will be prioritised according to their value to Copernicus and their excellence (according to EC criteria). An outline for the scope and rationale for HPC tasks has been drafted and is attached with the minutes. It is very probable that this will be one of the tasks selected to be included in the proposal. Comments on this draft are welcome.     
+
+7.      Agree which benchmark configurations to adopt  
+It was agreed to start with two configurations. The GYRE configuration (at a couple of resolutions) and Barcelona’s ORCA025-LIM3 configuration. These configurations will probably be made available as tar files. 
+
+Action 2:  Miguel, Silvia, Tim and Claire to coordinate the set up and review of these two configurations 
+
+8.      AoB 
+
+Action 3: Mike to put information from group meeting on the NEMO HPC WG wiki page
+
+Action 4: Mike to set up an email circulation list (to which all members can send emails) 
+
+Action 5: Mike to set up a doodle poll for the next meeting between 11 and 22 Jan.  
+