New URL for NEMO forge!   http://forge.nemo-ocean.eu

Since March 2022 along with NEMO 4.2 release, the code development moved to a self-hosted GitLab.
This present forge is now archived and remained online for history.
WorkingGroups/HPC (diff) – NEMO

Changes between Version 2 and Version 3 of WorkingGroups/HPC


Ignore:
Timestamp:
2014-11-03T15:12:25+01:00 (9 years ago)
Author:
smasson
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • WorkingGroups/HPC

    v2 v3  
    2929* define a simple configuration (with IO and complex geometry) that will serve as a proof of concept for validating the proposed approach for the future system. 
    3030 
     31== Some ideas...:[[BR]] == 
     32A strong improvement of NEMO scalability is needed to be able to take advantage of the new machines. This means a deep review/rewrite of NEMO code at some point in the futur (beyond 5 years). At the same time, we already know that CMIP7 won't use an ocean model that has not been strongly tested and validated and will stick to a NEMO model close to the existing one. [[BR]] 
     33This means that we need to: 
     34 1) keep improving the current structure of NEMO so it works quite efficiently for almost 10 more years (until the end of CMPI7). [[BR]] 
     35 2) start to work on a new structure that would be tested from year 5 and fully tested and validated for CMIP8 in about 10 years. [[BR]] 
     36 
     37Based on this, we propose to divide the work according to 3 temporal window [[BR]] 
     38 
     39'''0-3 years''': improvements with existing code: Do basic optimisation work: [[BR]] 
     40 1) reduce the number of communications: remove useless communications (a lot), do less and bigger communications (group communications, use large halo) [[BR]] 
     41 2) check code vectorization (AVX instructions for example) [[BR]] 
     42 3) reduce computations over halo (avoid systematic computation from 1 to jpi) [[BR]] 
     43 
     44'''0-5 years''': improvements through the introduction of OpenMP:  [[BR]] 
     45 work initialed by CMCC.  
     46 "Basic" implementation such as tiling may be efficient with many cores processors?  
     47 test different way to find new sources of parallelism. 
     48 
     49'''beyond 5 years''': [[BR]] 
     50 GungHo or not GungHo, that is the question... 
     51 
    3152 
    3253== Agenda:[[BR]] ==