Changes between Version 3 and Version 4 of WorkingGroups/HPC
- Timestamp:
- 2014-11-04T16:07:51+01:00 (9 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
WorkingGroups/HPC
v3 v4 38 38 39 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]] 40 1) reduce the number of communications: remove useless communications (a lot) 41 2) reduce the number of communications: do less and bigger communications (group communications, use large halo) [[BR]] 42 3) introduce asynchronous communications 43 3) check code vectorization (SIMD instructions) [[BR]] 43 44 44 45 '''0-5 years''': improvements through the introduction of OpenMP: [[BR]] … … 46 47 "Basic" implementation such as tiling may be efficient with many cores processors? 47 48 test different way to find new sources of parallelism. 49 OpenMP4, OpenACC? 48 50 49 51 '''beyond 5 years''': [[BR]]