Wiki
Clone wikigs2 / GS2_Development_Meetings
GS2 Development Meetings
2019
2018
The Cosener's House, Abingdon, 6th-7th August 2018
2016
GS2 Development Meeting: Oxford, 4 May 2016
This meeting brought together developers and users from the UK, the US and Europe to discuss new physics capabilities and code features in GS2 and to flag possible algorithmic and structural improvements that would be useful.
[https://bitbucket.org/gyrokinetics/wikifiles/raw/master/... Agenda for the day].
New capabilities featured:
Treatment of up-down asymmetric flux surfaces within the local (Miller) framework: Justin Ball
Inclusion of non-Maxwellian equilibrium distribution functions: George Wilkie
Radial profile variation and vpar-mu coordinate representation: Michael Barnes
2014
GS2 Development Meeting: CCFE, 26--27 November 2014
This meeting was to: review substantial recent progress from software development for GS2 and its sister codes TRINITY and GS3; pull together running projects; set priorities for coming months; and minimise the duplication of effort. One major discussion thread was on maximising GS2's scaling performance, which is motivated by the need for more demanding physics simulations. Meeting Summary.
Presentations:
- Objectives,Scope of the Meeting: C M Roach
- GS2 Resolution Considerations: Ion Scales in MAST to Multiscale: F van Wyk
- Effect of electron-ion collisionality on ETG turbulence: G Colyer
- Overview of recent GS2 Optimisations and Future Plans: D Dickinson
- eCSE Project to Optimise Field Solve in GS2: A Jackson
- Exploiting Shared Memory to Extend the Scalability of GS2 L Anton
- Hybridisation: Adding OpenMP to MPI in GS2: A Jackson
- Xeon Phi and Experience from Porting GS2: A Jackson
- GS2's New Eigenvalue Solver: D Dickinson
- New Diagnostics Module, Unit Tests, and Future Plans: E G Highcock
- GS3 Extension to GS2 for Low Flows with Global Effects: M Barnes
- TRINITY/GS2 Optimisation: H Leggate
- Simulating transport in MAST using Trinity and GS2: G Colyer
- Speeding up GS2 using a PARAREAL approach?: D Samaddar
2013
GS2 Development Meeting: Oxford, 17 December 2013
This meeting focussed on discussing Lucian Anton's progress on exploiting “shared memory segments”, and David Dickinson's impressive recent optimisations of GS2's implicit field solve and initialisation routines. Here is a summary.
Presentations:
- Optimising the Field Solve Routines in GS2: D Dickinson
GS2 Development Meeting: CCFE, 13 September 2013
This meeting discussed progress since May, mainly on the OpenMP MSc project at EPCC, but also covering gs2's status on BlueJoule, module tests, and the possibility of exploiting SHM. Here is a summary of the presentations and discussion.
Presentations included:
- Introduction: C M Roach
- Hybrid GS2 using OpenMP: A Jackson
Recent and Planned GS2 Development Projects: Meeting at CCFE, 21st May 2013
This meeting reviewed the current status and plans for various GS2 development projects. Highlights included a number of optimisations that improve GS2's scaling performance, and the new explicit algorithm based on a Discontinuous Galerkin approach. See this summary of the main presentations and discussions.
Presentations:
- Introduction: C M Roach
- Towards Multiscale Simulations: F van Wyk
- Collision Physics Fixes : G Colyer
- Single GS2 Restart File and Trinity Load Balancing: H Leggate
- Optimised Redistributes, next bottlenecks: A Jackson
- Recent Improvements to GS2's Scaling: collisions and v-space integrals: D Dickinson
- Status of New Explicit DG algorithm for GS2: P J Knight
- Exploiting OpenMP to Improve Scaling Performance: L Anton
Updated