1. INFSO-RI-508833 Enabling Grids for E-sciencE www.eu-egee.org Fusion Status Report Francisco Castejón (francisco.castejon@ciemat.es) CIEMAT. Madrid, Spain.

2. Grid Users Forum, CERN, March 2006 2 Enabling Grids for E-sciencE INFSO-RI-508833 Outline Strategy. Fusion Deployment and VO setup. –The problem of the name Present Applications: Computing in Plasma Physics. Future Applications in the grid. Data storage and handling. Conclusions.

3. Grid Users Forum, CERN, March 2006 3 Enabling Grids for E-sciencE INFSO-RI-508833 Strategy Computing: –Identify common Codes suitable for GRID. (Ongoing) –Adapt codes to the GRID. (Ongoing) –Set up VO (Ongoing) –Production phase. Data handling: Define strategies for data storage. & database organization. Protocol for data Access.

4. Grid Users Forum, CERN, March 2006 4 Enabling Grids for E-sciencE INFSO-RI-508833 ITER: Making decisions in real Time !! Data Acquisition and Storage (Grid, Supercomputers) Data Analysis and Reduction: Artificial Intelligence, Neural Network, Pattern Recognition Simulations: Large codes in different platforms (Grid, Supercomputers) Decision for present/next shot One half an hour shot every one hour and a half: Decisions in real time.

5. Grid Users Forum, CERN, March 2006 5 Enabling Grids for E-sciencE INFSO-RI-508833 ITER Partners Distributed Participation. Data access. Remote Control Rooms?

6. Grid Users Forum, CERN, March 2006 6 Enabling Grids for E-sciencE INFSO-RI-508833 International Tokamak (ITPA) and Stellarator (SIA) collaborations. USA: Alcator C-Mod (MIT) DIII-D (San Diego) NSTX (Princeton) NCSX (Princeton) HSX (Wisconsin) QPS (Oak-Ridge) USA Fusion Grid Russia: T-10 (Kurchatov) Globus (Ioffe) T-11M (TRINITI) L-2 (Gen. Inst. Phys.) EGEE Project EU: JET (EFDA) ASDEX (Ger.) TORE SUPRA (Fran.) MAST (UK) TEXTOR (Ger.) TCV (Switz.) FTU (Italy) W7-X (Ger.) TJ-II (Spain) EGEE Project Japan: JT-60 (Naka) LHD (Toki) CHS (Nagoya) H-J (Kyoto) GRID Project ? China, Brazil, Korea, India: KSTAR (Korea) TCBRA (Bra.) H-7 (China) U2A (China) SST1 (India) EGEE Project

7. Grid Users Forum, CERN, March 2006 7 Enabling Grids for E-sciencE INFSO-RI-508833 PARTNERS and Resources for VO SW Federation: CIEMAT, BIFI, UCM, INTA (Spain) Kurchatov (Russia). Culham Laboratory- UKAEA (UK) KISTI (South Korea). ENEA (Italy). CEA-Cadarache (France). … Experience in using and developing Fusion Applications. Experience in porting applications and developing Grid Technologies.  Connection with EELA (Some fusion partners: Brazil,Mexico,Argentina)  Needed: Join IPP-Max Planck (Germany) and other EFDA Associations.  Also needed: contact with USA, China, Japan,…

8. Grid Users Forum, CERN, March 2006 8 Enabling Grids for E-sciencE INFSO-RI-508833 VO Deployment Present: CIEMAT: 27 KSpecInts; BIFI: 8 KSpecInts; INTA: 6 KSpecInts Resource Broker in BIFI (Spain) VO Manager: I. Campos (BIFI. Spain) Resource Broker in BIFI (Spain) VO Manager: I. Campos (BIFI. Spain) http://grid.bifi.unizar.es/egee/fusion-vo/ http:.ciemat.es/colla http: //www-fusion.ciemat.es/colla boration/egee/ Within less than 6 months: JET: 38 KSpecInts; BIFI: 32 KSpecInts; CEA-Cadarache ? KISTI?, INTA?, ENEA? Within less than 6 months: JET: 38 KSpecInts; BIFI: 32 KSpecInts; CEA-Cadarache ? KISTI?, INTA?, ENEA? Beginning of 2007: JET: 32 additional cores; BIFI: 32 additional cores; CIEMAT ?; CEA-Cadarache ?(second phase already committed). Beginning of 2007: JET: 32 additional cores; BIFI: 32 additional cores; CIEMAT ?; CEA-Cadarache ?(second phase already committed).

9. Grid Users Forum, CERN, March 2006 9 Enabling Grids for E-sciencE INFSO-RI-508833 VO Deployment: The problem of the name. Russian Grid has adopted the same name of Fusion, as we have done. The works that are sent by our resource broker go to such Grid. Our VO deployment is hindered. They should change the name in short term (~1 week). A suitable name: Fusion-RIDG Otherwise: We have to change the name. Consequences on Russian certificates.

10. Grid Users Forum, CERN, March 2006 10 Enabling Grids for E-sciencE INFSO-RI-508833 COMPUTING in the GRID: Present Applications –Applications with distributed calculations: Monte Carlo, Separate estimates, … –Multiple Ray Tracing: e. g. TRUBA. –Stellarator Optimization: VMEC – Transport and Kinetic Theory: Monte Carlo Codes.

11. Grid Users Forum, CERN, March 2006 11 Enabling Grids for E-sciencE INFSO-RI-508833 Multiple Ray Tracing: TRUBA Single Ray (1 PE): Hamiltonian Ray Tracing Equations. Beam Simulation: Bunch of rays with beam waist far from the critical layer (100-200 rays) Bunch of rays with beam waist close to the critical layer (100-200 rays) x (100-200 wave numbers) ~10 5 GRID PROBLEM

12. Grid Users Forum, CERN, March 2006 12 Enabling Grids for E-sciencE INFSO-RI-508833 TRUBA: Multiple Ray Tracing TRUBA for EBW: TRUBA for EBW: -Real geometry in TJ-II:Coming from a supercomputer (VMEC). Real geometry in TJ-II:Coming from a supercomputer (VMEC).Real geometry in TJ-II:Coming from a supercomputer (VMEC). - A single Non-relativistic ray (about 18’). A single Non-relativistic ray (about 18’). A single Non-relativistic ray (about 18’). - A single relativistic ray (about 40’). A single relativistic ray (about 40’). A single relativistic ray (about 40’). - Some problems with Geometry libraries. Some problems with Geometry libraries. Some problems with Geometry libraries. - Ported to the grid using Grid Way (for the moment). Ported to the grid using Grid Way (for the moment). Ported to the grid using Grid Way (for the moment). -See: See: J. L. Vázquez-Poleti. “Massive Ray Tracing in Fusion Plasmas on EGEE”. User Forum, 2006. J. L. Vázquez-Poleti. “Massive Ray Tracing in Fusion Plasmas on EGEE”. User Forum, 2006.

13. Grid Users Forum, CERN, March 2006 13 Enabling Grids for E-sciencE INFSO-RI-508833 Optimised Stellarators QPS and NCSX Supercomputer Optimization NCSXQPS

14. Grid Users Forum, CERN, March 2006 14 Enabling Grids for E-sciencE INFSO-RI-508833 Stellarator optimization in the Grid - A lot of different Magnetic Configurations operating nowadays. OPTIMIZATION NECESITY BASED ON KNOWLEDGE OF STELLARATOR PHYSICS. Every variant computed on a separate processor (~10’) VMEC (Variational Momentum Equilibrium Code) 120 Fourier parameters are varied. Plasma configuration may be optimised numerically by variation of the field parameters.

15. Grid Users Forum, CERN, March 2006 15 Enabling Grids for E-sciencE INFSO-RI-508833 VMEC on Kurchatov GRID LCG-2 - based Russian Data Intensive Grid consortium resources. About 7.500 cases computed (about 1.500 was not VMEC- computable, i.e. no equilibrium). Each case took about 20 minutes. Up to 70 simultaneous jobs running on the grid. Genetic Algorith used to select the optimum case. -See: V. Voznesensky. “Genetic Optimisations in Grid”. User Forum, 2006. LCG-2 - based Russian Data Intensive Grid consortium resources. About 7.500 cases computed (about 1.500 was not VMEC- computable, i.e. no equilibrium). Each case took about 20 minutes. Up to 70 simultaneous jobs running on the grid. Genetic Algorith used to select the optimum case. -See: V. Voznesensky. “Genetic Optimisations in Grid”. User Forum, 2006.

16. Grid Users Forum, CERN, March 2006 16 Enabling Grids for E-sciencE INFSO-RI-508833 Kinetic Transport Following independent particle orbits Montecarlo techniques: Particles distributed according to experimental density and ion temperature profiles (Maxwellian distribution function) SUITABLE PROBLEM FOR CLUSTER AND GRID TECHNOLOGIES

17. Grid Users Forum, CERN, March 2006 17 Enabling Grids for E-sciencE INFSO-RI-508833 Kinetic Transport Example of orbit in the real 3D TJ-II Geometry (single PE). ~1 GBy data, 24 h x 512 PE Distribution function of parallel velocity at a given position (Data Analysis).

18. Grid Users Forum, CERN, March 2006 18 Enabling Grids for E-sciencE INFSO-RI-508833 Kinetic transport No collisions: 0.5 ms of trajectory takes 1 sec. CPU.. No collisions: 0.5 ms of trajectory takes 1 sec. CPU.. Collisions: 1 ms of trajectory takes 4 sec CPU. Collisions: 1 ms of trajectory takes 4 sec CPU. Particle life: 150 - 200 ms. Single particle ~ 10 min. Particle life: 150 - 200 ms. Single particle ~ 10 min. Necessary statistics for TJ-II 10 7 particles. Necessary statistics for TJ-II 10 7 particles.

19. Grid Users Forum, CERN, March 2006 19 Enabling Grids for E-sciencE INFSO-RI-508833 COMPUTING in the GRID: Future applications –EDGE2D Application for tokamaks –Transport Analysis of multiple shots (typically 10 4 shots) or Predictive Transport with multiple models: e. g. ASTRA. CIEMAT(Spa) + IPP(Ger) + Kurchatov(Rus) + EFDA(UE) + … –Neutral Particle Dynamics: EIRENE: CIEMAT(Spa) + IPP(Ger)

20. Grid Users Forum, CERN, March 2006 20 Enabling Grids for E-sciencE INFSO-RI-508833 JET – Flagship of Worldwide Fusion: EDGE2D Equilibrium code.

21. Grid Users Forum, CERN, March 2006 21 Enabling Grids for E-sciencE INFSO-RI-508833 Cross section of present EU D-shaped tokamaks compared to the ITER project EDGE2D: Determine plasma shape from Measurements: Plasma current, Pressure, Magnetic field… -EDGE2D code solves the 2 D fluid equations for the conservation of energy, momentum and particles in the plasma edge region. -Ions, electrons and all ionisation stages of multiple species are considered. -Interaction with the vessel walls is simulated by coupling to monte-carlo codes, to provide the neutral ion and impurity sources.

22. Grid Users Forum, CERN, March 2006 22 Enabling Grids for E-sciencE INFSO-RI-508833 Massive Transport Calculations For Instance: Enhanced heat Confinement in TJ-II. Lower heat diffusivity for low electron density and high absorbed power density. A different case on every PE.

23. Grid Users Forum, CERN, March 2006 23 Enabling Grids for E-sciencE INFSO-RI-508833 EIRENE Code Trayectory of a He atom in TJ-II. Vertical and horizontal proyections. It starts in the green point and is absorbed in the plasma by an ionization process. The real 3D geometry of TJ-II vacuum chamber is considerd.

24. Grid Users Forum, CERN, March 2006 24 Enabling Grids for E-sciencE INFSO-RI-508833 DATA HANDLING Storage: Large data flux: 10 4 sensors x 20-50 kHz sampling= 1-10 GBy per second raw data x 0.5 h= 3 TBy per shot in ITER every 1,5 h Data Access & Sharing in Large Cooperative Experiments: Strategy to be defined: -Database with distributed access or distributed storage? -The evolution of technologies until ITER works. If distributed storage: We need a standard representation for experimental data in LCG-2/gLite CE middleware. Storage should allow to do some basic processing: neural network, clustering…

25. Grid Users Forum, CERN, March 2006 25 Enabling Grids for E-sciencE INFSO-RI-508833 DAS Tools: Visualization, DAQ and processing To add grid-aware protocols for: Data navigation and mining Data exchange Data search Event catch

26. Grid Users Forum, CERN, March 2006 26 Enabling Grids for E-sciencE INFSO-RI-508833 Conclusions VO of Fusion grid is almost ready (the problem of the name). Effort to get more partners in and outside EFDA is being done. Several applications are running in the grid. Future applications for the grid are identified. A deep discussion and investigation on large amount of data handling is needed. Cook book for data handling in the grid is desirable.