© 2006 IBM Corporation Computación de Alto Rendimiento en IBM Research Septiembre 2006 Dr. José G. Castaños IBM T.J. Watson Research.

Presentación del tema: "© 2006 IBM Corporation Computación de Alto Rendimiento en IBM Research Septiembre 2006 Dr. José G. Castaños IBM T.J. Watson Research."— Transcripción de la presentación:

1 © 2006 IBM Corporation Computación de Alto Rendimiento en IBM Research Septiembre 2006 Dr. José G. Castaños castanos@us.ibm.com IBM T.J. Watson Research Center

2 IBM Research © 2006 IBM Corporation El Projecto Blue Gene En Diciembre 1999, IBM Research anuncia el Blue Gene Producir nuevos adelantos en simulaciones biomoleculares Investigar nueva tecnologias en hardware y software para producir computadoras de alto rendimiento Blue Gene sigue un enfoque modular, donde el bloque basico (o célula) puede copiarse ad infinitum Procesadores de bajo consumo – permite rendimientos combinados mas altos –PowerPC 440 System-on-a-chip ofrece ventajas en costo/rendimiento –Menor complejidad –Alta densidad (2048 procesadores por rack, enfriado por aire) –Redes integradas para gran escala Ambiente de software familiar, simplificado para HPC Mucha atención a RAS (reliability, availability, and serviceability) en todo el sistema

3 IBM Research © 2006 IBM Corporation El Chip Blue Gene/L (ASIC) IBM CU-11, 0.13 µm 11 x 11 mm die size 25 x 32 mm CBGA 474 pins, 328 signal 1.5/2.5 Volt

4 IBM Research © 2006 IBM Corporation Arquitectura de Blue Gene/L

5 IBM Research © 2006 IBM Corporation Blue Gene/L en Lawrence Livermore National Laboratory BG/L Number of Racks64 Number of Nodes65536 Processor Frequency700 Mhz Peak Performance Rack5.7 TF Machine360 TF Linpack280 TF App101 TF MemoryRack256 GB Machine16 TB Power~2 MW Size250 sq.m. Bisection Bandwidth700 GB/s CablesNumber5,000 Length25 km StorageDisk200 TB

6 IBM Research © 2006 IBM Corporation Source: www.top500.org # Ven- dor Rmax TFlops Installation 1IBM280.6 DOE/NSSA/LLNL (64 racks BlueGene/L) 2IBM91.2 BlueGene at Watson (20 racks BlueGene/L) 3IBM75.8 ASC Purple LLNL (1280 nodes p5 575) 4SGI51.9NASA/Columbia (Itanium2) 5Bull42.90 CEA/DAM Tera10 (Itanium2) 6Dell38.27 Sandia -Thunderbird (EM64T/Infiniband) 7Sun38.18 Tsubame Galaxy TiTech (Opteron/Infiniband) 8IBM37.33 FRZ – Juelich (8 racks BlueGene/L) 9Cray36.19 Sandia – Red Storm (XT3 Opteron) 10NEC35.86 Japan Earth Simulator (NEC) # Ven- dor Rmax TFlops Instalation 11IBM27.91 MareNostrum Barcelona Supercomputer (JS20) 12IBM27.45 ASTRON Netherlands (6 racks BlueGene/L) 13Cray20.52 ORNL – Jaguar (XT3 Opteron) 14Calif Dig19.94 LLNL (Itanium2) 15IBM18.20 AIST - Japan (4 rack BlueGene/L) 16IBM18.20 EPFL - Switzerland (4 rack BlueGene/L) 17IBM18.20 KEK – Japan (4 rack BlueGene/L) 18IBM18.20 KEK – Japan (4 rack BlueGene/L) 19IBM18.20 IBM – On Demand Ctr (4 rack BlueGene/L) 20Cray16.97 ERDC MSRC (Cray XT3 Opteron) Blue Gene en los Top500

7 IBM Research © 2006 IBM Corporation Motivacion del Software de Sistema Nodos de Computación dedicados a ejecutar una sola aplicacion, and casi nada más Compute node kernel (CNK) Simplicidad! Nodos de I/O corren Linux and proveen servicios de OS – files, sockets, comenzar programas, señales, debugging, and fianalización de tareas Solution estandar: Linux Nodos de Servicio ejecutan todos los servicios de administración (e.g., latidos, checkean errores) transparente para el programa de los usuarios

8 IBM Research © 2006 IBM Corporation Blue Gene/L: Architectura del Software de Sistema Functiona l Gigabit Ethernet I/O Node 0 Linux ciod C-Node 0 CNK I/O Node 1023 Linux ciod C-Node 0 CNK C-Node 63 CNK C-Node 63 CNK Control Gigabit Ethernet IDo chip LoadLeveler System Console CMCS JTAG torus tree DB2 Front-end Nodes Pset 1023 Pset 0 I2CI2C File Servers fs client Service Node app

9 IBM Research © 2006 IBM Corporation Classical MD – ddcMD 2005 Gordon Bell Prize Winner!! Scalable, general purpose code for performing classical molecular dynamics (MD) simulations using highly accurate MGPT potentials MGPT semi-empirical potentials, based on a rigorous expansion of many body terms in the total energy, are needed in to quantitatively investigate dynamic behavior of d-shell and f- shell metals. 524 million atom simulations on 64K nodes achieved 101.5 TF/s sustained. Superb strong and weak scaling for full machine - (very impressive machine says PI) Visualization of important scientific findings already achieved on BG/L: Molten Ta at 5000K demonstrates solidification during isothermal compression to 250 GPa 2,048,000 Tantalum atoms

10 IBM Research © 2006 IBM Corporation Resolidificación Rapida del Tantalum (ddcMD) Nucleation of solid is initiated at multiple independent sites throughout each sample cell Growth of solid grains initiates independently, but soon leads to grain boundaries which span the simulation cell: size of cell is now influencing continued growth 2,048,000 simulation recently performed indicates formation of many more grains