HPCx: Multi-Teraflops in the UK A World-Class Service for World-Class Research Dr Arthur Trew...
Transcript of HPCx: Multi-Teraflops in the UK A World-Class Service for World-Class Research Dr Arthur Trew...
HPCx: Multi-Teraflops in the UK
A World-Class Service for World-Class Research
Dr Arthur Trew
Director
HPCx rationale
UK academic research is increasingly dependent upon high-end compute facilities
recent technological and Grid advances highlighted the need to upgrade UK resources
HPCx objectives are thus:– aim “to deliver optimum service resulting in world-
leading science”– address “the problems involved in scaling to the
capability levels required”
What is HPCx?
Consortium of leading UK organisations committed to creating and managing the new resource for the next 6 years– led by University of Edinburgh
multi-stage project to deliver a world-class academic computing resource, the largest in Europe
£54M/$100M budget Grid-enabled, a key component in the UK e-Science
programme
HPCx Consortium Members
Daresbury Laboratory, CCLRC
IBM
EPCC, University of Edinburgh
UoE HPCx Ltd – wholly-owned subsidiary of the University of Edinburgh and lead contractor
the University of Edinburgh is one of the top 5 research universities in the UK
EPCC is the leading computer centre in Europe, bridging the gap between academia and industry
… and provides both HPC and novel computing solutions to a wide range of problems and users
long experience of providing national HPC services including:– Meiko Computing Surfaces– Thinking Machines CM200– Cray T3D/T3E
EPCC overview
Academic:o National HPC facilities o Researcho Support
Training:o Academiao Industry
European leadership:o Visitor programmeso Technology Transfero Strategic Planning
Industry:o Projectso Consultancyo Standards
Technology Transfer
Daresbury Laboratory
A multi disciplinary research lab with over 500 people
Provides large-scale research facilities both for UK academic and industrial research communities
Runs the UK’s Collaborative Computational Projects
Daresbury hosts the HPCx hardware
IBM provides the technology for HPCx Long standing involvement in HPC including the
development of a number of ASCI machines and 4 of the top dozen machines in the 21st TOP500 list:– ASCI White: Rmax = 7.3 TFlop/s– SP Power3 (6656 Processors): Rmax = 7.3 TFlop/s– xSeries (1920 Processors): Rmax = 6.6 TFlop/s– HPCx (1280 processors): Rmax = 3.2 TFlop/s
IBM has the long term technology road map essential to a 6 year project such as HPCx
HPCx: Phase 1
System will be commissioned in three main stages Phase 1 covering 2002-2004 consists of:– 40 Regatta-H SMP nodes, 1280 processors– Peak performance 6.6 TFlop/s, with 3.4 Tflop/s sustained
Linpack• currently 16th in the Top500
– 1.28 TB total memory capacity– Over 50 TB of storage capacity– Double plane Colony switch with total
peak bandwidth of 250 MB/s per processor
HPCx Phases 2 & 3
Phase 2 (2004-2006)– aiming for 6 TFlop/s sustained on Linpack and 2.5 TFlop/s on
sPPM– O(48) Regatta-H+ SMP nodes– interconnect upgraded to Federation switch– doubling of I/O and storage
• already built a cluster with 8 Regatta-H+ frames and a pre-release Federation switch
• undertaking a phased upgrade during 1H2004 Phase 3 (2006-2008)
– target of 12 TFlop/s sustained on Linpack– may be additional nodes or alternative technology
HPCx Science Support
Outreach
Life sciences
New applications
Applications Support
Helpdesk Training
Liaising with users
Users
Technology
Software EngineeringUnderpinning technology Grid/e-Science
Systems & NetworkingFlexible and responsive capability computing service
Smooth transitions between phases
Terascaling Capability applications
Scalable algorithms Performance optimisation
18 staff in 5 dual-centre functional support teams
HPCx Status: Usage
Utilisation of Capability Region
0%
20%
40%
60%
80%
100%
Jan
-03
Fe
b-0
3
Ma
r-03
Ap
r-03
Ma
y-03
Jun
-03
Jul-0
3
Au
g-0
3
Se
p-0
3
Oct-0
3
No
v-03
De
c-03
CPU Usage by Job Size
0
500000
1000000
1500000
2000000
Jan
-03
Fe
b-0
3
Ma
r-03
Ap
r-03
Ma
y-03
Jun
-03
Jul-0
3
Au
g-0
3
Se
p-0
3
Oct-0
3
No
v-03
De
c-03
AU
s
>1024
1024
512
256
128
64
32
16
8
currently 28 active consortia, and over 450 users
Life Sciences outreach activity supported by IBM
HPCx Status: Application Areas
Usage by Application Area
Chemistry
Engineering
Environmental Science
Materials
Science Support
Atomic & Molecular Physics
Atomic and Molecular Physics
The UK Multiphoton, Electron Collision and Bose Einstein Condensates (MECBEC) HPC Consortium
Two flagship projects model two-electron atoms (helium) and molecules (hydrogen) exposed to intense, ultra-short laser pulses
Modelling involves the grid solution of multi-dimensional, time-dependent partial differential equations)
Visualisation techniques crucial in extracting information
Simulation of Double Ionization of laser-driven helium performed at Queen’s University Belfast
Requires large amounts of computing power
Environmental Science: POLCOMS
POLCOMS is a multi-disciplinary model developed at the Proudman Oceanographic Laboratory
… a 3-D hydrodynamic model integrating coasts and oceans using a wide range of associated models
A simulation of Chlorophyll density in UK
waters using the POLCOMS model
POLCOMS is a step towards real time modelling of coastal zones, enabling better analysis of impacts to, and sustainability of, the marine environment
Material Science: Crystal
CRYSTAL computes electronic structure and related properties of periodic systems
Developed jointly by Daresbury and the University of Turin A Fortran90 and MPI program that performs Hartree-Fock, density
functional and other approximation calculations On HPCx, CRYSTAL used to calculate the structure of the Crambin
molecule, the largest Hartree-Fock calculation ever converged (1284 atoms)– next, the Rusticyanin molecule (6284 atoms)
Engineering: UKTC UK Turbulence Consortium is developing world
leading turbulence simulation codes using HPCx essential that they can easily facilitate the
scientific usage of the resulting data the calculated data are transferred from HPCx to a
remote site for analysis so, the Grid is becoming increasingly important
Chemistry: RealityGrid RealityGrid is a UK collaboration
– which aims to grid-enable the realistic modelling and simulation of complex condensed matter structures at the meso and nanoscale levels
HPCx terascaling team has worked on parallelisation & optmisation of RealityGrid codes, such as LB3D
RealityGrid also uses the Grid for data transfer, computational steering, and remote visualisation– using Globus on HPCx– TeraGyroid won award at SC2003
Summary
HPCx is the new UK HPC resource– capability computing for world-leading science– largest academic computer in Europe
Series of IBM pSeries clusters– 6.6TF 12TF 22TF– remaining very competitive until 2008+
Very successful first year
HPCx: Supporting theFuture of UK Research