Zenodo (CERN European Organization for Nuclear Research), Aug 30, 2008
This document contains an analysis of a major survey of HPC systems and applications across the P... more This document contains an analysis of a major survey of HPC systems and applications across the PRACE 1 partners (see [1]). The survey data was used to produce an overall utilisation matrix characterised by scientific area and algorithm. This matrix represents one of the best snapshots of HPC utilisation produced, and could be an invaluable basis to predict the likely utilisation of future European Petaflop/s systems. We discuss a methodology for weighting subset lists of applications to maximise their match with the utilisation matrix. This process is used to guide the selection of applications for inclusion in the representative benchmark suite. By selecting highly used applications from a range of scientific areas and algorithms, we were able to produce a list of no more than 16 applications that are generally representative of European HPC usage and fitted the utilisation matrix well. Drawing on the myriad PRACE expertise in existing and emerging applications areas, we produced a recommended list of nine representative applications, based on the best-fitting list, for use in benchmarking future Petaflop/s systems. The future of HPC in Europe is also considered. Programming multi-core architectures is seen as a major challenge alongside parallel I/O and techniques for scaling applications to thousands of cores. By comparing large and small systems in the survey it would appear that Particle Physics may well be the main scientific area in a future Petaflop/s system, alongside Materials and Computational Chemistry. The study detailed in this report is a detailed snapshot of current European HPC, enabling both the choice of a representative benchmark suite and providing insight into what might run on future systems.
We have studied the light hadron spectrum and decay constants for quenched QCD at β = 6.2 on a 24... more We have studied the light hadron spectrum and decay constants for quenched QCD at β = 6.2 on a 24 3 × 48 lattice. We compare the results obtained using a nearest-neighbour O(a)-improved ("clover") fermion action with those obtained using the standard Wilson fermion action on the same gauge configurations. For pseudoscalar meson masses in the range 330-800 MeV, we find no significant difference between the results for the two actions. The scales obtained from the string tension and mesonic sector are consistent, but higher than those derived from baryon masses. The ratio of the pseudoscalar decay constant to the vector meson mass is roughly independent of quark mass as observed experimentally.
Second NASA/ESA Conference on Adaptive Hardware and Systems (AHS 2007), 2007
We describe the FPGA-based supercomputer Maxwell built by the FPGA High-Performance Computing All... more We describe the FPGA-based supercomputer Maxwell built by the FPGA High-Performance Computing Alliance at the University of Edinburgh. Winner of the silver medal in the BT Flagship Award for Innovation at the 2007 British Computer Society Awards, Maxwell is a general-purpose 64 FPGA computer designed for high-performance parallel computing. This paper describes the machine itself, its hardware and software environment and presents benchmark results from runs of three commercial demonstration applications from the oil, medical and finance sectors.
First-principles molecular dynamics simulations having a duration of 8 ps have been used to study... more First-principles molecular dynamics simulations having a duration of 8 ps have been used to study the static, dynamic and electronic properties of ℓ-Ga at the temperatures 702 K and 982 K. The simulations use the densityfunctional pseudopotential method and the system is maintained on the Born-Oppenheimer surface by conjugate gradients relaxation. The static structure factor and radial distribution function of the simulated system agree very closely with experimental data, but the diffusion coefficient is noticeably lower than measured values. The long simulations allow us to calculate the dynamical structure factor S(q, ω). A sound-wave peak is clearly visible in S(q, ω) at small wavevectors, and we present results for the dispersion curve and hence
EPCC was one of three science support centres set up under HPCI, along with Daresbury Laboratory ... more EPCC was one of three science support centres set up under HPCI, along with Daresbury Laboratory and the University of Southampton. The major role for the centres was to provide HPC and computational science expertise for the various national consortia.
Second NASA/ESA Conference on Adaptive Hardware and Systems (AHS 2007), 2007
This paper reviews the current state of the art in highperformance reconfigurable computing (HPRC... more This paper reviews the current state of the art in highperformance reconfigurable computing (HPRC) from the perspective of EPCC, the high-performance computing centre at the University of Edinburgh. We look at architectural and programming trends and assess some of the challenges that HPRC needs to address in order to drive itself across the chasm from the optimistic early adopters to the pragmatic early majority.
Recent Advances in Parallel Virtual Machine and Message Passing Interface, 2001
hydrampi is a portable parallel N-body solver, based on the adaptive P3M algorithm. This Fortran9... more hydrampi is a portable parallel N-body solver, based on the adaptive P3M algorithm. This Fortran90 code is parallelised using a non-trivial task-farm and two domain decompositions: a 2D cycle of blocks and a slab distribution, using both MPI-1.1 and MPI-2 communications routines. Specifically, MPI_Put and MPI_Get are employed extensively in association with the communication epochs MPI_Fence and MPILock/MPIUnlock. The 1D
Abstract We present the first physics results for quenched QCD and for pure SU (2) gauge theory f... more Abstract We present the first physics results for quenched QCD and for pure SU (2) gauge theory from the UKQCD collaboration. We report preliminary results for static quark potentials, the string tension and the quenched hadron spectrum on a 24 3× 48 lattice at β ...
We study the approach to the continuum limit of SU(2) lattice gauge theory by accurately calculat... more We study the approach to the continuum limit of SU(2) lattice gauge theory by accurately calculating se~ eral physical quanntles on latnces up to 2& and for couphngs up to fl=4/gZ=2 6 We find that the dependence of mass ranos on the lattice spacing weakens as the spacing decreases, although the two-loop relation between the spacing and g~ ~s not ;yet attained for~<2 6 We findglueballmasses, andthestrlngtensaonloberoughl~lntheranom(2-) m(0) m(2 +) m(O +) \ K~7 65 55 38 1 for the smallest latnce spacings
We employ a nonrelativistic version of QCD (NRQCD) to study heavy qq-bound states in the lowest a... more We employ a nonrelativistic version of QCD (NRQCD) to study heavy qq-bound states in the lowest approximation without fine structure. We use gluon configurations on a 16 3 ×48-lattice at β = 6.2 from the UKQCD collaboration. For a bare quark mass near that of the b-quark (M a = 1.6) we obtain the bound state masses for the S, P and both types of D-waves. We also detect signals for two types of hybrids (qqg-states). The results are sufficiently accurate to confirm that the values of the D-wave mass from both D-waves coincide thus indicating that the cubical invariance of the lattice is restored to full rotational invariance to a good approximation. We also study S and P-wave masses for values of the range of bare quark mass M a = 1.0, 1.3, 1.6 & 1.9. The results confirm the idea that the S/P-splitting is relatively insensitive to the value of the bare quark mass.
We present the first study of the light hadron spectrum and decay constants for quenched QCD usin... more We present the first study of the light hadron spectrum and decay constants for quenched QCD using an O(a)-improved nearest-neighbour Wilson fermion action at β = 6.2. We compare the results with those obtained using the standard Wilson fermion action, on the same set of 18 gauge field configurations of a 24 3 ×48 lattice. For pseudoscalar meson masses in the range 330-800 MeV, we find no significant difference between the results for the two actions. The scales obtained from the string tension and mesonic sector are consistent, but differ from that derived from baryon masses. The ratio of the pseudoscalar decay constant to the vector meson mass is roughly independent of quark mass as observed experimentally, and in approximate agreement with the measured value.
... PHYSICAL REVIEW LETTERS First Principles Simulations of Silicon Nanoindentation RubenPerez an... more ... PHYSICAL REVIEW LETTERS First Principles Simulations of Silicon Nanoindentation RubenPerez and Michael C. Payne TCM, Cavendish Laboratory, University of Cambridge, Madingley Road, Cambridge CB3 OHE, United Kingdom Alan D. Simpson Edinburgh Parallel ...
We present results for heavy-light decay constants, using both propagating quarks and the static ... more We present results for heavy-light decay constants, using both propagating quarks and the static approximation, in O(a)-improved, quenched lattice QCD. At β = 6.2 on a 24 3 ×48 lattice we find f D = 185 + 4 − 3 (stat) + 42 − 7 (syst) MeV, f B = 160
Zenodo (CERN European Organization for Nuclear Research), Aug 30, 2008
This document contains an analysis of a major survey of HPC systems and applications across the P... more This document contains an analysis of a major survey of HPC systems and applications across the PRACE 1 partners (see [1]). The survey data was used to produce an overall utilisation matrix characterised by scientific area and algorithm. This matrix represents one of the best snapshots of HPC utilisation produced, and could be an invaluable basis to predict the likely utilisation of future European Petaflop/s systems. We discuss a methodology for weighting subset lists of applications to maximise their match with the utilisation matrix. This process is used to guide the selection of applications for inclusion in the representative benchmark suite. By selecting highly used applications from a range of scientific areas and algorithms, we were able to produce a list of no more than 16 applications that are generally representative of European HPC usage and fitted the utilisation matrix well. Drawing on the myriad PRACE expertise in existing and emerging applications areas, we produced a recommended list of nine representative applications, based on the best-fitting list, for use in benchmarking future Petaflop/s systems. The future of HPC in Europe is also considered. Programming multi-core architectures is seen as a major challenge alongside parallel I/O and techniques for scaling applications to thousands of cores. By comparing large and small systems in the survey it would appear that Particle Physics may well be the main scientific area in a future Petaflop/s system, alongside Materials and Computational Chemistry. The study detailed in this report is a detailed snapshot of current European HPC, enabling both the choice of a representative benchmark suite and providing insight into what might run on future systems.
We have studied the light hadron spectrum and decay constants for quenched QCD at β = 6.2 on a 24... more We have studied the light hadron spectrum and decay constants for quenched QCD at β = 6.2 on a 24 3 × 48 lattice. We compare the results obtained using a nearest-neighbour O(a)-improved ("clover") fermion action with those obtained using the standard Wilson fermion action on the same gauge configurations. For pseudoscalar meson masses in the range 330-800 MeV, we find no significant difference between the results for the two actions. The scales obtained from the string tension and mesonic sector are consistent, but higher than those derived from baryon masses. The ratio of the pseudoscalar decay constant to the vector meson mass is roughly independent of quark mass as observed experimentally.
Second NASA/ESA Conference on Adaptive Hardware and Systems (AHS 2007), 2007
We describe the FPGA-based supercomputer Maxwell built by the FPGA High-Performance Computing All... more We describe the FPGA-based supercomputer Maxwell built by the FPGA High-Performance Computing Alliance at the University of Edinburgh. Winner of the silver medal in the BT Flagship Award for Innovation at the 2007 British Computer Society Awards, Maxwell is a general-purpose 64 FPGA computer designed for high-performance parallel computing. This paper describes the machine itself, its hardware and software environment and presents benchmark results from runs of three commercial demonstration applications from the oil, medical and finance sectors.
First-principles molecular dynamics simulations having a duration of 8 ps have been used to study... more First-principles molecular dynamics simulations having a duration of 8 ps have been used to study the static, dynamic and electronic properties of ℓ-Ga at the temperatures 702 K and 982 K. The simulations use the densityfunctional pseudopotential method and the system is maintained on the Born-Oppenheimer surface by conjugate gradients relaxation. The static structure factor and radial distribution function of the simulated system agree very closely with experimental data, but the diffusion coefficient is noticeably lower than measured values. The long simulations allow us to calculate the dynamical structure factor S(q, ω). A sound-wave peak is clearly visible in S(q, ω) at small wavevectors, and we present results for the dispersion curve and hence
EPCC was one of three science support centres set up under HPCI, along with Daresbury Laboratory ... more EPCC was one of three science support centres set up under HPCI, along with Daresbury Laboratory and the University of Southampton. The major role for the centres was to provide HPC and computational science expertise for the various national consortia.
Second NASA/ESA Conference on Adaptive Hardware and Systems (AHS 2007), 2007
This paper reviews the current state of the art in highperformance reconfigurable computing (HPRC... more This paper reviews the current state of the art in highperformance reconfigurable computing (HPRC) from the perspective of EPCC, the high-performance computing centre at the University of Edinburgh. We look at architectural and programming trends and assess some of the challenges that HPRC needs to address in order to drive itself across the chasm from the optimistic early adopters to the pragmatic early majority.
Recent Advances in Parallel Virtual Machine and Message Passing Interface, 2001
hydrampi is a portable parallel N-body solver, based on the adaptive P3M algorithm. This Fortran9... more hydrampi is a portable parallel N-body solver, based on the adaptive P3M algorithm. This Fortran90 code is parallelised using a non-trivial task-farm and two domain decompositions: a 2D cycle of blocks and a slab distribution, using both MPI-1.1 and MPI-2 communications routines. Specifically, MPI_Put and MPI_Get are employed extensively in association with the communication epochs MPI_Fence and MPILock/MPIUnlock. The 1D
Abstract We present the first physics results for quenched QCD and for pure SU (2) gauge theory f... more Abstract We present the first physics results for quenched QCD and for pure SU (2) gauge theory from the UKQCD collaboration. We report preliminary results for static quark potentials, the string tension and the quenched hadron spectrum on a 24 3× 48 lattice at β ...
We study the approach to the continuum limit of SU(2) lattice gauge theory by accurately calculat... more We study the approach to the continuum limit of SU(2) lattice gauge theory by accurately calculating se~ eral physical quanntles on latnces up to 2& and for couphngs up to fl=4/gZ=2 6 We find that the dependence of mass ranos on the lattice spacing weakens as the spacing decreases, although the two-loop relation between the spacing and g~ ~s not ;yet attained for~<2 6 We findglueballmasses, andthestrlngtensaonloberoughl~lntheranom(2-) m(0) m(2 +) m(O +) \ K~7 65 55 38 1 for the smallest latnce spacings
We employ a nonrelativistic version of QCD (NRQCD) to study heavy qq-bound states in the lowest a... more We employ a nonrelativistic version of QCD (NRQCD) to study heavy qq-bound states in the lowest approximation without fine structure. We use gluon configurations on a 16 3 ×48-lattice at β = 6.2 from the UKQCD collaboration. For a bare quark mass near that of the b-quark (M a = 1.6) we obtain the bound state masses for the S, P and both types of D-waves. We also detect signals for two types of hybrids (qqg-states). The results are sufficiently accurate to confirm that the values of the D-wave mass from both D-waves coincide thus indicating that the cubical invariance of the lattice is restored to full rotational invariance to a good approximation. We also study S and P-wave masses for values of the range of bare quark mass M a = 1.0, 1.3, 1.6 & 1.9. The results confirm the idea that the S/P-splitting is relatively insensitive to the value of the bare quark mass.
We present the first study of the light hadron spectrum and decay constants for quenched QCD usin... more We present the first study of the light hadron spectrum and decay constants for quenched QCD using an O(a)-improved nearest-neighbour Wilson fermion action at β = 6.2. We compare the results with those obtained using the standard Wilson fermion action, on the same set of 18 gauge field configurations of a 24 3 ×48 lattice. For pseudoscalar meson masses in the range 330-800 MeV, we find no significant difference between the results for the two actions. The scales obtained from the string tension and mesonic sector are consistent, but differ from that derived from baryon masses. The ratio of the pseudoscalar decay constant to the vector meson mass is roughly independent of quark mass as observed experimentally, and in approximate agreement with the measured value.
... PHYSICAL REVIEW LETTERS First Principles Simulations of Silicon Nanoindentation RubenPerez an... more ... PHYSICAL REVIEW LETTERS First Principles Simulations of Silicon Nanoindentation RubenPerez and Michael C. Payne TCM, Cavendish Laboratory, University of Cambridge, Madingley Road, Cambridge CB3 OHE, United Kingdom Alan D. Simpson Edinburgh Parallel ...
We present results for heavy-light decay constants, using both propagating quarks and the static ... more We present results for heavy-light decay constants, using both propagating quarks and the static approximation, in O(a)-improved, quenched lattice QCD. At β = 6.2 on a 24 3 ×48 lattice we find f D = 185 + 4 − 3 (stat) + 42 − 7 (syst) MeV, f B = 160
Uploads
Papers by Alan Simpson