Explorations of the viability of ARM and Xeon Phi for physics processing

  title={Explorations of the viability of ARM and Xeon Phi for physics processing},
  author={David Abdurachmanov and Kapil Arya and Joshua Lorne Bendavid and Tommaso Boccali and Gene Cooperman and Andrea Dotti and Peter Elmer and Giulio Eulisse and Francesco Giacomini and Christopher D. Jones and Matteo Manzali and Shahzad Muzaffar},
  journal={Journal of Physics: Conference Series},
We report on our investigations into the viability of the ARM processor and the Intel Xeon Phi co-processor for scientific computing. We describe our experience porting software to these processors and running benchmarks using real physics applications to explore the potential of these processors for production physics processing. 

Heterogeneous High Throughput Scientific Computing with APM X-Gene and Intel Xeon Phi

This paper examines the Intel Xeon Phi Many Integrated Cores (MIC) co-processor and Applied Micro X-Gene ARMv8 64-bit low-power server system-on-a-chip (SoC) solutions for scientific computing applications and evaluates the potential for use of such technologies in the context of distributed computing systems such as the Worldwide LHC Computing Grid (WLCG).

Floating-point performance of ARM cores and their efficiency in classical molecular dynamics

This work presents the analysis of the floating point performance of the latest ARM cores and their efficiency for the algorithms of classical molecular dynamics.

Towards an algorithmic skeleton framework for programming the Intel R Xeon PhiTM processor

projects PTDC/EIA- EIA/113613/2009 (Synergy-VM) and PTDC/EEI-CTP/1837/2012 (SwiftComp) for financing the purchase of the Intel R Xeon PhiTM

Techniques and tools for measuring energy efficiency of scientific software applications

This work performs several physical and software-based measurements of workloads from HEP applications running on ARM and Intel architectures, and compares their power consumption and performance, and leverage several profiling tools to extract different characteristics of the power use.

Virtualizing high-end GPGPUs on ARM clusters for the next generation of high performance cloud computing

This work describes here how to accelerate inexpensive ARM-based computing nodes with high-end GPGPUs hosted on x86_64 machines using the GVirtuS general-purpose virtualization service.

Power-aware applications for scientific cluster and distributed computing

How power-aware software applications and scheduling might be used to reduce power consumption, both as autonomous entities and as part of a (globally) distributed system are discussed.

Exploiting multicore processors in PLCs using libraries for IEC 61131-3

The case study results show an explicit benefit of the multicore exploiting software in comparison to its singlecore counterpart, which is reflected with a faster processing of up to a factor of 3.5.

User-Space Process Virtualization in the Context of Checkpoint-Restart and Virtual Machines

This dissertation presents user-space process virtualization to decouple application processes from the external subsystems and an adaptive plugin based approach is used to implement the virtualization layers that allow the checkpoint-restart system to grow organically.

Fiducial cross-section measurements of the production of a prompt photon in association with a top-quark pair at $\sqrt{s}=13$ TeV with the ATLAS detector at the LHC

The cross sections for top-quark pair production in association with a photon are measured in a fiducial volume with the ATLAS detector at a centre-of-mass energy of 13 TeV. Results are presented

Эффективность процессоров ARM для расчетов классической молекулярной динамики \ast

Суперкомпьютерные вычисления экзафлопсной эры будут неизбежно ограничены энергоэффективностью. Сегодня в качестве возможных кандидатов для этих целей рассматриваются различные микропроцессорные



Initial explorations of ARM processors for scientific computing

The results of the initial investigations into the use of ARM processors for scientific computing applications are presented and ARM-specific issues regarding the software development environment, operating system, performance benchmarks and issues for porting High Energy Physics software are explored.

Use of checkpoint-restart for complex HEP software on traditional architectures and Intel MIC

This work analyzes both single- and multi-threaded applications and test on both standard Intel x86 architectures and on Intel MIC, considered an indicator of what the future may hold for many-core computing.

Computing for the Large Hadron Collider

  • I. Bird
  • Computer Science, Physics
  • 2011
The rationale for the design of a distributed system and how this environment was constructed and deployed through the use of grid computing technologies are discussed and the experience with large-scale testing and operation with real accelerator data shows that expectations have been met and sometimes exceeded.

The ATLAS Experiment at the CERN Large Hadron Collider

This paper describes the ATLAS experiment as installed in i ts experimental cavern at point 1 at CERN. It also presents a brief overview of the expec ted performance of the detector.

The CMS experiment at CERN

  • C. Wulz
  • Physics
    SPIE Optics + Optoelectronics
  • 2005
The search for new physics at high energies is the motivation for the construction of the CMS (Compact Muon Solenoid) experiment at CERN, the European Organization for Nuclear Research in Geneva. It

Stitched Together: Transitioning CMS to a Hierarchical Threaded Framework

This work will present CMS' effort to evolve the authors' single threaded framework into a highly concurrent framework, and outline the design of the new framework and how the design was constrained by the initial single threaded design.

Annual Review Of Nuclear And Particle Science

The contents of this review reflect some of the shifts of emphasis that are occurring among the fields of astrophysics, nuclear physics, and elementary particle physics. Particle physics has made

DMTCP: Transparent checkpointing for cluster computations and the desktop

Experimental results show that checkpoint time remains nearly constant as the number of nodes increases on a medium-size cluster, and DMTCP can be incorporated and distributed as a checkpoint-restart module within some larger package.

Optimization of the CMS software build and distribution system

This work describes how parallel build of software and minimal distribution size dramatically reduced the time gap between software build and installation on remote sites, and how producing few big binary products, instead of thousands of small ones, helped finding out the integration and runtime issues.