targetDP: an Abstraction of Lattice Based Parallelism with Portable Performance

@article{Gray2014targetDPAA,
  title={targetDP: an Abstraction of Lattice Based Parallelism with Portable Performance},
  author={Alan Gray and Kevin Stratford},
  journal={2014 IEEE Intl Conf on High Performance Computing and Communications, 2014 IEEE 6th Intl Symp on Cyberspace Safety and Security, 2014 IEEE 11th Intl Conf on Embedded Software and Syst (HPCC,CSS,ICESS)},
  year={2014},
  pages={312-315}
}
  • A. GrayK. Stratford
  • Published 29 April 2014
  • Computer Science
  • 2014 IEEE Intl Conf on High Performance Computing and Communications, 2014 IEEE 6th Intl Symp on Cyberspace Safety and Security, 2014 IEEE 11th Intl Conf on Embedded Software and Syst (HPCC,CSS,ICESS)
To achieve high performance on modern computers,it is vital to map algorithmic parallelism to that inherent inthe hardware. From an application developer's perspective, itis also important that code can be maintained in a portablemanner across a range of hardware. Here we present targetDP(target Data Parallel), a lightweight programming layer thatallows the abstraction of data parallelism for applications thatemploy structured grids. A single source code may be used totarget both thread level… 

Figures from this paper

A lightweight approach to performance portability with targetDP

Leading high performance computing systems achieve their status through use of highly parallel devices such as NVIDIA graphics processing units or Intel Xeon Phi many-core CPUs. The concept of

A Modular Approach to Performance, Portability and Productivity for 3D Wave Models

A modular approach using existing frameworks to target the productivity problem compiling into an IR language which addresses the performance portability problem is proposed and can be used as an example to extend to other physical simulations using similar algorithms.

Area Exam: General-Purpose Performance Portable Programming Models for Productive Exascale Computing

This area exam will survey many of these proposed general-purpose programming models, including libraries, parallel languages, directive-based language extensions, and source-to-source translators, and compare them in terms of use cases, performance, portability, and developer productivity.

Performance, Portability and Productivity for Room Acoustics Codes

Comparisons of more advanced versions of the benchmarks with the original versions determined that there is a similar performance profile when the room acoustics algorithms become more complex, which has shown that it is possible to develop more productive and performance portable codes for room acoustic simulations which simplify the problems of writing programmable and performant code across different platforms.

References

SHOWING 1-9 OF 9 REFERENCES

On the Performance Portability of Structured Grid Codes on Many-Core Computer Architectures

The results show that, contrary to conventional wisdom, using OpenCL it is possible to achieve a high degree of performance portability, at least for structured grid applications, using a set of straightforward techniques.

OCCA: A unified approach to multi-threading languages

Computational results using finite difference, spectral element and discontinuous Galerkin methods show OCCA delivers portable high performance in different architectures and platforms.

The BAGEL assembler generation library

  • P. Boyle
  • Computer Science
    Comput. Phys. Commun.
  • 2009

The Chroma Software System for Lattice QCD

Multiple GPUs for a fluid lattice Boltzmann application

  • Designing Scientific Applications on GPUs (Ed. R. Couturier), Chapman and Hall/CRC Scientific Computing Series
  • 2013

OpenMP 4.0 Complete Specifications, openmp.org/wp/openmpspecifications

  • OpenMP 4.0 Complete Specifications, openmp.org/wp/openmpspecifications
  • 2013

Better performance at lower occupancy

  • Proceedings of the 2010 GPU Technology Conference
  • 2010

OpenACC 2.0a Spec

  • 2013