J
J. A. Herdman
Researcher at Atomic Weapons Establishment
Publications - 21
Citations - 431
J. A. Herdman is an academic researcher from Atomic Weapons Establishment. The author has contributed to research in topics: CUDA & Software portability. The author has an hindex of 10, co-authored 21 publications receiving 399 citations.
Papers
More filters
Proceedings ArticleDOI
Accelerating Hydrocodes with OpenACC, OpeCL and CUDA
J. A. Herdman,Wayne Gaudin,Simon McIntosh-Smith,Michael Boulton,David Beckingsale,A. C. Mallinson,Stephen A. Jarvis +6 more
TL;DR: It is found that OpenACC is an extremely viable programming model for accelerator devices, improving programmer productivity and achieving better performance than OpenCL and CUDA.
Journal ArticleDOI
An investigation of the performance portability of OpenCL
Simon J. Pennycook,Simon D. Hammond,Steven A. Wright,J. A. Herdman,I. Miller,Stephen A. Jarvis +5 more
TL;DR: The development of an MPI/OpenCL implementation of LU, an application-level benchmark from the NAS Parallel Benchmark Suite, is reported, demonstrating the importance of memory arrangement and work-item/work-group distribution strategies when applications are deployed on different device types.
Proceedings ArticleDOI
WARPP: a toolkit for simulating high-performance parallel scientific codes
TL;DR: The Warwick Performance Prediction (WARPP) simulator is described, which is used to construct application performance models for complex industry-strength parallel scientific codes executing on thousands of processing cores.
CloverLeaf: Preparing Hydrodynamics Codes for Exascale
A. C. Mallinson,David Beckingsale,Wayne Gaudin,J. A. Herdman,J. M. Levesque,Stephen A. Jarvis +5 more
TL;DR: This work directly evaluates five candidate programming models for future exascale applications using a recently developed Lagrangian-Eulerian explicit hydrodynamics mini-application to better inform the exacsale planning at large HPC centres such as AWE.
Journal ArticleDOI
Parallel File System Analysis Through Application I/O Tracing
Steven A. Wright,Simon D. Hammond,Simon J. Pennycook,Robert Bird,J. A. Herdman,I. Miller,A. Vadgama,Abhir Bhalerao,Stephen A. Jarvis +8 more
TL;DR: The RIOT toolkit provides a powerful framework in which to analyse I/O and parallel file system behaviour and demonstrates the large file locking overhead of IBM's General Parallel File System, which can consume nearly 30% of the total write time in the FLASH-IO benchmark.