A Performance Analysis of Modern Parallel Programming Models Using a Compute-Bound Application

Poenaru, Andrei; Lin, Wenzhi; McIntosh–Smith, Simon

doi:10.1007/978-3-030-78713-4_18

Cited by 16 publications

(12 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The SYCL results again confirm that its results on the CPU tend to lag behind the other models. For MiniBUDE, this was observed previously [21], although the performance of DPC++ for this benchmark on CPUs seems to have regressed.…”

Section: B Minibudesupporting

confidence: 74%

See 1 more Smart Citation

Heterogeneous Programming for the Homogeneous Majority

Deakin

Cownie

Lin

et al. 2022

2022 IEEE/ACM International Workshop on Performance, Portability and Productivity in HPC (P3HPC)

View full text Add to dashboard Cite

In order to take advantage of the burgeoning diversity in processors at the frontier of supercomputing, the HPC community is migrating and improving codes to utilise heterogeneous nodes, where accelerators, principally GPUs, are highly prevalent in top-tier supercomputer designs. Programs therefore need to embrace at least some of the complexities of heterogeneous architectures. Parallel programming models have evolved to express heterogeneous paradigms whilst providing mechanisms for writing portable, performant programs. History shows that technologies first introduced at the frontier percolate down to local workhorse systems. However, we expect there will always be a mix of systems, some heterogeneous, but some remaining as homogeneous CPU systems. Thus it is important to ensure codes adapted for heterogeneous systems continue to run efficiently on CPUs. In this study, we explore how well widely used heterogeneous programming models perform on CPU-only platforms, and survey the performance portability they offer on the latest CPU architectures.

show abstract

Section: B Minibudesupporting

confidence: 74%

“…MiniBUDE is a proxy for the Bristol University Docking Engine (BUDE), a protein docking molecular dynamics code [21]. It is compute bound, typically achieving around 50% of the peak floating point performance of any given platform; as such it is measurably performance portable.…”

Section: Benchmarksmentioning

confidence: 99%

Heterogeneous Programming for the Homogeneous Majority

Deakin

Cownie

Lin

et al. 2022

2022 IEEE/ACM International Workshop on Performance, Portability and Productivity in HPC (P3HPC)

View full text Add to dashboard Cite

show abstract

“…MiniBUDE accepts an input deck, number of poses, and the iteration as parameters. We left everything as default, using the bm1 deck at 65536 poses for 8 iterations for the results to be consistent with our previous work in [15]. The mini-app contains built-in validation procedures where the computed values are checked against the known correct values of the input deck.…”

Section: B Minibudementioning

confidence: 88%

“…MiniBUDE is a compute bound HPC mini-app derived from the full-scale Bristol University Docking Engine (BUDE) molecular dynamics application [15]. The core virtualscreening algorithm computes the charge interactions between molecules when docked in different poses, a process used for drug discovery.…”

Section: B Minibudementioning

confidence: 99%

Comparing Julia to Performance Portable Parallel Programming Models for HPC

Lin

McIntosh–Smith

2021

2021 International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems (PMBS)

View full text Add to dashboard Cite

Julia is a general-purpose, managed, strongly and dynamically-typed programming language with emphasis on high performance scientific computing. Traditionally, HPC software development uses languages such as C, C++ and Fortran, which compile to unmanaged code. This offers the programmer near bare-metal performance at the expense of safety properties that a managed runtime would otherwise provide. Julia, on the other hand, combines novel programming language design approaches to achieve high levels of productivity without sacrificing performance while using a fully managed runtime.This study provides an evaluation of Julia's suitability for HPC applications from a performance point of view across a diverse range of CPU and GPU platforms. We select representative memory-bandwidth bound and compute bound mini-apps, port them to Julia, and conduct benchmarks across a wide range of current HPC CPUs and GPUs from vendors such as Intel ® , AMD ® , NVIDIA ® , Marvell ® , and Fujitsu ® . We then compare and characterise the results against existing parallel programming frameworks such as OpenMP ® , Kokkos, OpenCL ™ , and firstparty frameworks such as CUDA ® , HIP ™ , and oneAPI ™ SYCL ™ . Finally, we show that Julia's performance either matches the competition or is only a short way behind.

show abstract

“…In [7] the authors study OpenMP offload on NVIDIA V100 with a few mini-apps and various compilers, observe performance variations, and provide some OpenMP optimization techniques. In [8], the authors present the computebound mini-app miniBUDE and evaluate various programming models, including offload to GPUs. In [9], the authors present a performance analysis of CUDA, OpenACC, and OpenMP programming models on V100 GPU where they illustrate how it is easier to use OpenMP offloading and OpenACC compared to CUDA, and they measure the performance.…”

Section: Related Workmentioning

confidence: 99%

Evaluating GPU Programming Models for the LUMI Supercomputer

Markomanolis

Alpay

Young

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

It is common in the HPC community that the achieved performance with just CPUs is limited for many computational cases. The EuroHPC pre-exascale and the coming exascale systems are mainly focused on accelerators, and some of the largest upcoming supercomputers such as LUMI and Frontier will be powered by AMD Instinct™ accelerators. However, these new systems create many challenges for developers who are not familiar with the new ecosystem or with the required programming models that can be used to program for heterogeneous architectures. In this paper, we present some of the more well-known programming models to program for current and future GPU systems. We then measure the performance of each approach using a benchmark and a mini-app, test with various compilers, and tune the codes where necessary. Finally, we compare the performance, where possible, between the NVIDIA Volta (V100), Ampere (A100) GPUs, and the AMD MI100 GPU.

show abstract

A Performance Analysis of Modern Parallel Programming Models Using a Compute-Bound Application

Abstract: Please refer to any applicable terms of use of the publisher.

Cited by 16 publications

References 25 publications

Heterogeneous Programming for the Homogeneous Majority

Heterogeneous Programming for the Homogeneous Majority

Comparing Julia to Performance Portable Parallel Programming Models for HPC

Evaluating GPU Programming Models for the LUMI Supercomputer

Contact Info

Product

Resources

About