An official website of the United States government
Over the past few years, there has been an increased interest in including FPGAs in data centers and high-performance computing clusters along with GPUs and other accelerators. As a result, it has become increasingly important to have a unified, high-level programming interface for CPUs, GPUs and FPGAs. This has led to the development of compiler toolchains to deploy OpenCL code on FPGA. However, the fundamental architectural differences between GPUs and FPGAs have led to performance portability issues: it has been shown that OpenCL code optimized for GPU does not necessarily map well to FPGA, often requiring manual optimizations to improve performance. In this paper, we explore the use of thread coarsening - a compiler technique that consolidates the work of multiple threads into a single thread - on OpenCL code running on FPGA. While this optimization has been explored on CPU and GPU, the architectural features of FPGAs and the nature of the parallelism they offer lead to different performance considerations, making an analysis of thread coarsening on FPGA worthwhile. Our evaluation, performed on our microbenchmarks and on a set of applications from open-source benchmark suites, shows that thread coarsening can yield performance benefits (up to 3-4x speedups) to OpenCL code running on FPGA at a limited resource utilization cost.
more »
« less
First Experiences in Performance Benchmarking with the New SPEChpc 2021 Suites
Modern High Performance Computing (HPC) systems are built with innovative system architectures and novel programming models to further push the speed limit of computing. The increased complexity poses challenges for performance portability and performance evaluation. The Standard Performance Evaluation Corporation (SPEC) has a long history of producing industry-standard benchmarks for modern computer systems. SPEC’s newly released SPEChpc 2021 benchmark suites, developed by the High Performance Group, are a bold attempt to provide a fair and objective benchmarking tool designed for stateof-the-art HPC systems. With the support of multiple host and accelerator programming models, the suites are portable across both homogeneous and heterogeneous architectures. Different workloads are developed to fit system sizes ranging from a few compute nodes to a few hundred compute nodes. In this work we present our first experiences in performance benchmarking the new SPEChpc2021 suites and evaluate their portability and basic performance characteristics on various popular and emerging HPC architectures, including x86 CPU, NVIDIA GPU, and AMD GPU. This study provides a first-hand experience of executing the SPEChpc 2021 suites at scale on production HPC systems, discusses real-world use cases, and serves as an initial guideline for using the benchmark suites.
more »
« less
- Award ID(s):
- 1814609
- PAR ID:
- 10366117
- Date Published:
- Journal Name:
- 2022 22nd International Symposium on Cluster, Cloud and Internet Computing (CCGrid)
- Page Range / eLocation ID:
- 675 to 684
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Scientific communities are increasingly adopting machine learning and deep learning models in their applications to accelerate scientific insights. High performance computing systems are pushing the frontiers of performance with a rich diversity of hardware resources and massive scale-out capabilities. There is a critical need to understand fair and effective benchmarking of machine learning applications that are representative of real-world scientific use cases. MLPerf ™ is a community-driven standard to benchmark machine learning workloads, focusing on end-to-end performance metrics. In this paper, we introduce MLPerf HPC, a benchmark suite of large-scale scientific machine learning training applications, driven by the MLCommons ™ Association. We present the results from the first submission round including a diverse set of some of the world’s largest HPC systems. We develop a systematic framework for their joint analysis and compare them in terms of data staging, algorithmic convergence and compute performance. As a result, we gain a quantitative understanding of optimizations on different subsystems such as staging and on-node loading of data, compute-unit utilization and communication scheduling enabling overall >10× (end-to-end) performance improvements through system scaling. Notably, our analysis shows a scale-dependent interplay between the dataset size, a system’s memory hierarchy and training convergence that underlines the importance of near-compute storage. To overcome the data-parallel scalability challenge at large batch-sizes, we discuss specific learning techniques and hybrid data-and-model parallelism that are effective on large systems. We conclude by characterizing each benchmark with respect to low-level memory, I/O and network behaviour to parameterize extended roofline performance models in future rounds.more » « less
-
Operational ocean forecasting systems (OOFSs) are complex engines that must execute ocean models with high performance to provide timely products and datasets. Significant computational resources are then needed to run high-fidelity models, and, historically, the technological evolution of microprocessors has constrained data-parallel scientific computation. Today, graphics processing units (GPUs) offer a rapidly growing and valuable source of computing power rivaling the traditional CPU-based machines: the exploitation of thousands of threads can significantly accelerate the execution of many models, ranging from traditional HPC workloads of finite difference, finite volume, and finite element modelling through to the training of deep neural networks used in machine learning (ML) and artificial intelligence. Despite the advantages, GPU usage in ocean forecasting is still limited due to the legacy of CPU-based model implementations and the intrinsic complexity of porting core models to GPU architectures. This review explores the potential use of GPU in ocean forecasting and how the computational characteristics of ocean models can influence the suitability of GPU architectures for the execution of the overall value chain: it discusses the current approaches to code (and performance) portability, from CPU to GPU, including tools that perform code transformation, easing the adaptation of Fortran code for GPU execution (like PSyclone), the direct use of OpenACC directives (like ICON-O), the adoption of specific frameworks that facilitate the management of parallel execution across different architectures, and the use of new programming languages and paradigms.more » « less
-
Badia, Rosa M; Mohror, Kathryn (Ed.)In contemporary high-performance computing architectures, the integration of GPU accelerators has become increasingly prevalent. To harness the full potential of these accelerators, developers often resort to vendor-specific kernel languages, such as CUDA. While this approach ensures optimal efficiency, it inherently compromises portability and engenders vendor dependency. Existing portable programming models, such as OpenMP, while promising, demand extensive code rewriting due to their foundamental difference from kernel languages. In this work, we introduce extensions to LLVM OpenMP, transforming it into a versatile and performance portable kernel language for GPU programming. These extensions allow for the seamless porting of programs from kernel languages to high-performance OpenMP GPU programs with minimal modifications. To evaluate our extension, we implemented a proof-of-concept prototype that contains a subset of extensions we proposed. We ported six established CUDA proxy and benchmark applications and evaluated their performance on both AMD and NVIDIA platforms. By comparing with native versions (HIP and CUDA), our results show that OpenMP, augmented with our extensions, can not only match but also in some cases exceed the performance of kernel languages, thereby offering performance portability with minimal effort from application developers.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/CCGrid54584.2022.00077
