Truly Heterogeneous HPC: Co-design to Achieve What Science Needs from HPC

Cardwell, Suma; Vineyard, Craig Michael; Severa, William; Chance, Frances S.; Rothganger, Frederick H.; Wang, Felix; Musuvathy, Srideep; Teeter, Corinne; Aimone, James B.

doi:10.1007/978-3-030-63393-6_23

Cited by 11 publications

(4 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Inside the GPU kernel, threads access a and b pointers from the kernel function parameters. When accessing addresses based on a and b pointers (e.g., a[0], a [1], etc. ), GPU threads emit loads and stores with the host pointer values as the base address.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Porting HPC Applications to AMD Instinct™ MI300A using Unified Memory and OpenMP®

Tandon,

Grinberg,

Bercea

et al. 2024

ISC High Performance 2024 Research Paper Proceedings (39th International Conference)

View full text Add to dashboard Cite

Instinct TM MI300A is the world's first data center accelerated processing unit (APU) with memory shared between the AMD "Zen 4" EPYC TM cores and third generation CDNA TM compute units. A single memory space offers several advantages: i) it eliminates the need for data replication and costly data transfers, ii) it substantially simplifies application development and allows an incremental acceleration of applications, iii) is easy to maintain, and iv) its potential can be well realized via the abstractions in the OpenMP® 5.2 standard, where the host and the device data environments can be unified in a more performant way. In this article, we provide a blueprint of the APU programming model leveraging unified memory and highlight key distinctions compared to the conventional approach with discrete GPUs. OpenFOAM®, an open-source C++ library for computational fluid dynamics, is presented as a case study to emphasize the flexibility and ease of offloading a full-scale production-ready application on MI300 APUs using directivebased OpenMP programming.

show abstract

Section: Related Workmentioning

confidence: 99%

“…High-performance computing (HPC) node architecture designs are driven primarily by power management considerations and are increasingly reliant on high degrees of finegrained parallelism [1], [2]. Top500 list [3] mentions over 150+ known systems that use accelerator/co-processor technology.…”

Section: Introductionmentioning

confidence: 99%

Porting HPC Applications to AMD Instinct™ MI300A using Unified Memory and OpenMP®

Tandon,

Grinberg,

Bercea

et al. 2024

ISC High Performance 2024 Research Paper Proceedings (39th International Conference)

View full text Add to dashboard Cite

show abstract

“…When we design an HPC system and environment, multiple scientists, software developers, and hardware developers need to discuss an optimal system with collaborations; hardware based on existing technology and software running on it as well as new hardware and software not yet emerging result in challenging scientific goals (Cardwell et al) [38][39][40]. The best ideal HPC system requires the shared development of actionable information that engages all communities in charge and the values guiding their engagement.…”

Section: Co-designmentioning

confidence: 99%

High-Performance Computing in Meteorology under a Context of an Era of Graphical Processing Units

Nakaegawa

2022

Computers

View full text Add to dashboard Cite

This short review shows how innovative processing units—including graphical processing units (GPUs)—are used in high-performance computing (HPC) in meteorology, introduces current scientific studies relevant to HPC, and discusses the latest topics in meteorology accelerated by HPC computers. The current status surrounding HPC is distinctly complicated in both hardware and software terms, and flows similar to fast cascades. It is difficult to understand and follow the status for beginners; they need to overcome the obstacle of catching up on the information on HPC and connecting it to their studies. HPC systems have accelerated weather forecasts with physical-based models since Richardson’s dream in 1922. Meteorological scientists and model developers have written the codes of the models by making the most of the latest HPC technologies available at the time. Several of the leading HPC systems used for weather forecast models are introduced. Each institute chose an HPC system from many possible alternatives to best match its purposes. Six of the selected latest topics in high-performance computing in meteorology are also reviewed: floating points; spectral transform in global weather models; heterogeneous computing; exascale computing; co-design; and data-driven weather forecasts.

show abstract

“…In the context of Edge computing, domain-specific accelerators (ASICs and FPGA) and generalpurpose processors are commonly used together to perform near-data processing, thereby, unlocking various real-time use cases (e.g., edge AI and AR/VR applications) [2], [3]. In the HPC context, deploying various machine types with different architectures on HPC boards to fulfill the power and performance requirements is becoming a trend [7].…”

Section: Introductionmentioning

confidence: 99%

E2C: A Visual Simulator to Reinforce Education of Heterogeneous Computing Systems

Mokhtari¹,

Drake²,

Huynh³

et al. 2023

Preprint

View full text Add to dashboard Cite

Heterogeneity has been an indispensable aspect of distributed computing throughout the history of these systems. In particular, with the increasing popularity of accelerator technologies (e.g., GPUs and TPUs) and the emergence of domain-specific computing via ASICs and FPGA, the matter of heterogeneity and understanding its ramifications on the system performance has become more critical than ever before. However, it is challenging to effectively educate students about the potential impacts of heterogeneity on: (a) the performance of distributed systems; and (b) the logic of resource allocation methods to efficiently utilize the resources. Making use of the real infrastructure (such as those offered by the public cloud providers) for benchmarking the performance of heterogeneous machines, for different applications, with respect to different objectives, and under various workload intensities is cost-and time-prohibitive. Moreover, not all students (globally and nationally) have access or can afford such real infrastructure. To reinforce the quality of learning about various dimensions of heterogeneity, and to decrease the widening gap in education, we develop an open-source simulation tool, called E2C , that can help students researchers and practitioners to study any type of heterogeneous (or homogeneous) computing system and measure its performance under various system configurations. To make the learning curve shallow, E2C is equipped with an intuitive graphical user interface (GUI) that enables its users to easily examine system-level solutions (scheduling, load balancing, scalability, etc.) in a controlled environment within a short time and at no cost. In particular, E2C is a discrete event simulator that offers the following features: (i) simulating a heterogeneous computing system; (ii) implementing a newly developed scheduling method and plugging it into the system, (iii) measuring energy consumption and other output-related metrics; and (iv) powerful visual aspects to ease the learning curve for students. We used E2C as an assignment in the Distributed and Cloud Computing course. Our anonymous survey study indicates that students rated E2C with the score of 8.7 out of 10 for its usefulness in understanding the concepts of scheduling in heterogeneous computing. Moreover, our pre-and post-evaluations indicate that E2C has improved the students' understanding of heterogeneous computing systems by around 18%.

show abstract

Truly Heterogeneous HPC: Co-design to Achieve What Science Needs from HPC

Cited by 11 publications

References 52 publications

Porting HPC Applications to AMD Instinct™ MI300A using Unified Memory and OpenMP®

Porting HPC Applications to AMD Instinct™ MI300A using Unified Memory and OpenMP®

High-Performance Computing in Meteorology under a Context of an Era of Graphical Processing Units

E2C: A Visual Simulator to Reinforce Education of Heterogeneous Computing Systems

Contact Info

Product

Resources

About