A summary of Runnemede: An Architecture for Ubiquitous High-Performance Computing

Nicholas P. Carter et al; https://doi.org/10.1109/HPCA.2013.6522319

For the summary of the paper, go to the Summary section of this article.

Table of Contents

• A summary of Runnemede: An Architecture for Ubiquitous High-Performance Computing
  • Table of Contents
  • First Pass
    • Category
    • Context
    • Contributions
  • Second Pass
    • Background Work
    • Motivation
    • Figures, Diagrams, Illustrations, and Graphs
    • Clarity
    • Relevant Work
    • Author Assumptions
      • Correctness
  • Summary
  • Summarization Technique
  • Citations

First Pass

Discussion about the title, abstract, introduction, section and sub-section headings, and conclusion

The paper Runnemede: An Architecture for Ubiquitous High-Performance Computing by Nicholas P. Carter et al. [1] describes the Runnemede high performance computing architecture targeting extreme-scale systems. This architecture was developed for the DARPA’s Ubiquitous High-Performance Computing program. The authors describe multiple facets of the architecture including the networking, hardware and software design, the energy efficiencies of the architecture. They also evaluate the performance of the architecture as well. Their many contributions are a theoretical architecture that is well optimized for energy efficiency on extra-scale computers.

Category

What type of paper is this work?

This is a theoretical paper describing an architecture for HPC systems.

Context

What other types of papers is the work related to?

Similar works would involve HPC architecture descriptions as well as low powered computing architectures as well.

Contributions

What are the author’s main contributions?

Their contributions are a theoretical design and analysis of a HPC architecture focused on energy efficiency.

Second Pass

Background Work

What has been done prior to this paper?

Runnemede is a one of four architectures under the DARPA UHPC program. Additionally, work has been done before to build both low powered cluster computers, and HPC.

Motivation

Why should we care about this paper?

The justification for this work is that there exists a theory that larger and larger HPC computers will require more and more power, without fully utilizing the entire device array. Additionally, a test chip was designed, but never produced, called “Sunshine”. By designing this chip, the authors were able to theoretically test the ideas presented in the paper as well as develop new ones for the architecture.

Figures, Diagrams, Illustrations, and Graphs

Are the axes properly labeled? Are results shown with error bars, so that conclusions are statistically significant?

All of the figures and tables are clear and easy to understand.

Clarity

Is the paper well written?

The paper is well written and clear to understand.

Relevant Work

Mark relevant work for review

The following relevant work can be found in the Citations section of this article.

• [2]

Author Assumptions

What assumptions does the author(s) make? Are they justified assumptions?

The author’s assumed that, “… The power consumed by logic is expected to scale well as feature sizes shrink, but not as well as transistor density, leading to the design of over provisioned, energy-limited systems that contain more hardware than they can operate simultaneously”. In other words, systems will have more and more power hungry hardware that cannot be utilized in its entirety. Additionally, they assume that the current trend with DRAM will cause power consumption to decrement over time, but not fast enough.

Correctness

Do the assumptions seem valid?

As the year is 2022, the current next generation hardware from NVIDIA, Intel, and AMD has been announced, all of which require immense power draw to operate. Additionally, DDR5 DRAM exists and consumes less power than the previous DDR4 DRAM. Therefore, I agree with the assumptions of the authors.

Summary

A summary of the paper

The paper Runnemede: An Architecture for Ubiquitous High-Performance Computing by Nicholas P. Carter et al. [1] describes the Runnemede high performance computing architecture targeting extreme-scale systems. This architecture was developed for the DARPA’s Ubiquitous High-Performance Computing program to address over provisioned, energy limited HPC architecture designs. The authors proposed a theoretical architecture design, and justify it via bench marking that they performed with simulations. Their work assumes (correctly in my opinion) that systems will continue to require more power to operate in order to achieve better performance.

Summarization Technique

This paper was summarized using a modified technique proposed by S. Keshav in his work How to Read a Paper [0].

Citations