CAREER: Fast, Energy Efficient Irregular Kernels via Neural Accerlation
职业:通过神经加速实现快速、节能的不规则内核
基本信息
- 批准号:2044633
- 负责人:
- 金额:$ 47.92万
- 依托单位:
- 依托单位国家:美国
- 项目类别:Continuing Grant
- 财政年份:2021
- 资助国家:美国
- 起止时间:2021-04-01 至 2026-03-31
- 项目状态:未结题
- 来源:
- 关键词:
项目摘要
High-performance computing suffers from a performance bottleneck that wastes computation time, money, and energy, as processing cores on multicore systems sit idle waiting for memory accesses from irregular kernels. These irregular kernels normally accomplish little computational work despite the high cost of accessing memory. These costly bottlenecks must be remedied by a new approach to high-performance computing. But, at the same time, computing is evolving and is becoming less dependent on the low-level programming languages that cause these bottlenecks and more dependent on learning algorithms such as neural networks to attain the necessary efficiency. This project builds the foundation for accelerating irregular kernels by replacing them with neural networks that run on accelerators optimized for neural networks. These neural networks offer better performance and energy consumption. Additionally, these networks are tuned in high-level programming languages (e.g., Python) that are easier for novice users to learn. This allows more computer scientists to aid the scientific and high-performance computing communities. This project also builds a new curriculum such as adding neural accelerators and expanding neural network algorithm materials into traditional undergraduate courses. This project, in both its research and educational aspects, significantly reduces the development time and costs of high-performance computing while simultaneously reducing performance bottlenecks. Furthermore, this project will support graduate and undergraduate students as they engage in cross-disciplinary involvement to match accuracy and performance constraints from the scientific-modeling and big-data-analysis communities that currently depend on irregular kernels for areas such as climate modeling, large scale circuit design, and drug analysis on infectious diseases. The goals and scope of this project are to build a framework that allows irregular kernels to be optimized in terms of both their performance and energy usage using the technique of neural acceleration, i.e., being represented and executed as a neural network. The methods used to meet the project’s goals and scope include the following: 1) The development of an approximation-bound characteristic that quantifies and qualifies acceptable error bars on the developed neural networks along with performance and energy requirements; 2) The development of initial neural networks for commonly used irregular kernels that can be used as starting networks for more complex irregular kernels and be used by individuals tuning their irregular kernels (which will be made available by a public database that is created and maintained by the investigator to support research in this area); and 3) The construction of a toolchain to aid in identifying irregular kernels in code, constructing neural networks based on user input, and deciding how the neural networks should be scheduled. The deliverable toolchain has support for popular libraries like TensorFlow and will be disseminated via an open-source repository. The transformative impact of this project’s effort generates a completely new optimization option for irregular kernels and a base set of tools (i.e., a public database and scheduling toolchain) that will foster future advances into using neural acceleration for various codes and lead to significant advancements in science and engineering. As such, this new optimization option may inspire a new computational model in a post-Moore era that provides timely scientific data for urgent government policy, such as climate change and foreign affairs.This project is jointly funded by CAREER Software and Hardware Foundations HPC program and the Established Program to Stimulate Competitive Research (EPSCoR).This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
高性能计算受到性能瓶颈的困扰,这会浪费计算时间、金钱和能量,因为多核系统上的处理内核会处于空闲状态,等待来自不规则内核的内存访问。尽管访问内存的成本很高,但这些不规则的内核通常只完成很少的计算工作。必须通过一种新的高性能计算方法来弥补这些代价高昂的瓶颈。但是,与此同时,计算也在不断发展,越来越不依赖于导致这些瓶颈的低级编程语言,而更多地依赖于神经网络等学习算法来获得必要的效率。这个项目通过用运行在为神经网络优化的加速器上的神经网络代替不规则内核,为加速不规则内核奠定了基础。这些神经网络提供了更好的性能和能耗。此外,这些网络使用高级编程语言(例如Python)进行调优,使新手更容易学习。这使得更多的计算机科学家能够帮助科学和高性能计算社区。本项目还建立了新的课程,如在传统的本科课程中增加神经加速器和扩展神经网络算法材料。该项目在研究和教育方面都显著减少了高性能计算的开发时间和成本,同时减少了性能瓶颈。此外,该项目将支持研究生和本科生从事跨学科参与,以匹配科学建模和大数据分析社区的准确性和性能限制,这些社区目前依赖于不规则核,如气候建模、大规模电路设计和传染病药物分析。该项目的目标和范围是建立一个框架,允许不规则内核在性能和能量使用方面使用神经加速技术进行优化,即作为神经网络表示和执行。用于满足项目目标和范围的方法包括以下内容:1)开发一种近似边界特征,该特征可以量化和限定已开发神经网络上的可接受误差条以及性能和能量要求;2)为常用的不规则核开发初始神经网络,可以用作更复杂的不规则核的起始网络,并可用于个人调整其不规则核(这将通过由研究者创建和维护的公共数据库提供,以支持该领域的研究);3)构建工具链,以帮助识别代码中的不规则核,根据用户输入构建神经网络,并决定如何调度神经网络。可交付的工具链支持流行的库,如TensorFlow,并将通过开源存储库传播。该项目的变革性影响为不规则内核产生了一个全新的优化选项和一套基本工具(即公共数据库和调度工具链),这将促进未来在各种代码中使用神经加速的进步,并导致科学和工程领域的重大进步。因此,这种新的优化选择可能会在后摩尔时代激发一种新的计算模型,为紧急的政府政策(如气候变化和外交事务)提供及时的科学数据。本项目由职业软件和硬件基金会HPC项目和促进竞争性研究的既定项目(EPSCoR)共同资助。该奖项反映了美国国家科学基金会的法定使命,并通过使用基金会的知识价值和更广泛的影响审查标准进行评估,被认为值得支持。
项目成果
期刊论文数量(1)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
Neural Acceleration of Graph Based Utility Functions for Sparse Matrices
稀疏矩阵的基于图的效用函数的神经加速
- DOI:10.1109/access.2023.3262453
- 发表时间:2023
- 期刊:
- 影响因子:3.9
- 作者:Booth, Joshua Dennis;Bolet, Gregory S.
- 通讯作者:Bolet, Gregory S.
{{
item.title }}
{{ item.translation_title }}
- DOI:
{{ item.doi }} - 发表时间:
{{ item.publish_year }} - 期刊:
- 影响因子:{{ item.factor }}
- 作者:
{{ item.authors }} - 通讯作者:
{{ item.author }}
数据更新时间:{{ journalArticles.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ monograph.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ sciAawards.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ conferencePapers.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ patent.updateTime }}
Joshua Booth其他文献
Nerve fibre organisation in the human optic nerve and chiasm: what do we really know?
人类视神经和视交叉中的神经纤维组织:我们究竟了解多少?
- DOI:
10.1038/s41433-024-03137-7 - 发表时间:
2024-06-07 - 期刊:
- 影响因子:3.200
- 作者:
Pratap R. Pawar;Joshua Booth;Andrew Neely;Gawn McIlwaine;Christian J. Lueck - 通讯作者:
Christian J. Lueck
Joshua Booth的其他文献
{{
item.title }}
{{ item.translation_title }}
- DOI:
{{ item.doi }} - 发表时间:
{{ item.publish_year }} - 期刊:
- 影响因子:{{ item.factor }}
- 作者:
{{ item.authors }} - 通讯作者:
{{ item.author }}
{{ truncateString('Joshua Booth', 18)}}的其他基金
Collaborative Research: SHF: Small: Learning Fault Tolerance at Scale
合作研究:SHF:小型:大规模学习容错
- 批准号:
2135310 - 财政年份:2022
- 资助金额:
$ 47.92万 - 项目类别:
Standard Grant
相似国自然基金
基于FAST搜寻及观测的脉冲星多波段辐射机制研究
- 批准号:12403046
- 批准年份:2024
- 资助金额:0 万元
- 项目类别:青年科学基金项目
FAST连续观测数据处理的pipeline开发
- 批准号:
- 批准年份:2024
- 资助金额:0.0 万元
- 项目类别:省市级项目
基于神经网络的FAST馈源融合测量算法研究
- 批准号:12363010
- 批准年份:2023
- 资助金额:31 万元
- 项目类别:地区科学基金项目
使用FAST开展河外中性氢吸收线普查
- 批准号:12373011
- 批准年份:2023
- 资助金额:52.00 万元
- 项目类别:面上项目
基于FAST的射电脉冲星搜索和候选识别的深度学习方法研究
- 批准号:12373107
- 批准年份:2023
- 资助金额:54 万元
- 项目类别:面上项目
基于FAST观测的重复快速射电暴的统计和演化研究
- 批准号:12303042
- 批准年份:2023
- 资助金额:30 万元
- 项目类别:青年科学基金项目
利用FAST漂移扫描多科学目标同时巡天宽带谱线数据研究星系中性氢质量函数
- 批准号:12373012
- 批准年份:2023
- 资助金额:52.00 万元
- 项目类别:面上项目
基于FAST望远镜及超级计算的脉冲星深度搜寻和研究
- 批准号:12373109
- 批准年份:2023
- 资助金额:55.00 万元
- 项目类别:面上项目
基于FAST高灵敏度和高谱分辨中性氢数据的暗星系的系统搜寻与研究
- 批准号:12373001
- 批准年份:2023
- 资助金额:52.00 万元
- 项目类别:面上项目
基于FAST的纳赫兹引力波研究
- 批准号:LY23A030001
- 批准年份:2023
- 资助金额:0.0 万元
- 项目类别:省市级项目
相似海外基金
CAREER: Unary Computing in Memory for Fast, Robust and Energy-Efficient Processing
职业:内存中的一元计算,实现快速、稳健和节能的处理
- 批准号:
2339701 - 财政年份:2024
- 资助金额:
$ 47.92万 - 项目类别:
Continuing Grant
An innovative hybrid infrastructure system delivering both electric and hydrogen for vessel fast charging/refuelling using off grid renewable energy and onsite wastewater.
一种创新的混合基础设施系统,利用离网可再生能源和现场废水,为船舶快速充电/加油提供电力和氢气。
- 批准号:
10099143 - 财政年份:2024
- 资助金额:
$ 47.92万 - 项目类别:
Collaborative R&D
DMREF: Design of fast energy storage pseudocapacitive materials
DMREF:快速储能赝电容材料的设计
- 批准号:
2324326 - 财政年份:2023
- 资助金额:
$ 47.92万 - 项目类别:
Standard Grant
Energy filter with fast-counting direct electron detector
带有快速计数直接电子探测器的能量过滤器
- 批准号:
525040890 - 财政年份:2023
- 资助金额:
$ 47.92万 - 项目类别:
Major Research Instrumentation
SBIR Phase I: Fast and low-energy manufacturing of high-performance, fiber-reinforced composites
SBIR 第一阶段:快速、低能耗地制造高性能纤维增强复合材料
- 批准号:
2304621 - 财政年份:2023
- 资助金额:
$ 47.92万 - 项目类别:
Standard Grant
Structure formation and energy relaxation by counter fast electron flow and background plasma
通过反快速电子流和背景等离子体形成结构和能量弛豫
- 批准号:
23H01150 - 财政年份:2023
- 资助金额:
$ 47.92万 - 项目类别:
Grant-in-Aid for Scientific Research (B)
CDS&E: Data-driven fast methods for high-energy plasma astrophysics
CDS
- 批准号:
2307684 - 财政年份:2023
- 资助金额:
$ 47.92万 - 项目类别:
Standard Grant
Energy And Chemical Gas Chromatography For Fast Refinery Gas Analysis
用于快速炼厂气分析的能源和化学气相色谱法
- 批准号:
RTI-2023-00336 - 财政年份:2022
- 资助金额:
$ 47.92万 - 项目类别:
Research Tools and Instruments
CAREER: Fast-Charging Energy Storage Devices Enabled by Modulating Internal Electric Field of Heterostructure
职业:通过调制异质结构内部电场实现快速充电储能装置
- 批准号:
2144708 - 财政年份:2022
- 资助金额:
$ 47.92万 - 项目类别:
Continuing Grant
Advancing Machine-Learning Augmented Free-Energy Density Functionals for Fast and Accurate Quantum Simulations of Warm Dense Plasmas
推进机器学习增强自由能密度泛函,以实现快速、准确的热致密等离子体量子模拟
- 批准号:
2205521 - 财政年份:2022
- 资助金额:
$ 47.92万 - 项目类别:
Standard Grant