Avatar

Alon Amid

Hardware/Software Engineer

Computer Architecture PhD

Biography

I am currently a senior SoC and deep learning systems architect at NVIDIA, working on server and data-center SoCs for deep learning systems. I was previously a senior hardware engineer in Microsoft Azure Hardware Architecture, working on cloud-scale hardware acceleration for artificial intelligence.

I earned my PhD in computer architecture from the UC Berkeley EECS department in 2021, with a dissertation titled ``Generator-Based Design of Custom Systems-on-Chip for Numerical Data Analysis”. I was supervised by Borivoje Nikolić and Krste Asanović, and affiliated with the ADEPT Lab (formerly ASPIRE), and the Berkeley Wireless Research Center (BWRC).

I am a recipient of the 2020-2021 Demetri Angelakos Memorial Achievement Award from the UC Berkeley EECS Department, and was active as the co-president and treasurer of the Electrical Engineering Graduate Student Association (EEGSA) in the UC Berkeley EECS Department and as a delegate to the UC Berkeley Graduate Assembly (GA).

I'm interested in most parts of the digital computing stack, from circuits to software. I believe that the end of Moore's law and the never-ending demand for additional computing resources require us to solve the interesting challenges of distributed, parallel, and heterogeneous computing in an energy efficient manner. These challenges span across the computing stack from the single-core and single-chip level up to clusters of thousands of many-core heterogeneous machines, and require system-level design.

Specifically, I've been recently focusing on data-parallel and numerical analysis applications, hardware-software co-design, large-scale distributed clusters, and the trade-offs between general purpose data parallel processors and custom specialized accelerators for emerging machine learning and data-analytics workloads.

I am a supporter of the open-source hardware movement as an innovation catalyst, and I try to contribute to the RISC-V ecosystem. As part of my PhD, I was a user and developer of the open-source “Berkeley ecosystem” for agile hardware design developed in the ASPIRE and ADEPT labs. I am a developer of the FireSim and Chipyard projects and was a contributor to the Hurricane/Eagle series of test chips, demonstrating new capabilities in generator-based SoC development.

Interests

  • Computer Architecture
  • Parallel Computing
  • Machine Learning
  • Energy Efficient Computing
  • Distributed Systems
  • Software Systems
  • Design Methodology and Automation

Education

  • PhD in Electrical Engineering and Computer Sciences, 2021

    University of California, Berkeley

  • MSc in Electrical Engineering and Computer Sciences, 2019

    University of California, Berkeley

  • BSc in Electrical Engineering, 2016

    Technion - Israel Institute of Technology

Recent Publications

Accelerating General-Purpose Linear Algebra on DNN Accelerators

Deep learning inference and training tasks are often accompanied by additional numerical data analysis tasks such as clustering, …
PDF

Gemmini: Enabling Systematic Deep-Learning Architecture Evaluation via Full-Stack Integration

DNN accelerators are often developed and evaluated in isolation without considering the cross-stack, system-level effects in real-world …
Preprint PDF

Accessible, FPGA Resource-Optimized Simulation of Multi-Clock Systems in FireSim

Given the complexity of modern systems-on-chip (SoCs), hardware-assisted verification is an integral part of the chip-design process. …
Preprint PDF

Vertically Integrated Computing Labs Using Open-Source Hardware Generators and Cloud-Hosted FPGAs

The design of computing systems has changed dramatically over the past decade, but most courses in advanced computer architecture …
Preprint PDF

COBRA: A Framework for Evaluating Compositions of Hardware Branch Predictors

We present COBRA, a framework which enables a realistic hardware-guided methodology for evaluating compositions of hardware branch …
Preprint PDF
See all publications

Recent & Upcoming Talks

Generator-Based Design of Custom Systems-on-Chips for Numerical Data Analysis

With the end of Moore's Law and Dennard scaling, the continuous demand for higher computing performance and efficiency is increasingly …
Jun 30, 2021 3:30 PM Berkeley, CA, USA
Slides

Vertically Integrated Computing Labs Using Open-Source Hardware Generators and Cloud-Hosted FPGAs

The design of computing systems has changed dramatically over the past decade, but most courses in advanced computer architecture …
May 28, 2021 6:00 PM Daegu, Korea

Secondary Use of DNN Accelerators on Edge SoCs (BLAS on Gemmini)

Deep learning models are often accompanied by additional numerical data analysis tasks such as clustering, dimensionality reduction, …
May 26, 2021 10:40 AM Virtual

Chipyard: An Integrated SoC Design, Simulation, and Implementation Environment - Release 1.4 Updates

Continued improvement in computing efficiency requires functional specialization of hardware designs. Agile design methodologies have …
Jan 15, 2021 9:00 AM Virtual

Poster: Chipyard – Integrated Design, Simulation, and Implementation of Custom RISC-V SoCs

We present Chipyard - an open-source integrated SoC design, simulation, and implementation environment for specialized RISC-V compute …
Aug 16, 2020 8:00 AM Virtual Hot Chips
See all talks

CV

alt text

*I currently work full-time as a senior SoC and deep learning systems architect at NVIDIA, working on server and data-center SoCs for deep learning systems.

*I previously worked full-time as a senior hardware engineer in Microsoft Azure Hardware Architecture, working on acceleration of artificial intelligence.

*During my PhD, I spent a summer as a research intern in Microsoft Azure working on accelerating data analytics systems using cloud-hosted FPGAs, and another summer interning at Google as a software engineering intern in datacenter platforms working on machine-learning-based solutions for predicting far-memory operations.

*I worked as an undergraduate student at Qualcomm Inc. in the milimeter waves 802.11ad System Team as a PHY ASIC Digital Design Engineer. We made cutting edge Wi-Fi (WiGig) chipsets in the 60 GHz band, leading the world into multi-Gbps Wi-Fi data rates.

*I was an Infosys Instep research intern at Infosys Labs in Bangalore, India. I worked on a research project about “Hybrid Access Control Based Solutions for Cloud Services”.

alt text

*I earned a PhD and a M.S. in electrical engineering and computer sciences at UC Berkeley.

*I completed a B.Sc. in electrical engineering (cum laude) in Technion - Israel Institute of Technology.

*I graduate high school from the Hebrew Reali School of Haifa.

alt text

*Served as an IDF (Israel Defense Forces) officer in an elite intelligence unit.

*If I tell you what I did, I'll have to kill you… but if you insist: I was a system officer, leading a team developing automatic intelligence analysis processes for an operational expert system. Later, I was a staff officer, responsible for strategic planning, analysis, and product development of technology mega-projects.

alt text

*A long time ago (after high school), I was a network administrator in B.M Carmel, helping administer the Priority ERP system, and helped set up the network and ERP system for a new remote site in Vermont.

*A longer time ago, I was a youth guide in Madatech - the Israel National Museum of Science (I got to meet Neil Armstrong there!)

alt text

*High Level Software: Python, Matlab, R, Shell (Bash), SQL, Scala

*Low Level Software: C, C++, CUDA, RISC-V Assembly

*Hardware Design: Verilog, VHDL, Chisel, Vivado, Synopsys/Cadence EDA tools

*Basic Profficiency: Perl, Pascal, Prolog, Visual Basic .NET, Java

alt text

I played trumpet in the Technion Symphony Orchestra, Haifa Youth Wind Orchestra, and the Anderson High School Trojan Marching Band in Austin, TX

alt text

English, Hebrew

alt text

I am fortunate enough to have called several awesome cities ``Home”:

Haifa (Israel), Portland (OR, USA), Austin (TX, USA), Melbourne (VIC, Australia), Berkeley (CA, USA)

You can also download my full résumé in PDF.

Misc

I love traveling. I've been to over 60 countries across 6 continents, and (hopefully) still going strong. Whether it's mountains, beaches, wildlife, outdoors, urban or historic, I'm always happy to explore new places and new cultures.

Antarctica is still an elusive destination, so if you know of a short-term opportunity in Antarctica, I'll be happy to hear about it.

Here is a sample of my travel map:

My dad got us a family domain a very long time ago (before it was popular). We had a web server in our house. We were also some of the first users of Google Apps (now known as Google G-suite). The web server no longer exists, but a small commemorative website is still out there: www.amidfamily.net

I was active in the Berkeley Hillel graduate student community. Throughout my time in Berkeley, I considered myself on the frontlines of fighting modern Anti-Semitism and Anti-Zionism on US college campuses (and specifically in UC Berkeley). If you have any questions about that, come and talk to me.

Contact