Name: Ben Feinberg
Pronouns: he/him
Biography:
Ben Feinberg is a Senior Member of Technical Staff in the Scalable Computer Architecture group at Sandia National Laboratories. He works on a variety of topics related to energy-efficient and reliable computing with an emphasis on memory-centric computing and architectures for autonomous systems. Prior to joining Sandia in 2019, Ben completed his PhD in Electrical Engineering at the University of Rochester.
Institution/Lab: Sandia National Laboratories
Website:
SRP Collaboration Topic/Title: Software and System Architectures for Analog Linear Algebra Accelerators
Field or research area: Computer Systems
Please select all the topical areas that apply to your project:
Computer Science (i.e., architectures, compilers/languages, networks, workflow/edge, experiment automation, containers, neuromorphic computing, programming models, operating systems, sustainable software); Machine Learning and AI; National Security
Brief Abstract:
Analog computing has significant potential to enable new applications on autonomous systems by providing order-of-magnitude improvements in computational efficiency. Unfortunately, most of the work on analog hardware has focused on device and circuit optimization rather than flexible and programmable systems. To remedy this issue, this project will develop components of a system software toolchain including tools to effectively map applications to the target hardware and manage other runtime components. Additionally, this project will use Sandia’s Structural Simulation toolkit to evaluate the developed software on a simulated system and evaluate potential architectural optimizations. As part of this evaluation, we will consider trade offs between programmability and system-level performance and efficiency for a variety of applications. This project will involve interdisciplinary collaboration with both application and circuit/device experts to ensure that the developed toolchain is useful for application developers and provide feedback on architectural challenges that could be mitigated at the circuit or device level.
Desired relevant skills, background, or interests:
-Experience with C/C++ -Basic knowledge of computer architecture, compilers, or programming models -Interest in novel computing concepts
Other comments:
The project described above has numerous individual components which can be tailored to specific skills and interests of the SRP participant. I do not expect an SRP participant have knowledge or interest in all aspects of this project.
Do any special requirements apply? Minimum GPA (specify what GPA in comments below); other
Other, specify: Neither US Citizen or In-Person are required but both are strongly preferred
Keywords:
Computer Architecture; Computer Systems; System Architecture; Post-Moore Computing; Novel Architectures; Accelerators
Lightning Talk Title: Analog Accelerators for Linear Algebra