Hardware acceleration for fluid flow simulation
Abstract/Contents
- Abstract
- Over the past 35 years, the speed of fluid flow simulations reflected the increase in transistor densities as predicted by Moore's law. Recent improvements in compute capability, however, highlight a slowing trend: difficulties in extracting fine-grain parallelism, the divergence between flops and memory bandwidth, and the additional burden on solver developers who are now expected to manage complex memory hierarchies explicitly are all contributing factors. Yet the speed demands on fluid simulations for applications that require multiple evaluations including design exploration and optimization, uncertainty quantification, AI/ML, reduced-order modeling, digital twins, etc., are continuing to increase. In this dissertation we focus on the study of a possible alternative to bridge the gap between the predicted shortcomings in Moore's law and the requirements of many of these applications. We discuss the possibility of developing application-specific integrated circuits (ASIC) to solve the equations of fluid flow using a Lattice-Boltzmann Method (LBM) approach. We begin with a brief discussion of the LBM approach, focused on computation and communication requirements, and compare the relative performance and merits of implementations on GPUs, FPGAs, and ASIC. Preliminary conclusions indicate that performance improvements of 74x, relative to a state-of-the-art GPU, may be possible using and ASIC with feature sizes of 7 nm or smaller
Description
Type of resource | text |
---|---|
Form | electronic resource; remote; computer; online resource |
Extent | 1 online resource |
Place | California |
Place | [Stanford, California] |
Publisher | [Stanford University] |
Copyright date | 2020; ©2020 |
Publication date | 2020; 2020 |
Issuance | monographic |
Language | English |
Creators/Contributors
Author | Altoyan, Wesson Abdulaziz |
---|---|
Degree supervisor | Alonso, Juan José, 1968- |
Thesis advisor | Alonso, Juan José, 1968- |
Thesis advisor | Horowitz, Mark (Mark Alan) |
Thesis advisor | Murmann, Boris |
Degree committee member | Horowitz, Mark (Mark Alan) |
Degree committee member | Murmann, Boris |
Associated with | Stanford University, Department of Electrical Engineering |
Subjects
Genre | Theses |
---|---|
Genre | Text |
Bibliographic information
Statement of responsibility | Wesson Altoyan |
---|---|
Note | Submitted to the Department of Electrical Engineering |
Thesis | Thesis Ph.D. Stanford University 2020 |
Location | electronic resource |
Access conditions
- Copyright
- © 2020 by Wesson Abdulaziz Altoyan
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...