Grant: Discovery Grants Program - Individual
University of Guelph
$40,000.00
Using Artificial Intelligence to Enhance Field Programmable Gate Array Design Speed and Quality
EF 2024-25
Résumé
Field Programmable Gate Arrays (FPGAs) are integrated circuits composed of programmable logic blocks and interconnects that can be configured to implement any digital hardware circuit. Compared to other semiconductor technologies, FPGAs offer several advantages such as faster deployment, reduced cost, higher performance, and lower power consumption. Today, these advantages are being leveraged to implement a wide range of applications that are revolutionizing society. These include accelerating 4K video streaming workloads in data centers, performing sensor fusion and object detection in autonomous vehicles, enabling efficient medical imaging and gene processing, among many others. Nevertheless, FPGAs have their own challenges, among which are long placement and routing runtimes. Large, industrial designs can easily take hours (or even days) to place-and-route on modern FPGA devices. Furthermore, achieving design closure may require numerous design modifications increasing compilation times and adversely affecting the time-to-market advantage of FPGAs, thus limiting their full deployment.
The primary objectives of this research program are to decrease FPGA compilation times and expedite design closure. The emphasis is on placement and routing, which are the two most critical and time-consuming stages in the FPGA design process. This work combines the benefits of algorithmic optimizations with emerging artificial intelligence to develop transformative placement-and-routing tools that are intelligent, adaptive, and surpass existing tools in terms of speed and solution optimality. These innovative tools will be data-driven and will utilize machine learning, deep learning, reinforcement learning and graph neural networks to uncover hidden patterns that connect the design, FPGA device, and dynamic behaviour of placement-and-routing optimizations. These patterns will be used to achieve faster runtimes and improved quality-of-result for individual optimizations, as well as to automatically construct a placement-and-routing framework that accelerates design closure by selecting and utilizing the ideal optimizations and optimization strategy for a given design and target FPGA device. This is expected to result in innovative tools that learn and improve over time, produce near-optimal solutions, and significantly reduce design time. This will have a substantial impact on the research community, the Canadian semiconductor industry, and will benefit industries worldwide by accelerating the development process and enabling faster design-space exploration, leading to more efficient and effective use of FPGAs in various applications.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
RGPIN-2024-03720
Name nom
Grewal, Gary
Department département
Computer Science, School of - Computer Science, School of
Organization id
24
Institution établissement
University of Guelph
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
40000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1510
Committee name en
Electrical and Computer Engineering
Committee name fr
Génie électrique et informatique
Area of application code
801
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Computer software
Area of application fr
Logiciel
Research subject code
2519
Research subject group en
Electrical and electronic engineering
Research subject group fr
Génie électrique et électronique
Research subject en
Microelectronics
Research subject fr
Microélectronique
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Using Artificial Intelligence to Enhance Field Programmable Gate Array Design Speed and Quality
Keywords
field programmable gate arrays (fpga); computer aided design (cad); placement; routing; machine learning; deep learning; reinforcement learning; algorithm design and analysis; artificial intelligence; verilog-to-routing (vtr)
Application summary
Field Programmable Gate Arrays (FPGAs) are integrated circuits composed of programmable logic blocks and interconnects that can be configured to implement any digital hardware circuit. Compared to other semiconductor technologies, FPGAs offer several advantages such as faster deployment, reduced cost, higher performance, and lower power consumption. Today, these advantages are being leveraged to implement a wide range of applications that are revolutionizing society. These include accelerating 4K video streaming workloads in data centers, performing sensor fusion and object detection in autonomous vehicles, enabling efficient medical imaging and gene processing, among many others. Nevertheless, FPGAs have their own challenges, among which are long placement and routing runtimes. Large, industrial designs can easily take hours (or even days) to place-and-route on modern FPGA devices. Furthermore, achieving design closure may require numerous design modifications increasing compilation times and adversely affecting the time-to-market advantage of FPGAs, thus limiting their full deployment.
The primary objectives of this research program are to decrease FPGA compilation times and expedite design closure. The emphasis is on placement and routing, which are the two most critical and time-consuming stages in the FPGA design process. This work combines the benefits of algorithmic optimizations with emerging artificial intelligence to develop transformative placement-and-routing tools that are intelligent, adaptive, and surpass existing tools in terms of speed and solution optimality. These innovative tools will be data-driven and will utilize machine learning, deep learning, reinforcement learning and graph neural networks to uncover hidden patterns that connect the design, FPGA device, and dynamic behaviour of placement-and-routing optimizations. These patterns will be used to achieve faster runtimes and improved quality-of-result for individual optimizations, as well as to automatically construct a placement-and-routing framework that accelerates design closure by selecting and utilizing the ideal optimizations and optimization strategy for a given design and target FPGA device. This is expected to result in innovative tools that learn and improve over time, produce near-optimal solutions, and significantly reduce design time. This will have a substantial impact on the research community, the Canadian semiconductor industry, and will benefit industries worldwide by accelerating the development process and enabling faster design-space exploration, leading to more efficient and effective use of FPGAs in various applications.