Designing a RISC-V Based ASIP for Accelerating Third-Generation DNA Sequence Alignment on Embedded Systems
Team
- E/19/057, Chamath Colombage, email
- E/19/453, Akash Withanaarachchi, email
- E/17/083, Mahela Ekanayake, email
Supervisors
Table of content
- Abstract
- Related works
- Methodology
- Experiment Setup and Implementation
- Results and Analysis
- Conclusion
- Publications
- Links
Abstract
This project explores the design and implementation of a RISC-V-based Application-Specific Instruction-set Processor (ASIP) to accelerate DNA sequence alignment for embedded systems. By focusing on the computationally intensive chaining stage of Minimap2, the project aims to achieve real-time, energy-efficient processing for third-generation DNA sequencing technologies.
Related works
The project builds upon existing research in DNA sequence alignment, hardware acceleration, and RISC-V-based processor design. Key references include:
- Minimap2: A widely used tool for long-read DNA sequence alignment.
- Studies on hardware acceleration for genomics using FPGA and ASIC platforms.
- Open-source RISC-V ISA and its applications in custom processor design.
Methodology
- Performance Analysis: Profiling Minimap2 to identify bottlenecks in the chaining stage.
- ASIP Design: Developing a custom co-processor using the RISC-V Rocket Chip platform and RoCC interface.
- Simulation and Testing: Verifying the ASIP design using simulation tools and benchmarking its performance.
- Integration: Incorporating the ASIP into an embedded system for real-world testing.
Experiment Setup and Implementation
- Hardware: RISC-V Rocket Chip platform with RoCC interface.
- Software: Minimap2, Verilog/Chisel for hardware design, and Linux for embedded system testing.
- Tools: Simulation tools for performance evaluation and FPGA for prototyping.
Results and Analysis
- Performance Gains: Preliminary results indicate significant speedup in the chaining stage of Minimap2.
- Energy Efficiency: The ASIP design demonstrates reduced power consumption compared to traditional processors.
- Scalability: The approach is scalable to other stages of DNA sequence alignment.
Conclusion
The project successfully demonstrates the feasibility of using a RISC-V-based ASIP to accelerate DNA sequence alignment on embedded systems. Future work includes extending the design to other stages of Minimap2 and deploying the solution on FPGA for real-world applications.