Radiation hardened RISC-V processor
This thesis focuses on designing and validating an open-source 32-bit RISC-V processor and implementing using NTU’s in-house RHBD cell library. We inves tigate, analyse and compare the final layout in timing, area, and power based on three different libraries: (a) Full triple-module-redundancy (T...
Saved in:
Main Author: | |
---|---|
Other Authors: | |
Format: | Thesis-Master by Coursework |
Language: | English |
Published: |
Nanyang Technological University
2022
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/156208 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
Summary: | This thesis focuses on designing and validating an open-source 32-bit RISC-V
processor and implementing using NTU’s in-house RHBD cell library. We inves tigate, analyse and compare the final layout in timing, area, and power based on
three different libraries:
(a) Full triple-module-redundancy (TMR) technique using the GF 65nm library
(b) DICE library
(c) NTU’s in-house library
The results show that the RHBD RISC-V design using NTU’s in-house library
passes functional tests at 250MHz, and can run higher if a faster SRAM is used.
This suggests that NTU’s in-house library is applicable for high-speed applica tions.
The results also show that the implementation using our in-house library is about
33% smaller than the implementation using DICE and about 60% smaller than
the implementation applying full TMR using the GF 65nm library. In terms
of power consumption, the implementation using our in-house library is almost
equal to the implementation using the DICE library and 30% smaller than the
implementation applying full TMR in the worst case.
Overall, in the implementation of the RHBD RISC-V processor, we conclude that
the application of NTU’s in-house RHBD library is superior to reported RHBD
methodologies, including DICE and TMR. |
---|