TY - GEN
T1 - Effective Processor Model Generation from Instruction Set Simulator to Hardware Design
AU - Kappes, Johannes
AU - Kunzelmann, Robert
AU - Emrich, Karsten
AU - Foik, Conrad
AU - Mueller-Gritschneder, Daniel
AU - Ecker, Wolfgang
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - As the complexity of modern processing systems continues to increase, there has been a growing interest in the application of reusable code generators for hardware design. However, a formal system specification must be provided for such generators to be effective. Code generation has become a promising solution in the processor domain, where the Instruction Set Architecture (ISA) provides a natural system-level abstraction. In this paper, we present a holistic generator framework that produces variable microarchitecture designs, as well as relevant models for validation and formal verification. Our approach involves generating multiple functionally consistent models across different levels of abstraction using a unified formal ISA specification. By adhering to the 4-eyes principle of verification, we demonstrate that our framework minimizes the risk of common-mode errors that would typically escape the design verification process. Additionally, our code generators significantly reduce the manual development effort and generate high-performing instruction set simulators.
AB - As the complexity of modern processing systems continues to increase, there has been a growing interest in the application of reusable code generators for hardware design. However, a formal system specification must be provided for such generators to be effective. Code generation has become a promising solution in the processor domain, where the Instruction Set Architecture (ISA) provides a natural system-level abstraction. In this paper, we present a holistic generator framework that produces variable microarchitecture designs, as well as relevant models for validation and formal verification. Our approach involves generating multiple functionally consistent models across different levels of abstraction using a unified formal ISA specification. By adhering to the 4-eyes principle of verification, we demonstrate that our framework minimizes the risk of common-mode errors that would typically escape the design verification process. Additionally, our code generators significantly reduce the manual development effort and generate high-performing instruction set simulators.
KW - Code generation
KW - design verification
KW - formal specification
KW - instruction set simulation
KW - system validation
UR - http://www.scopus.com/inward/record.url?scp=85179503226&partnerID=8YFLogxK
U2 - 10.1109/NorCAS58970.2023.10305465
DO - 10.1109/NorCAS58970.2023.10305465
M3 - Conference contribution
AN - SCOPUS:85179503226
T3 - 2023 IEEE Nordic Circuits and Systems Conference, NorCAS 2023 - Proceedings
BT - 2023 IEEE Nordic Circuits and Systems Conference, NorCAS 2023 - Proceedings
A2 - Nurmi, Jari
A2 - Ellervee, Peeter
A2 - Koch, Peter
A2 - Moradi, Farshad
A2 - Shen, Ming
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 9th IEEE Nordic Circuits and Systems Conference, NorCAS 2023
Y2 - 31 October 2023 through 1 November 2023
ER -