Imperas Presentation
summary
Shows using Imperas’ instruction-accurate simulator to profile a full C application, find hot spots, and convert hot routines into C-modeled RISC-V custom instructions, with the ChaCha20 example quantified end to end. Workflow includes cycle-accurate timing annotations, simulation of timing-annotated cycles, trace and block-level debug, code and line coverage, compliance tests, and auto-generated documentation for handoff. Compares baseline vs custom-instruction performance and iterates until the instruction set and timing meet target gains.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 0:46Two types of developers
- 1:12Custom instructions
- 2:19Design flow
- 3:07Development environment
- 3:25Custom instruction simulation
- 4:03Custom instruction performance
- 5:26Debugging
- 6:16Further analysis
- 6:46Documentation
- 7:13Summary
- 7:36Outro
related talks
11mGetting Started with RISC-V Custom Instructions
11mGetting Started with RISC-V Custom Instructions
06mImperas Interview
10mImperasFPM Fast Processor Models
15mA Common Software Development Environment for Many-core RISC-V based Hardware and Virtual Platforms
09mDemo: Imperas/OVP RISC-V Models: Not Your Grandparents’ ISS
21mThe New Verification Ecosystem that Supports RISC-V Verification fo...
13mIntroduction to RISC-V Processor Verification
