Sponsored Workshop: Verification of Inferencing Algorithm Accelerators

This short workshop will cover the verification of a custom AI accelerator as it is migrated from a machine learning framework in Python to RTL. Using High-Level synthesis provides a C++ version of the algorithm being verified. We will show how the original Python can be verified, and subsequent implementations, C++ and RTL, can be shown to be equivalent to the original Python. This allows for greater verification at the algorithmic level, thus requiring less verification at the RTL level. Please see attached proposal.

Russel Klein