Embedded Systems

Modeling the Nvidia NVDLA Machine Learning Accelerator in ACADL for Performance Predictions

Bachelor’s Thesis / Master’s Thesis / Student Research Project

Abstract

Abstract modeling of HW/SW systems is a relatively new research topic. This technique aims to capture only the essential parameters of software and hardware that influence their timing behavior.

This student project’s goal is to model the Nvidia NVDLA Machine Learning Accelerator using the Python-based Abstract Computer Architecture Description Language (ACADL) and use different methods for runtime/performance prediction and compare those against the cycle-accurate model.

Block diagram of the NVDLA Architecture (source)

NVDLA Accelerator

References

Requirements

  • Python
  • C++
  • Linux
  • Deep Learning
  • Successfully atteded the lecture “Grundlagen der Rechnerarchitektur” and/or “Parallele Rechnerarchitekturen” (optional)

Contact

Lübeck, Konstantin

Jung, Alexander

Bringmann, Oliver