Bit Fusion for Acceleration of Deep Neural Networks Computation

This was my final project for my graduate ECE-564 course. The project uses hardware to process matrix multiplication such as how neural networks operate. I designed a very cost-effective and low area utilization using the Verilog hardware descriptive language. I then later synthesized the design using a program called Synopsys. This modeled the performance and efficiency of my design, and also allowed for estimation of manufacturing costs.

The project in its entirety can be found in the Git repository below.
(You must be logged in to view the contents of the repository)

https://git.robsengineeringadventures.com/rmhunter/bit-fusion

Leave a comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.