FPGA vs. GPU for Deep Learning —Introduction

Hello everyone,

I want to write an article about a subject that has been on my mind for a while. I think that doing this interactively, not alone, will enable me to deal with the subject in many ways. Although I have knowledge about FPGAs but I am a newbie of deep learning. If the industry is heading somewhere, we wouldn’t want to lag behind, would we?

When I saw Intel’s “Why Choose an FPGA for Deep Learning?” article, I decided to research this topic. I wonder if a big difference can really be made, or will we get the same result with more effort? Thinking about this topic, I started researching. I realized that I have many options in FPGA, such as RTL, Vivado HLS tool, DPU. This indicated that my research should be multi-faceted. My challenge would be not only between GPU and FPGA but also within FPGA. I decided to start with small steps. Simple projects may not help me compare in the first step, but I think it will help me get used to the environment. I think it would be helpful to start with one of the first CNNs, Lenet. I can find many examples on this subject and compare and quickly implement them.

Next week, we will start the Lenet implementation together in the article I will write next week. I’m waiting for articles and applications to help while I continue my research! So stay healthy until my next post. We have a lot to do and learn.