Hardware Description Languages (HDLs) such as VHDL are very difficult to master. On the other hand, there has been a demand in the market for a simplified solution for developers as it is essential to utilize the computing performance of FPGAs in digital designs.

In this webinar, a methodology and examples for creating necessary specifications or defining functions using AI language models such as ChatGPT and generating VHDL code that meets the requirements were disclosed.

Liquid Instruments (hereinafter referred to as Liquid) recently hosted a webinar at e4ds EEWebinar on the topic of 'ChatGPT: Essential Strategies for Smarter Testing Using AI'.

FPGAs used in test and measurement are widely used because their necessity is recognized for their advantages in high-speed processing, custom design, flexibility, and reliability.
Ben Nizette, Product Director at Liquid Instruments, demonstrated how to build a custom hardware solution that dramatically simplifies the design of desired features by combining the capabilities of ChatGPT and Moku Cloud Compile.

I shared examples of how to generate VHDL using ChatGPT, implementing a signal limiter, implementing an event counter, and converting Python to VHDL: sqrt(x).

Each step is explained in the process of 'Prompt → Adjust → Copy → Build', and you only need to copy the code written by ChatGPT to Moku Cloud Compile. You can build your bitstream into Moku Cloud Compile, where you can edit the VHDL and improve your program.

Even if you are not a VHDL expert, you will have free access to 13 software-defined instruments from Moku, expanding the possibilities of instruments that can expand test capabilities with FPGAs, high-precision embedded instruments, and portable hardware.

Liquid explains that these examples are just a few, and that there is plenty of room for creation, from simple time-saving programs to complex prototypes that can be refined or iterated on through different prompts and methodologies.

“We are confident that future versions of ChatGPT will allow you to upload screenshots and even tell us where errors occurred,” said Ben Nijt, director of Moku Cloud Compile and ChatGPT. “This will make Moku Cloud Compile and ChatGPT a useful part of everyday testing scenarios.”