How to Learn Chip Design with Open-Source EDA Tools

Spam, the promised content is behind a paywall.

It appeared from the large number of comments on HackerNews and LinkedIn that people shared the post's general stance. Particularly, the culture of secrecy appeared to mesh well with other people's experiences. The way the semiconductor EDA industry operates and the associated licensing fees were also widely disliked. Some said that understanding the basics is essential, which supports a theory-first approach to education. I continue to think that it is possible to teach strong fundamentals in a practical-first manner.

My claim that hardware and software paid about the same and that it's difficult to determine which pays more drew the greatest criticism. I have changed my position on this issue as a result of some strong reasons.

This is my updated stance.

    Most of the time, but not always, software does pay more than hardware especially if you normalize for the actual and opportunity cost of learning.

    Software is a better option because of the sheer number of job opportunities available, even if hardware engineering and software engineering jobs paid the same (except economic downturns, where everything is a toss-up).

    With a wealth of accessible resources and educational possibilities, software engineering offers a comparatively simpler route to up-skilling.

I’ll assume that you are only continuing to read this post because you like hardware engineering. Not everything revolves around work and money. It is satisfying to solve difficult problems. This is why individuals choose careers in science and mathematics.

Hardware engineering offers significant benefits. When scaled with sufficient effort, the steep learning curve works as its own skill-moat, making it difficult to breach. Electrical engineering and semiconductors are so diverse that it is possible to excel in many different things. This can serve as a basis for a long and fulfilling career.

In this post, we will look at actionable steps you can take to scale the learning curve in the chip design industry.

For free subscribers:

    A survey of EDA tools (trial and open-source) for learning chip design

        Ways to get access to industry-standard tools

        Going open source: simulators, solvers, layout editors, and PDKs

        TinyTapeout

        Getting involved with open-source silicon

For paid subscribers:

    Step-by-step actionable guide to get started with chip design:

        How to decide on a focus area and pick a project

        How to find circuits to design for projects — from simple circuits to cutting edge designs

        How to pick tools, plan and implement design projects

        What kinds of documentation you should maintain

        How to showcase your design skills to future employers

Read time: 16 mins

You can only learn chip design by doing. It does not matter how many YouTube videos of lectures you watch, or how many courses you download from MIT OpenCourseWare (OCW). Unless you try and build circuits from scratch, you will never develop the intuition and proficiency needed for chip-design.

Unfortunately, EDA tools are expensive and chip design companies pay millions of dollars every year to tool vendors like Synopsys, Cadence, Siemens, and Keysight. This causes a significant barrier to entry for new learners.

But fortunately, there are some ways around this. Ways to access industry-standard EDA tools