Electronic Design Automation : What’s Next?
Today, we are surrounded by a plethora of computer devices which have become part of our every-day life. It is not uncommon for individuals to own multiple laptops, and other smart devices. This massive scaling in the number of personal electronics would not have been possible without Electronic Design Automation or EDA.
Electronic Design Automation Industry has been a crucial driving force behind the growth of the Semiconductor Industry. EDA companies deliver software products, tools, and platforms which assist in designing and verifying Integrated Circuits(ICs) and Printed Circuit Boards(PCB). Today, computers have billions of transistors and manual design and layout are not a feasible option anymore. EDA drastically reduces the cost and time as the chips can be tested, verified and simulated before the actual manufacturing.
To explore this industry from the lens of first design principles, we need to know a high-level overview of the chip design process. The process starts by writing the specifications for a chip using a Hardware Description Language. Chips can have millions of lines of specifications. The next step is to generate gates from these specifications and make the chips testable by using techniques like adding special gates to send signals to the chip. This is followed by simulations and several rounds of verification of the schematic. Next, we have the blueprint and actual layout of all the gates and their connections. As we can imagine, the sheer number of components have made it mandatory to use software automation for most if not all of these steps. EDA tools are also used in further steps for optimizing and making the layout ready for manufacturing.
Despite being at the heart of innovation, the EDA industry faces its own set of challenges which it needs to overcome in order to increase its growth rate. One of the major challenges that restrict the growth of the EDA industry is its limited customer base. The primary customers of EDA are big semiconductor companies, and with acquisitions happening inside the semiconductor industry EDA consumer base is further declining. While EDA companies try to solve complex design automation challenges of ever-evolving semiconductor technologies, it also needs to attempt to make software tools more powerful. This may include improving user interface, tools featuring version control, continuous integration, etc. Moreover, as the designs are becoming increasingly complex, the EDA tools are becoming more and more computation intensive. This mandates huge computing environments which reduces the tool’s affordability.
EDA should attempt to leverage the power of open source software, crowd-sourced developments, and standard platforms. Some computation intensive tools could be made available as web services to users. Another interesting future perspective is the design and verification of IoT devices. IoT consists of several connected devices that gather information from the environment and possibly performing some actuation. Majority of these individual devices are less complex and current EDA tools can be used for their design. However, a lot of these devices will be created by individuals, communities and small companies. Hence, the business model of EDA companies might change and the products would need to be more user-friendly, affordable and adaptable. Like several other domains, Machine learning algorithms can help in improving and adding features to EDA tool-chains like choosing the best available design options, optimize existing designs, improve circuit and power performance, etc.
EDA attempts to solve a really complex problem and requires knowledge from various other domains like physics, manufacturing processes, application software, etc. EDA tools have to support increasing chip design complexity, new semiconductor technologies, enhanced debugging and verification techniques while improving the performance of their tools and maintaining ease of use. However, since this industry is at the start of a chain of industries and innovations, it needs to reinvent itself and adopt newer technologies and better practices to enable rapid advances in other fields.
