Samsung was the first among processor manufacturers to create them using artificial intelligence

In mid-August 2021, it became known that the company Samsung began to use software artificial intelligence Synopsys with to create. processors Although other companies, including and, Google Nvidia previously tried to use AI in the development of chips, Samsung turned out to be the first manufacturer to bring this project to a massive level.

Samsung introduced the Synopsys DSO.ai tool, which can accelerate the development of semiconductors and discover new chip designs. Samsung said that the company uses Synopsys software to develop its Exynos chips, which are used in smartphones, as well as in other gadgets.

Samsung began to create processors using artificial intelligence

Mike Demler, a senior analyst at Linley Group, specializing in the semiconductor industry, says AI is well suited for organizing billions of transistors in a chip, and it will also become a standard part of a suite of computing tools. The use of AI is usually expensive, since it takes a lot of cloud computing power to train a powerful algorithm. But he expects that as the cost of computing decreases and the efficiency of models increases, they will become more affordable. He adds that many tasks related to the design of chips cannot be automated, so specialists are still needed.

The AI approach used by Synopsys, as well as Google, Nvidia and IBM, uses a machine learning method called reinforcement training to develop chip design. Reinforcement training involves training the algorithm to perform a task through reward or punishment and it has proven to be an effective way to capture subtle and difficult-to-modify human judgments.

The method can automatically form the basis of a design, including the placement of components and how to connect them, trying different designs in the simulation and studying which ones give the best results. This can speed up the chip design process and allow the engineer to experiment more effectively with new designs. In a blog post in June 2021, Synopsys announced that one North American integrated circuit manufacturer improved chip performance by 15% with this software.

Stelios Diamantidis, Senior Director of AI Solutions at Synopsys, says the software can DSO.ai be prioritized for different goals, such as performance or energy efficiency.^[1]

Notes