Artificial Intelligence (AI) have become more powerful over the years. So much so that Samsung confirmed it will be relying on AI to design its upcoming Exynos chips.
In a confirmation told to Wired, a Samsung spokesperson said that they will be using an AI software from Synopsis called DSO.ai to design its new and upcoming Exynos chips.
Going with Synopsys is no surprise considering that they have already partnered with a lot of other companies in the past and present. Albeit, Samsung hasn’t revealed yet if such chips are already into production and when they are coming out.
Using AI to design chips could become a new trend in the industry. In fact, Google has previously detailed using AI to help create its Tensor chips, which have been used to train and run AI programs for its data centers. Also, Google’s upcoming Pixel 6 series smartphones will be powered by an identical custom Tensor CPU.
Besides Samsung and Google, big-name tech hardware companies like NVIDIA and IBM are also playing with AI-backed chip designs.
Other than Synopsys, competitor company Cadence is also developing similar technology.
So, how does AI help design a chip? Brands like Synopsys are utilizing a machine-learning method called Reinforcement Learning to help work out the chip’s design. That includes the training of an algorithm to perform tasks through reward or punishment.
With the help of running simulations, this technology can help manufacturers draw the basics of a chip’s design, which includes the placement of the components and their wiring.
This could result in much smaller, powerful, and power-efficient processors in the future.
Via: Wired