New techniques are being used to improve chip designs as manufacturing processes continue to mature and evolve. Researchers at Cornell University have applied AI in the floorplanning stage, where the memory block layout of the macroblock is decided. The AI not only designed it quicker than humans but also increased the potential speed and efficiency of the chip. This stage usually takes days or weeks, but the AI designed it in under six hours. It was trained with reinforcement learning to assess the state of the chip when making decisions. Google has already used these findings for its next-generation artificial-intelligence processors.
Success or failure in designing microchips depends heavily on steps known as floorplanning and placement. These steps determine where memory and logic elements are located on a chip. The locations, in turn, strongly affect whether the completed chip design can satisfy operational requirements such as processing speed and power efficiency.
Sources: Nature, Cornell University, TechPowerUp