There has been a shortage of microchips globally resulting from the Covid-19 pandemic, which forced some Chip producing factories to shut down for a certain period. The companies are currently overwhelmed with the increased demand. Microchips are essential to electronic devices, and this shortage may affect the industry in the coming days.
However, there are some positive signs as tech giants invest more in microchip research and development. One search research that has yielded fruit is Alphabet Inc (NASDAQ: GOOGL) AI approach.
AI Produces six-month designs in Only 6 hours
Machine learning is helping Google design its next-generation machine learning chips. The algorithm’s designs are rapidly generated and are far more superior compared to the human-created ones.
Google engineers say AI completes work that takes humans six months in only 6 hours. For years, Google has made attempts to involve machine learning in the creation of chips. This latest move marks the first time its research has delivered a commercial product, the TPU (Tensor Processing Unit)
A publication in the journal Nature reports that Google engineers believe the chip industry is on its way to witnessing a significant impact. The development will help companies speed up exploring their architecture space while easily customizing chips for upcoming tasks.
Is this the end of Moore’s law?
An editorial piece in the same paper notes that what the research has achieved is monumental and could be the last nail in Moore’s Law coffin. Moore’s principle states that the number of transistors on a chip doubles every two years as the cost of computers is halved.
Whether AI will solve the challenge of adding more transistors to the chips is yet known, but it could find alternative performance paths.
Google’s new algorithm has found a solution to floorplanning, a process of determining the location of elements of memory and logic on a chip. The procedure which has defied all automation attempts
involves expert human engineers laboring for months deciding the exact location for components like CPU, GPU, and memory.
Thanks to AI, Floorplans designing will save engineers a load of time and effort.
Please make sure to read and completely understand our disclaimer at https://www.wallstreetpr.com/disclaimer. We may be compensated for posting this content on our website by EDM Media LLC. For questions, comments or suggestions please contact firstname.lastname@example.org.