Very impressive! Cheaper, more specialized, and better chips!
" ... a learning-based approach to chip placement, one of the most complex and time-consuming stages of the chip design process ... To achieve these results, we pose placement as a Reinforcement Learning (RL) problem and train an agent to place the nodes of a chip netlist onto a chip canvas. ... Our objective is to minimize PPA (power, performance, and area) ... "
Google claims its AI can design computer chips in under 6 hours | VentureBeat: In a paper, Google researchers reveal that they've developed an AI framework that can produce chip designs in less than six hours.
Here is the preprint paper:
https://arxiv.org/abs/2004.10746
Here is the blog:
https://ai.googleblog.com/2020/04/chip-design-with-deep-reinforcement.html
No comments:
Post a Comment