These blocks can be described by a netlist, a graph of circuit components, such as macros (memory components) and standard cells (logic gates like NAND, NOR, and XOR), all of which are connected by wires. While we show that we can generate optimized placements for Google accelerator chips ( TPUs), our methods are applicable to any kind of chip ( ASIC).Ī computer chip is divided into dozens of blocks, each of which is an individual module, such as a memory subsystem, compute unit, or control logic system. Whereas existing baselines require human experts in the loop and take several weeks to generate, our method can generate placements in under six hours that outperform or match their manually designed counterparts. In particular, as we train over a greater number of chip blocks, our method becomes better at rapidly generating optimized placements for previously unseen chip blocks. Unlike prior methods, our approach has the ability to learn from past experience and improve over time. In “ Chip Placement with Deep Reinforcement Learning”, we pose chip placement as a reinforcement learning (RL) problem, where we train an agent (i.e, an RL policy) to optimize the quality of chip placements. What if ML itself could provide the means to shorten the chip design cycle, creating a more integrated relationship between hardware and ML, with each fueling advances in the other? Dramatically shortening the chip design cycle would allow hardware to adapt to the rapidly advancing field of ML. However, today’s chips take years to design, resulting in the need to speculate about how to optimize the next generation of chips for the machine learning (ML) models of 2-5 years from now. With the slowing of Moore’s Law and Dennard scaling, the world is moving toward specialized hardware to meet the exponentially growing demand for compute. The revolution of modern computing has been largely enabled by remarkable advances in computer systems and hardware. Update, June 9, 2021: Today in Nature, we've published methods that improve on what is discussed below, and that have been used in production to design the next generation of Google TPUs. Posted by Anna Goldie, Senior Software Engineer and Azalia Mirhoseini, Senior Research Scientist, Google Research, Brain Team
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |