LLMs, autonomous AI agents speed up Nvidia chip design
Nvidia has trained a large language model LLM on Verilog and autonomous agents to speed up the design cycle for its GPU, CPU and networking chips.
AI models are supporting the work of chip designers by improving design quality and productivity, boosting the efficiency of manual processes and automating some time-consuming tasks. The models include prediction and optimization tools to help engineers rapidly analyze and improve designs, as well as LLMs that can assist engineers with answering questions, generating code, debugging design problems and more.
“Our researchers have basically developed a large language model that can be used to basically accelerate the creation of our Verilog code to describe our systems,” said Dave Salvator, director of accelerated computing products at Nvidia.
“We will be using this in future generations of our products to help build those out right and it can do a number of things. It can help basically speed up design and verification processes. It can speed up sort of the manual aspects of design and essentially automate a number of those tasks, and then the other piece of this is that you’ve heard us talk about our one GPU per year cadence.”
“If you’re familiar with the Blackwell architecture, you know that it is the most complex architecture we’ve ever built, and you know the architectures and the things we are building for future generations, you know, pose even bigger challenges in terms of what we’re trying to build. And so having a tool like this one will help us maintain that one product per year cadence while delivering you know, next generation architectures that continue to drive the state of the art forward.
Mark Ren, director of design automation research at Nvidia, will provide an overview of these models and their uses in a tutorial at the Hot Chips conference in the US this week. In a second session, he will focus on agent-based AI systems for chip design.
AI agents powered by LLMs can be directed to complete tasks autonomously. Researchers are developing agent-based systems that can reason and take action using customized circuit design tools, interact with experienced designers, and learn from a database of human and agent experiences.
Ren will share examples of how engineers can use AI agents for timing report analysis, cell cluster optimization processes and code generation. The cell cluster optimization work recently won best paper at the first IEEE International Workshop on LLM-Aided Design.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
