Nvidia uses chat robots in the chip design process to improve communication and testing efficiency

Nvidia, a chipmaker, released a new study on Monday that uses chatbots in chip design to improve communication and testing efficiency.

Modern chips are large-scale integrated circuits made up of hundreds of billions of transistors, and arranging them on small silicon wafers is one of the most difficult tasks in the technology industry, which takes tens of thousands of engineers up to two years to complete.

Nvidia chips are very complex and are at the core of artificial intelligence technologies such as ChatGPT.

The new study uses the big language model behind chatbots and 30 years of chip design archive data from Nvidia. One of the applications is to use the company's long history of chip design to answer questions.

Bill Daley, chief scientist at Nvidia, said: "We found that many senior designers spend a lot of time answering questions from junior designers. Now we let designers ask chat robots questions, which can save senior designers a lot of time."

The key finding of the study is that even ordinary chatbots can be more accurate than advanced chatbots by adding large amounts of company-specific data. Nvidia says this helps control system costs.

Nvidia also demonstrated another feature, even using artificial intelligence to generate code. 'Engineers spend most of their time looking for possible problems in the chip and using testing tools to find out why, 'Mr. Daley said.

The artificial intelligence system can quickly generate the script of the test tool to test the designed chip.

"our goal is not to automate processes or replace people, but to improve the existing capabilities of our employees and make them more efficient," Daley said. "