Chipmaker NVIDIA released a new research study in which they use chatbots in the chip design process 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 tiny silicon wafers is one of the most difficult tasks in the tech industry, requiring tens of thousands of engineers to spend up to two years to complete.
The NVIDIA chip is so complex that it’s also at the heart of artificial intelligence technologies like ChatGPT.
This new research utilizes the big language models behind chatbots and data from NVIDIA’s 30-year archive of chip designs. One application uses the company’s long history of chip design to answer questions.
We’ve found that many senior designers spend a lot of time answering questions from junior designers,” said Bill Dally, NVIDIA’s chief scientist. Now we’re letting designers ask questions to chatbots, which saves senior designers a lot of time.”
The key finding of the study is that by adding a lot of company-specific data, even a regular chatbot can be more accurate than a senior chatbot. NVIDIA says this helps control system costs.
Another feature NVIDIA demonstrated was the use of artificial intelligence to generate code. Daley said engineers spend most of their time looking for possible problems in chips and using test tools to find out why.
The AI system can quickly generate scripts for test tools to test the designed chips.
Daley said, “The goal is not to automate processes or replace human roles, but to enhance the existing capabilities of employees and make them more efficient.”
THIS IS A SPONSOR PROMOTION: >>>>>>>>>>>>>
Geekwills is an online shop that connects consumers with millions of products and brands around the world with the mission to empower them to live their best lives. Geekwills is committed to offering the most affordable quality products to enable consumers and sellers to fulfill their dreams in an inclusive environment.
