Nvidia boss pitches generative AI for chip manufacturing
Chip manufacturing is an “ideal application” for GPUs running generative AI algorithms says Jensen Huang, CEO of Nvidia.
Speaking at imec’s ITF World 2023 semiconductor conference in Antwerp, Belgium, by video, Huang points to customers such as KLA, Applied Materials and Hitachi.
The company is also working with TSMC, ASML and Synopsys to accelerate computational lithography. cuLitho is a software library with optimized tools and algorithms to simulate Maxwell’s equations to model the behaviour of light passing through optics and interacting with photoresists to improve chip manufacturing.
- ChatGPT leaking Samsung chip secrets is iceberg’s tip
- ChatGPT Gets Its “Wolfram Superpowers”
- Test tool adds ChatGPT for FPGA configuration .
Demand for computing capability is soaring. “As a result, global demand for cloud computing is causing data centre power consumption to skyrocket,” he said. This is driving processing into end equipment. Offloading and accelerating compute-intensive algorithms can speed up applications by 10-100x while reducing power and cost by an order of magnitude, says Huang (above).
This is where the “We are experiencing two simultaneous platform transitions, accelerated computing and generative AI,” Huang said. He points to advanced chip manufacturing that requires over 1,000 steps and each step must be nearly perfect to yield functional output.
“Sophisticated computational sciences are performed at every stage to compute the features to be patterned and to do defect detection for in-line process control,” said Huang. “Chip manufacturing is an ideal application for Nvidia accelerated and AI computing.”
Computational lithography is a key example. “We have already accelerated the processing by 50 times,” said Huang. “Tens of thousands of CPU servers can be replaced by a few hundred NVIDIA DGX systems, reducing power and cost by an order of magnitude.”
The savings will reduce carbon emissions or enable new algorithms to push beyond 2 nanometers, he said.
IMS Nanofabrication and NuFlare build mask writers to create the photomasks that that transfer patterns onto wafers using electron beams. Nvidia GPUs are used by D2S to accelerate the computationally demanding tasks of pattern rendering and mask process correction for these mask writers.
Semiconductor manufacturer TSMC and equipment providers KLA and Lasertech use extreme ultraviolet light (EUV) for 3nm processes and deep ultraviolet light (DUV) for other leading edge process, and GPUs are used to process classical physics modeling and deep learning to generate synthetic reference images and detect defects.
KLA, Applied Materials, and Hitachi High-Tech also use Nvidia GPUs in their e-beam and optical wafer inspection and review systems.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
