Last year, NVIDIA launched the cuLitho software library, which is expected to speed up photomask development by 40 times. Today, NVIDIA announced a partnership with TSMC and Synopsys to implement its computational lithography platform for production use and use the company’s next-generation Blackwell GPUs for AI and HPC applications.
Photomask development is a critical step for every chip, and NVIDIA’s cuLitho platform, enhanced with new generative AI algorithms, can significantly speed up this process. NVIDIA says computational lithography consumes tens of billions of hours on CPUs every year. By leveraging GPU-accelerated computational lithography, cuLitho provides significant improvements over traditional CPU-based methods. For example, 350 NVIDIA H100 systems can now replace 40,000 CPU systems, reducing production time, cost, and space and power requirements.
NVIDIA claims that its new generative AI algorithm provides an additional 2x acceleration on top of the already accelerated process implemented through cuLitho. This enhancement is particularly beneficial for the Optical Proximity Correction (OPC) process, which creates a near-perfect inverse mask to resolve light diffraction problems.
TSMC said that integrating cuLitho into its workflow has increased the speed of curve processes by 45 times and Manhattan processes by nearly 60 times. Curvilinear flow involves mask shapes represented by curves, while Manhattan mask shapes are limited to horizontal or vertical orientations.
Synopsys, a leading developer of electronic design automation (EDA), says its Proteus mask synthesis software, running on the NVIDIA cuLitho software library, accelerates computational workloads compared to current CPU-based methods. This acceleration is critical to enabling angstrom-scale scaling and reducing turnaround times in chip manufacturing.
The collaboration between NVIDIA, TSMC and Synopsys represents a significant advancement in semiconductor manufacturing, specifically the adoption of cuLitho. By leveraging accelerated computing and generative artificial intelligence, the partners are advancing what is possible in semiconductor scaling and opening up new innovation opportunities in chip design.