NVIDIA cuLitho is a library with optimized tools and algorithms designed to GPU-accelerate computational lithography and the semiconductor manufacturing process. This library significantly enhances the performance of computational lithography, a critical step in manufacturing computer chips, by leveraging the power of NVIDIA GPUs. Computational lithography involves complex computations related to electromagnetic physics, photochemistry, computational geometry, iterative optimization, and distributed computing. As silicon feature sizes become smaller, the need for advanced computational lithography techniques increases, making cuLitho an essential tool for the semiconductor industry.

Features

• Acceleration Libraries: Targets the emerging challenges of nanoscale computational lithography, providing up to 40X acceleration for inverse lithography technology (ILT).
• High Performance: Generates accurate photomasks faster, with a 40X performance speedup.
• Productivity: Processes photomasks that previously took two weeks overnight, enabling the generation of 3X to 5X more masks per day.
• Cost Savings: 500 NVIDIA Hopper™ GPU systems running cuLitho can perform the work of 40,000 CPU systems, using 1/9 the power and 1/8 the space.
• Future Silicon Scaling: Enables new lithography innovations to address the miniaturization of next-generation semiconductors, including subatomic modeling, curvilinear OPC, and high-numerical aperture extreme-ultraviolet (high-NA EUV) lithography.

Benefits

• Enhanced Performance: Delivers significant performance improvements, making computational lithography faster and more efficient.
• Increased Productivity: Reduces the time required to process photomasks, allowing for higher throughput and faster turnaround times.
• Cost Efficiency: Reduces the number of systems required, leading to substantial cost savings in power and space.
• Scalability: Supports the miniaturization of next-generation semiconductors, enabling continued advancements in semiconductor technology.