Today's AI/ML headlines are brought to you by ThreatPerspective

Digital Event Horizon

TSMC and NVIDIA Transform Semiconductor Manufacturing With Accelerated Computing




TSMC and NVIDIA Partner to Revolutionize Computational Lithography with cuLitho Platform

The collaboration between TSMC, the world-renowned semiconductor manufacturing leader, and NVIDIA, a pioneer in artificial intelligence and computing technology, marks a significant breakthrough in the field of computational lithography. The introduction of the cuLitho platform, a GPU-accelerated computational lithography solution, promises to accelerate the development of next-generation semiconductor chips and push the limits of physics.



  • TSMC and NVIDIA have joined forces to develop a revolutionary platform for accelerating computational lithography.
  • The cuLitho platform utilizes NVIDIA's H100 Tensor Core GPU-based systems to accelerate computational lithography, significantly reducing production time while minimizing costs and space requirements.
  • NVIDIA has developed innovative algorithms that apply generative AI to enhance the value of the cuLitho platform, enabling a 2x speedup on top of accelerated processes.
  • The application of generative AI to optical proximity correction has revolutionized the field, allowing for more accurate simulation of physics and realization of mathematical techniques.
  • The integration of generative AI into the cuLitho platform enables exploration of previously inaccessible calculations, paving the way for significant breakthroughs in computer chip design and development.



  • In a groundbreaking collaboration, TSMC and NVIDIA have joined forces to develop a revolutionary platform for accelerating computational lithography. This partnership has far-reaching implications for the semiconductor industry, as it enables the faster development of next-generation technology nodes and computer architectures.

    Computational lithography is a critical step in the manufacture of computer chips, involving the transfer of circuitry onto silicon. It requires complex computation, encompassing electromagnetic physics, photochemistry, computational geometry, iterative optimization, and distributed computing. The process consumes tens of billions of hours per year on CPUs in leading-edge foundries, necessitating large data centers within semiconductor foundries.

    Traditional methods for resolving these complex calculations are time-consuming and resource-intensive, often taking several months or even years to complete. In contrast, the cuLitho platform utilizes NVIDIA's H100 Tensor Core GPU-based systems to accelerate computational lithography. By leveraging the power of accelerated computing, 350 cuLitho systems can now replace 40,000 CPU systems, significantly reducing production time while minimizing costs and space requirements.

    Moreover, NVIDIA has developed innovative algorithms that apply generative AI to enhance the value of the cuLitho platform. These advanced techniques enable the creation of near-perfect inverse masks or inverse solutions, accounting for diffraction of light involved in computational lithography. This results in a 2x speedup on top of the accelerated processes enabled through cuLitho.

    The application of generative AI to optical proximity correction has revolutionized the field, allowing for the more accurate simulation of physics and the realization of mathematical techniques that were once impractical. This enormous speedup accelerates the creation of every single mask in the fab, which speeds the total cycle time for developing a new technology node.

    Furthermore, the integration of generative AI into the cuLitho platform enables the exploration of previously inaccessible calculations. Leading-edge foundries will utilize this capability to ramp up inverse and curvilinear solutions, creating the next generation of powerful semiconductors.

    As Dr. C.C. Wei, CEO of TSMC, noted at the GTC conference earlier this year, "Our work with NVIDIA to integrate GPU-accelerated computing in the TSMC workflow has resulted in great leaps in performance, dramatic throughput improvement, shortened cycle time, and reduced power requirements." This collaboration has redefined the boundaries of what is possible in semiconductor manufacturing, paving the way for significant breakthroughs in computer chip design and development.

    The introduction of cuLitho represents a seismic shift in the industry's approach to computational lithography. By harnessing the power of accelerated computing and AI, TSMC and NVIDIA have created a platform that will revolutionize the way semiconductors are designed and manufactured. As the semiconductor landscape continues to evolve, this partnership will play a pivotal role in shaping the future of computer chips and enabling groundbreaking innovations.



    Related Information:

  • https://blogs.nvidia.com/blog/tsmc-culitho-computational-lithography/


  • Published: Wed Oct 16 01:45:39 2024 by llama3.2 3B Q4_K_M











    © Digital Event Horizon . All rights reserved.

    Privacy | Terms of Use | Contact Us