Last year, NVIDIA introduced its cuLitho software library, which promises to speed up photomask development by up to 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.

The development of photomasks is a crucial step for every chip ever made, and NVIDIA's cuLitho platform, enhanced with new generative AI algorithms, significantly speeds up this process. NVIDIA says computational lithography consumes tens of billions of hours per year on CPUs. By leveraging GPU-accelerated computational lithography, cuLitho substantially improves over traditional CPU-based methods. For example, 350 NVIDIA H100 systems can now replace 40,000 CPU systems, resulting in faster production times, lower costs, and reduced space and power requirements.

NVIDIA claims its new generative AI algorithms provide an additional 2x speedup on the already accelerated processes enabled through cuLitho. This enhancement is particularly beneficial for the optical proximity correction (OPC) process, allowing the creation of near-perfect inverse masks to account for light diffraction.

TSMC says that integrating cuLitho into its workflow has resulted in a 45x speedup of curvilinear flows and an almost 60x improvement in Manhattan-style flows. Curvilinear flows involve mask shapes represented by curves, while Manhattan mask shapes are restricted to horizontal or vertical orientations.

Synopsys, a leading developer of electronic design automation (EDA), says that its Proteus mask synthesis software running on the NVIDIA cuLitho software library has accelerated computational workloads compared to current CPU-based methods. This acceleration is crucial for enabling angstrom-level scaling and reducing turnaround time in chip manufacturing.

The collaboration between NVIDIA, TSMC, and Synopsys represents a significant advancement in semiconductor manufacturing in general and cuLitho adoption in particular. By leveraging accelerated computing and generative AI, the partners are pushing semiconductor scaling possibilities and opening new innovation opportunities in chip designs.

Source: NVIDIA

Comments Locked


View All Comments

  • GeoffreyA - Tuesday, March 19, 2024 - link

    Possibly, their software has been of the older, tried-and-tested sort.
  • PeachNCream - Tuesday, March 19, 2024 - link

    Wow! NV is using computers to help with lithography?! Amazing! Thank goodness they finally figured that out and aren't making integrated circuit mockups on chains of perforated tractor-feed dot matrix printer paper taped together and spread out onto office floors. NV engineers can finally turn their desk fans on without risking displacement of critical product designs. Maybe next year they can get central air conditioning or something. The future is now!
  • kn00tcn - Tuesday, March 19, 2024 - link

    parallel gpus are wildly more efficient than parallel cpus in many tasks, that's what this is about not your delusional narrative

    you are sick, polluting the comments with insane aggressive tirades mocking or slurring users and products constantly, disgusting!
  • PeachNCream - Tuesday, March 19, 2024 - link

    Your joke detector's batteries may require replacement.
  • haplo602 - Tuesday, March 19, 2024 - link

    The problem here is the AI part. In a normal algorithmic solution, if there's an error, you can trace it back and fix it. With AI no such thing is possible easily as the link between algorithm and result is not straight forward.
  • ballsystemlord - Tuesday, March 19, 2024 - link

  • Dante Verizon - Tuesday, March 19, 2024 - link

    This is not an error-tolerant process, so it makes no sense to use GPUs.

Log in

Don't have an account? Sign up now