Nvidia has announced the development of inverse lithography, a tool that speeds up a computation-limited step in the chipmaking process, allowing chipmakers to print nanometer-scale features using light with a longer wavelength than the size of those features.
The tool, dubbed CuLitho is a set of algorithms designed for use with GPUs that turns two weeks of work into an overnight job. It was created in collaboration with Taiwan Semiconductor Manufacturing Co. and Synopsys. CuLitho, according to Nvidia, provides up to 40 times faster computations and reduces carbon footprints.
The technology “will allow fabs to increase throughput, reduce their carbon footprint, and set the foundations for 2-nanometer [manufacturing processes] and beyond,” said Nvidia CEO Jensen Huang. In fact, cuLitho can produce three to five times as many photomasks per day, drawing only 5 MW instead of 35 MW.
Inverse lithography is a scheme that calculates what pattern on the photomask would produce the desired pattern on the wafer, which is much more efficient than the current photomask projection method. However, the long computation time for lithography has slowed chip technology development and improvement.
According to Vivek K. Singh, Vice President in Nvidia’s advanced technology group working on silicon manufacturing, it has been a long pole in chip development. “If inverse lithography technology were sped up 40 times, would more companies use it on more layers? Surely,” Singh added.
Nvidia’s advancement may also result in the use of fewer photomasks, as what must now be done with a double pattern may now be done with only one. Furthermore, cuLitho creates otherwise difficult-to-calculate curvy polygons on the mask, resulting in a greater depth of focus for the pattern cast onto the wafer.
The sources for this piece include an article in Spectrumieee.
