by Brian Kelleher

NVIDIA’s labs are busy with engineers preparing to bring to market the next generation of GPUs based on 28nm process technology, the result of a close, 13-year partnership with Taiwan Semiconductor Manufacturing Co. (TSMC). The new family of 28nm GPUs are designed with a focus on performance per watt.

Engineers in our Advanced Technology Group (ATG) have been working for the past two years with TSMC to ensure that TSMC is ready to ramp our products aggressively.

We had established the ATG in the wake of the challenges in driving to production with the 40nm process, TSMC’s previous chip-making technology. We wanted to ensure this new 28nm process was ready by the time the NVIDIA product team started their work.

A TSMC tech in the 12-inch wafer fab
inspects micro-architecture.

With each new process technology, physical structures on a processor become smaller and smaller, requiring new and advanced techniques to create the structures while maintaining low power consumption. For reference, a human hair is about 100,000 nanometers, so imagine the complexity of creating structures that are only 28nm in size.

We send our ATG engineers into the battle early to eliminate design and manufacturing issues and ensure our products have high quality and high yield. As a direct result of the collaboration between the ATG and TSMC, NVIDIA’s 28nm processors will soon deliver an amazing gaming experience through our GeForce GTX products, enable a new level of creativity through our Quadro products and help solve some of the world’s most difficult computing problems.

The next-generation of NVIDIA processors will be capable of performing more calculations in a shorter period of time – all while drawing less power. That’s good for high-performance graphics, super phones, and tablets. And it’s good for supercomputers.