Revolutionary SiPh Chip Accelerates AI Processing Using Light Waves
Engineers at the University of Pennsylvania have developed a silicon-photonic chip that uses light instead of electricity for AI computations, promising faster speeds and reduced energy consumption.
- The silicon-photonic (SiPh) chip performs complex mathematical operations essential for AI training at the speed of light, enhancing processing speeds and cutting energy use.
- Designed with localized variations in silicon thickness, the chip controls light propagation to execute calculations, marking a shift from conventional electrical chips.
- The chip's compatibility with existing GPUs and potential for commercial use could revolutionize computing, offering unprecedented computational performance.
- Inherent privacy features of the SiPh chip eliminate the need to store sensitive data in working memory, making future systems virtually immune to hacking.
- Supported by grants from the U.S. Air Force Office of Scientific Research and the U.S. Office of Naval Research, the project highlights the impact of interdisciplinary research on technological advancement.
Open Access Government 