Skip to main content

Posts

Showing posts with the label Dual-IMC

Energy efficiency in AI data operations with dual-IMC

Investigating Solution to the Von Neumann Bottleneck in AI Models Introduction to the Research AI models like ChatGPT are driven by complex algorithms and an insatiable need for data, which they inter pret through machine learning. But what are the boundaries of their data- processing ca pacity? Led by Professor Sun Zhong, a team from Peking University is investigating solutions to the Von Neumann bottleneck, a key barrier to data- processing  performance. Dual-IMC Scheme for Enhanced Machine Learning In their Se ptember 12, 2024  publication in Device , the research team introduced a dual-IMC (in-memory com puting ) scheme that enhances machine learning s peed while significantly boosting the energy efficiency of conventional data o perations. Matrix-Vector Multi plication in Neural Networks Software engineers and com puter scientists utilize matrix-vector multi plication (MVM) o perations when designing algorithms to  power neural networks, a com putational architectu...