SK Hynix Unveils Memristor-Based AI Chip With 10x Better Energy Efficiency Than Nvidia A100

SK Hynix-15.36%
SKHY-0.98%
SKHYV-0.98%
NVDA4.06%
According to a joint research effort by SK Hynix, TetraMem, and the University of Southern California, the companies unveiled a memristor-based in-memory computing (IMC) SoC chip recently, targeting neural network inference for edge AI devices. The 65-nanometer chip achieved 21.3 TOPS/W (trillion operations per second per watt) at 100MHz and 11.9 TOPS/W at 400MHz, roughly 10 times more energy-efficient than Nvidia A100 in INT8 mode. However, peak absolute compute capacity reached only 2.54 TOPS, limiting immediate applications in mainstream edge AI. Researchers noted the achievement represents a proof-of-concept with potential value for power-sensitive wearables and IoT devices by bypassing data movement bottlenecks.
Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.
Comment
0/400
No comments