Samsung patent boosts AI server memory efficiency without adding GPUs
Samsung Electronics has disclosed a patent application targeting GPU memory shortages when multiple users simultaneously submit requests to AI servers. By calculating data-transfer costs between GPUs and CPUs and minimizing unnecessary transfers, the technology is designed to enable AI servers to process more user requests quickly within limited memory capacity, without adding GPUs.
Read original article ↗
Related Articles
Samsung chairman visits UAE, encourages employees
Samsung Electronics Chairman Lee Jae-yong visited the Al Dhafra gas-fired power plant construction site in Abu Dhabi, th
SDC supplies over 70% of major laptop brands' OLED panels
Samsung Display (SDC) currently supplies more than 70% of the OLED panels used by major global laptop brands, intensifyi
Samsung's Nvidia HBM bid raises four strategic concerns
Samsung Electronics is reportedly turning to TSMC's 3nm process to produce base dies for Nvidia's custom high-bandwidth
Russia considers reinstating parallel imports of Samsung, SK Hynix products amid memory shortages
Russia is reportedly considering reinstating parallel imports of products from Samsung Electronics and SK Hynix to ease