UPMEM

UPMEM Processing In-Memory chips solve the dominant cost of data movement and the Memory wall, at marginal cost and energy consumption for Large Analytics and Generative AI applications. UPMEM 1st generation solution readily integrates in app servers to offer the 1st efficient scalable programmable acceleration solution for data-intensive apps such as genomics, AI, anaytics, search... UPMEM first chip and programmable solution is avalaible and shipping to top tier customers - application owners and cloud players - in US, Asia and Europe, or accessible from our PIM cloud services in France for customers or research partners. UPMEM successfully released the first PIM on DRAM chip and server based solution with amazing programmability and scalability outlook. UPMEM 2nd generation PIM chip delivers 20x cheaper LLM inference for Generative AI in the cloud and computes with 10x less energy on mobile, opening the path to massive on device GenAI usage. UPMEM is maintaining a sustained...