UPMEM is a fabless semiconductor company founded in 2015. The company is building a UPMEM Processing In-Memory (PIM) solution based on an innovative processor architecture to enable the computation of data intensive operations into hundreds of programmable co-processors sitting right next to the data. Big data applications leveraging UPMEM’s PIM solution are modified to dispatch and orchestrate the tasks executed by the co-processors from the main CPU, using high level programming languages.

The PIM chip, integrating UPMEM’s proprietary RISC processors (DRAM Processing Units, DPUs) and main memory (DRAM), is the building block of the first efficient, scalable and programmable acceleration solution for big data applications. Associated with its Software Development Kit, the UPMEM PIM solution can accelerate data-intensive applications in the datacenter servers 20 times, with close to zero additional energy premium. This huge leap opens new horizons for Big Data players, in terms of costs and new services.

UPMEM’s innovative technology solves the Memory Wall and the dominant energy cost of data movement between the processor and its main memory in application servers. Thousands of UPMEM in-memory co-processors (DRAM Processing Units, aka DPUs) orchestrated by the main processor, localize most of data processing in the memory chips, while proposing familiar programmability. Besides, the UPMEM solution comes without any disruption of existing server hardware, standardized protocols, programming & compiling schemes, removing any barrier for fast & massive adoption. For instance, the UPMEM solution provides a full DNA mapping and variance analysis in minutes instead of hours, making affordable real-time personalized genomics a reality.