Accelerating AI Workloads
NPL (Neural Processor Lab) is responsible to come up with world-best innovative NPU (Neural Processing Unit) architectures and algorithms to be implemented in Samsung Exynos, or other Samsung products. For this, NPL works as a part of global Samsung NPU R&D ecosystem, composed of SAIT, S.LSI, and several other overseas Samsung R&D centers.
NPL is the fore-runner of these Samsung organizations to guide the future directions of Samsung NPU, by identifying major AI workloads that Samsung’s NPU needs to accelerate in 3-5 years. For this, NPL collaborates with world-renowned academia research groups in AI and NPU.
NPL is composed of motivated deep learning software researchers and hardware architecture engineers. We are a dozen of fast moving explorers in this crazy-paced AI field, and idea generators to innovate Samsung NPU’s performance. As such, we value speed, insight, and team-work. As a bridge between academia and industry, we also publish and participate in AI conferences like CVPR, ECCV, etc. Located in the heart of Silicon Valley AI ecosystem, we also sponsor and participate various local conferences like tinyML, MLSys, etc.
NPL originated from NPU project which started in 2015. The NPU project has been providing NPU architectures that were adopted in Samsung Exynos chips which powered Samsung AI solutions to Samsung Galaxy phones since 2019. Samsung NPU is a wholly home-grown IP that were made possible by the early actions taken before competitors came up with their own products. These successes lead into the foundation of NPL in 2019.