SESSION

Beyond Open Source Hardware, Open Source Processor REVIEW

PREVIEW
Can open source shaking the software world make a big difference in the semiconductor field? As renowned semiconductor companies such as Google, Samsung, IBM, and NVIDA showed interest in the recent open source chip design project RISC-V, things that could not have been imagined after a few years of reorganization of the processor field led by Intel and ARM are a reality. The emerging open source processor has evolved from a university education tool to the level of technology used in strategies for big players in the semiconductor field such as Google and Samsung Electronics. In this interest, it is being evaluated that it can be a threat to the market of Intel and ARM. In this session, we will explain these open processors and briefly introduce representative models of RISC-V and NVDLA. RISC-V has been applied in various fields such as SSD controller, 5G modem, IoT, etc. to show the future and potential of RISC-V, and recently, research to apply it to the automotive field is underway. NVDLA started as an open source project in various Neural Processing Units (NPUs), and the standardization of NPU was devised. In addition, this session will introduce the open source processor, the general introduction of open source hardware, and the future of the open source processor. He ask for the participation of interested people.
  • Sang-Soo Park
  • Sang-Soo Park / Hanyang University Ph. D Candidate Student
  • Sang-Soo Park is doing a doctoral course at Hanyang University. In graduate school, his research area is an accelerating neural networks with various methods. He is researching a hardware accelerator architecture dedicated to neural networks that accelerates neural network inference with low power and high efficiency, and researches on how to effectively accelerate neural network inference using existing processors (CPU/GPU/NPU). Along with research, He is working on various projects. First, Home appliances such as robot cleaners operate in very different environments. Various environments have a great influence on the motion as the robot cleaner is different from the correct humid environment. To this end, He designed a transfer learning framework (LG Electronics) that can respond to various operating environments in home appliances, a voice recognition hardware accelerator, and an NDP (Near data processing) project (Samsung Electronics) that computes and accelerates a recommendation system near memory. In addition, He is leading to the Neural Acceleration study, which researches on methods of neural network acceleration.

Feel free to leave comments/curiosity/opinions about the session.

* Comments (including swearwords and slander) that do not match the purpose of the bulletin are deleted by the administrator without notice.

* Comments can only be uploaded while the sessions are ON–AIR and updated every minutes.

0/100 characters WRITE