Logo
Samsung Research America

SOC Performance Engineer

Samsung Research America, Mountain View, CA, United States


Lab Overview:

The Samsung SOC Lab vision provides innovative SoC architecture, bus / memory subsystem, multimedia subsystems and key IP blocks for future Samsung Galaxy products (Smartphones, tablets and future devices). We are defining the high performance SoC architecture development for various Galaxy device lineups. This lab collaborates with Samsung's strategic SoC partners, Samsung MX headquarter team, and key R&D teams around the globe to innovate and reinvent technology that will positively impact millions of people around the world via the Galaxy flagship products.

Position Summary:

We are looking for a Performance Engineer to build benchmarks for characterizing SoC’s for compute, memory and power performance. The person will assist SoC architects to analyze the characterization data for new SoC architectures and features for next generation mobile products. Solid computer architecture foundation, strong coding skills, aptitude for embedded systems based upon Linux/Android and performance benchmarking experience is required for success in this role.

Position Responsibilities:

  • Understand high-level system architecture of SoC based embedded systems including HW and SW.
  • Board bring-up with a clear understanding of initialization sequence and interdependencies of various sub-systems.
  • Research performance benchmarks and analyze workloads for compute, memory and power characterization at Soc level as well as at IP level (CPU, GPU, NPU, Media, Connectivity etc.).
  • Execute workloads and develop tools to analyze and present findings from the characterization data.
  • Leverage Industry standard tools and/or create in house tools to speed up performance evaluation and analysis.
  • Work across the organization to analyze workload and decompose, map workload components to IPs to optimize system performance, energy.
  • Working with teams [IP design, modeling, system architecture] to identify, analyze and optimize critical complex use cases for 2D and 3D applications involving vision sensors, displays and AI based vision data processing.

Required Skills:

  • BSc/Masters with 3+ years of experience or PhD with 1+ years of relevant experience or equivalent combination of education, training, and experience.
  • Strong development skills using C++, python and similar programming languages. Some understanding of hardware description languages (e.g. SystemC) is a plus.
  • Clear understanding of operating systems principles. Understanding of Android architecture is a big plus.
  • Experience with tools, debuggers, logs and data analysis tools Android and/or Linux based products.
  • Experience with device driver development. Having experience of Display or ISP or GPU driver is a huge plus.
  • Experience with industry standard performance and power characterization tools, workloads and benchmarks with a focus on executing them and correlating scores/output with architectural constraints.
  • Understanding of computer architecture. ARM CPU architecture is desired. Understanding how CPU, GPU and NPU architecture compare, and contrast is nice to have.
  • Ability to translate high level architectural issues into relevant characterizable workloads.
#J-18808-Ljbffr