Datacenters are evolving at an outstanding rate of innovation. NVIDIA is rapidly expanding its offerings in the Datacenter SoC Product space with a varying array of workloads. We are looking for outstanding individuals who are driven to lead workload analysis and optimizations and drive the future of SoC data center products from NVIDIA.
What you’ll be doing:
In this role, your primary role will be to analyze workloads that are run by major data center CSPs. Your work will be integral to defining the performance specifications of NVIDIA’s next generation of datacenter server SoC products.
Profile and debug SoC application performance, tracing through complex application, driver, and OS software stacks.
Communicate your learnings to architecture and engineering teams to bring performance differentiation.
Use and develop pre- and post-Si architectural models, analytical models, and tools to drive hardware architecture decisions to examine performance, perf/watt, perf/area, TCO, ROI and analyze tradeoffs based on customer and segment workloads and use cases.
Where necessary, build micro benchmarks for critical data center use cases to help evaluate against key performance metrics.
What we’ll need to see:
A MS or BS in Electrical Engineering, Computer Science, Computer Engineering (or equivalent experience).
10+ years of experience in SoC, CPU, and system performance architecture.
Extensive experience in software profiling and optimization.
Experience with tools for performance analysis.
Knowledge of Linux kernel internals, Filesystem, and memory architecture,
Solid in coding using any of languages like python, C/C++. Experience with version control tools (e.g. git).
Ways to stand out from the crowd:
Good understanding of ARM and x86 platforms at architecture level
Experience in Cloud native applications optimization and debug
Solid knowledge of Architecture for Core, Mesh and IOs, such as PCIe is a huge plus
Additional software skills such as SystemC modeling, OS/driver/real time software development