CPU Core Performance Architect

NVIDIA has been transforming computer graphics, PC gaming, and accelerated computing for more than 25 years. It’s a unique legacy of innovation that’s motivated by great technology—and outstanding people. We are leading the way in groundbreaking developments in Artificial Intelligence, High-Performance Computing and Visualization. The GPU, our invention, serves as the visual cortex of modern computers and is at the heart of our products and services. As an NVIDIAN, you’ll be immersed in a diverse, supportive environment where everyone is inspired to do their best work. Come join the team and see how you can make a lasting impact on the world!

We are now looking for a CPU Architect in the CPU team! NVIDIA needs a passionate engineer who is comfortable in both the hardware and software worlds to help us push the boundaries of CPU and SoC simulation. Today’s complex CPU designs rely on detailed functional models to aid in verification, and to support the running of complex workloads for architectural exploration. Such workloads often span CPU and GPU, such as gaming, Deep Learning (DL), autonomous vehicles, and High-Performance Computing (HPC). These models and workloads enable important decisions to be made about future NVIDIA CPU architectures and allow us to gain confidence in our HW implementations.

What you'll be doing:

Develop cycle-accurate architectural performance and power models for the CPU core of a complex high performance system-on-a-chip.

Use models to analyze architectural tradeoffs related to CPU Performance and Power, identifying performance bottlenecks on deep learning (DL) workloads, high-performance computing (HPC) workloads, as well as industry standard benchmarks like SPECInt and SPECFp.

Propose micro-architectural alternatives that fit within area, timing and power constraints for the design.

Work with the performance verification team to validate the model, and to root cause disconnects between model projections and design (RTL).

What we need to see:

BS/MS in EE, CE, or CS

4+ years of experience with a focus on CPU microarchitecture, or equivalent experience

Excellent C/C++ programming and debugging skills

Understanding of computer architecture

Ways to stand out from the crowd:

Knowledge in CPU memory & caching microarchitecture

Knowledge in OOO microarchitecture for hyperscale CPU’s

A history of generating innovative microarchitectural ideas to improve performance, from an idea to a working feature on silicon

Knowledge of the ARM ISA & microarchitecture

Understanding of SoC and GPU architecture

NVIDIA invented the GPU in 1999 and sparked the growth of PC gaming, redefined computer graphics, and revolutionized parallel computing. Today, NVIDIA develops chips and software to tackle challenges no one else can solve. Our work is transforming the world’s largest industries—from gaming to robotics, self-driving cars to life-saving healthcare, climate change to virtual worlds where we can all connect and create. Our team plays a key role in setting power and performance goals for our in-house developed CPUs, as well as ensuring that the resulting design meets these goals. We collaborate with other groups within and outside of the greater CPU team to help arriving at the most efficient design. As tools to achieve these goals, we maintain a CPU lab for silicon experiments and a detailed performance/power CPU simulator for simulation studies. These are driven by workloads that we identify and bring up both on silicon and in full-system simulation.