GPU Next-Gen Architecture Modeling Engineer

Advanced Micro Devices, Inc.•Santa Clara, CA

62d•Hybrid

About The Position

At AMD, our mission is to build great products that accelerate next-generation computing experiences-from AI and data centers, to PCs, gaming and embedded systems. Grounded in a culture of innovation and collaboration, we believe real progress comes from bold ideas, human ingenuity and a shared passion to create something extraordinary. When you join AMD, you'll discover the real differentiator is our culture. We push the limits of innovation to solve the world's most important challenges-striving for execution excellence, while being direct, humble, collaborative, and inclusive of diverse perspectives. Join us as we shape the future of AI and beyond. Together, we advance your career. THE ROLE: The Architectural Modeling (AM) team is looking for a GPU Next-Gen Architecture Modeling Engineer who would be responsible for the only software simulator for all graphics products at AMD (Instinct and Graphics products). The tool is utilized to predict the performance of the next-gen product across different GPU families at AMD. The role of the Architecture Modeling Engineer/Scientist can be of the following categories: Problem identification: Data from the AM tool is used by GPU architects used to identify points that can be enhanced in our next-gen products. AM team develops a variety of tools and datapoints makes finding performance problems easier. Feature Development: The AM team works with architects on ideas, called features, to improve the graphics architecture. The AM Engineer is responsible for the gathering of requirements, design, and implementation of new features within the model. The AM team has the core responsibility of proving or disproving new features inclusion in next-gen products. Model Development: As an AM engineer you will develop the knowledge of the entire graphics and compute pipeline for all our off-the-shelve products, and details of the upcoming changes and their impact. Infrastructure / Model Enhancement: is a C++ simulator coupled with python postscripts for data visualization and analysis. As an AM engineer, you would get the chance to bring your expertise in these languages to improve the overall usability of the AM tools in terms of wall-clock runtime, memory consumption, code quality and complexity. SW Shift Left: Help drive our SW shift-left initiative, where we have a faster functional model and driver stack to facilitate our architecture analysis and IP verification needs sooner in the development cycle, prior to detailed architecture modeling of target software. THE PERSON: As a GPU modeling engineer, you will work in a hardware/software co-design environment to craft industry leading simulation platforms and tools, have a strong voice in next generation of AMD products, and get to learn and contribute to the entire graphics pipeline and software/hardware stack. The ideal candidate is a highly skilled software designer and engineer, who is proficient in C++, and familiar with hardware architecture. Knowledge of the GPU hardware and the efficient use of discrete GPUs, APUs, or SoC devices is a must. The candidate will work in a big, multi-geo team environment on a variety of opportunities, so teamwork and drive are essential.

Requirements

Expertise in C++ and python Programming
Experience of Hardware architecture
Experience of GPU architecture, drivers or compilers
Experience of modeling and simulation tools

Nice To Haves

Experience of C++ build tools such as CMake, Meson, Bazel is a plus
Experience of D3D, Vulcan and OpenGL graphics applications is a plus
Experience of OpenCL/CUDA/ROCm development is a plus

Responsibilities

Develop tools and algorithms to identify and eliminate performance limiters.
Design, implement, assess, document and test new features for next-gen products.
Be a developer, leader, and owner of the Architectural Modeling Tool.
Model and analyze hardware and software features to improve products performance, abilities, and cost.