Enabling Excellence: Optimizing GPU Applications, AI Workloads, and Supporting the HPC Community
Quentarius Moore, Advanced Micro Devices (AMD)
As a recipient of the prestigious Howes Scholar Award from the DOE CSGF, I am honored to share my journey since being a fellow. This award, established in memory of Frederick Anthony Howes, celebrates technical excellence, leadership, and integrity—values I strive to embody in my work.
In this talk, I will provide a high-level overview of my current role and responsibilities during the past 10 months at Advanced Micro Devices (AMD), focusing on the optimization and porting of high-performance applications to enhance their efficiency on AMD graphics processing units (GPU) clusters. I have worked on a range of scientific and AI applications, involving detailed performance profiling, scalability improvements, and innovative solutions to complex computational challenges.
Additionally, I will discuss my involvement in responding to Requests for Proposals (RFPs) aimed at profiling and optimizing applications for large-scale deployment. My efforts have been geared towards enabling high-performance applications to achieve their full potential on AMD GPU-based clusters, thereby benefiting both the developer and end-user communities.
Throughout the talk, I will highlight the significance of close support and comprehensive documentation as vital components of successful application optimization. By documenting our work meticulously, we create valuable assets for future training and development, ensuring that our methodologies benefit a broader audience.
I will conclude with personal experiences and lessons learned since transitioning from graduate studies to professional life, offering practical advice for new and current fellows. My goal is to inspire and encourage the next generation of computational scientists to seek the excellence already within themselves and make meaningful contributions to their personal lives and fields.