Profiling Hybrid CUDA/Graphics Applications using Nsight Graphics

In this session, you'll learn how to use NVIDIA Nsight Graphics to profile applications that combine the use of CUDA and graphics APIs , including ray tracing and AI-enhanced workloads, focusing on optimizing performance across compute and graphics tasks. We’ll explore the connection between high-level CUDA C, shader code (HLSL, GLSL), and GPU hardware concepts, while inspecting performance traces of CUDA, DirectX, Vulkan, and NVIDIA SDK calls. By the end of the session, you’ll be better equipped to identify and resolve performance bottlenecks at both the source code and GPU assembly (SASS) levels, ensuring your applications run with maximum efficiency.