Profiling Deep Learning Applications with NVIDIA Tools Extension (NVTX)

Profiling Deep Learning Applications with NVIDIA Tools Extension (NVTX)

Discover how to leverage the NVIDIA Tools Extension Library (NVTX) to enhance deep learning application profiling in this 24-minute webinar from Nvidia. Explore techniques for annotating code to enrich profiler timelines, gaining insights into network performance optimization. Learn about the Nsight product family, NVTX syntax, and best practices for memory pinning and asynchronous data copy. Dive into practical examples of backward propagation with and without Automatic Mixed Precision (AMP), and understand how to identify the right performance knobs to tune your deep learning applications effectively.