Talk · Part of NVIDIA GTC EMEA
Optimizing for WARP Speed: Fine-Grained Kernel Optimization With Nsight Tools and NVTXD
Key note by Eden Biran, Head of GPU Optimization at Decart.
About
Understand how use NVIDIA Nsight™ Perf to profile production workloads live without interruption.
Learn how developers can fine-tune workloads starting at the cluster scale using Nsight Perf and Nsight Systems and then drill down to fine-grained kernel optimization using Device side NVTX(NVTXD) in Nsight Compute.
Explore how Decart uses all of the above to optimize performance of their models.