Profiling CUDA Applications with nvprof, nsight systems and nsight compute.

Demo on howto use nvprof, NVIDIA Nsight Systems and Nsight Compute to profile and analyse CUDA code. Development and compiling (nvcc compiler) are used on Google Colab.

This video is part of lecturer of "Hardware Accelerated Computing" by Vincent Claes

[fpgabe]

When you get the error: /bin/bash: line1: nsys: command not found you have to run the following code in a Jupyter cell:

!apt update
!apt install
!apt --fix-broken install