For NVIDIA cards, you can use NSight. There's also RenderDoc that works on a large number of GPUs.

nsys and nvtx are awesome.

many don't know but you can use them without GPUs :)

	▲	KeplerBoy 4 hours ago \| parent [-]
		That's true. It's pretty good for all kinds profiling. I especially like the python GIL profiler of nsys.