nvprof --metrics branch_efficiency..... Why no metrics ?

I am trying yo run
nvprof --metrics branch_efficiency warp_divergence.out
but I got this output

C:\Users\sulo4\source\repos\warp_divergence\warp_divergence>nvprof --metrics branch_efficiency warp_divergence.out
======== Warning: Skipping profiling on device 0 since profiling is not supported on devices with compute capability 7.5 or higher. Profiling features on these devices are supported in the next generation GPU profiling tool NVIDIA Nsight Compute. Refer https://developer.nvidia.com/nsight-compute for more details.

-----------------------WARP DIVERGENCE EXAMPLE------------------------

==13820== NVPROF is profiling process 13820, command: warp_divergence.out
==13820== Profiling application: warp_divergence.out
==13820== Profiling result:
No events/metrics were profiled.

I should have metrics?

As the nvprof warning message says - you need to use Nsight Compute for metric collection on GPU devices with compute capability 7.5 or higher.

Note that the metric names in Nsight Compute are different than nvprof.

Please refer to the metric mapping table in the Nsight Compute document:
https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html#nvprof-metric-comparison

You can also refer to the blog post:
https://devblogs.nvidia.com/using-nsight-compute-to-inspect-your-kernels/

To be honestly I didn’t understand your advice.
But I am using RTX2060 and it has 7.5 compute capability

nvprof does not support metric collection on RTX2060 (compute capability 7.5).
You need to use Nsight Compute.