I use it - perf is wonderful. Back in the dark days of grad school, getting performance counter data for your program required hacking your code to call into PAPI (http://icl.cs.utk.edu/papi/overview/). Being able to tell a program, "Hey, run this for me, and tell me how many branch mispredictions there are. And where they happen" is like magic.

If you do any performance analysis of software, learn perf.

Official manual: https://perf.wiki.kernel.org/index.php/Main_Page

Brendan Gregg's perf page: http://www.brendangregg.com/perf.html

I beg to disagree. For instance, it's been running for years sampling Google's fleet [1] (the paper mentions oprofile, but that has been replaced quite a while ago).

[1] http://courses.cs.washington.edu/courses/cse590g/11sp/36575....

I used it to debug some performance issues in a graphics engine we use at my company. It helped me understand what the low-hanging fruit was.