Profiling with gdb

Why not profiling your application by sampling backtraces using gdb? This intuitive idea is easily adapted to remote debugging as well. [2]

If the sampling rate is too slow, quickstack can go faster. It is also a good reference for implementing a faster version of a gdb script. [3]

Bonus: gdb convenience variables and especially this one:

(gdb) break funcA if $_caller_is("funcB")

Wouldn't it be nice to setup a little Raspberry with gdb compiled for your target in order to sample backtraces all the time? (especially for Windows users) [4]

Wait! gdb is super slow because of the hundreds of megabytes of symbols it is loading each time. Solution! -> DebugFission [5] ([6] [7] [8] [9])

Profiling with callgrind

Why not using a real profile?

Valgrind Callgrind [10]

The output can be read with qtcreator, but it does not work as well as kcachegrind. Good news, kcachegrind includes qcachegrind, which is a Qt-only application without dependencies to KDE. The official website looks old and out-dated [11], but if you look at the git repository [12] shows that the tool is still updated.

References