Updated about 4 years ago by Knödlseder Jürgen

How to use valgrind?¶

Valgrid is a very useful tool for finding bottle necks in the execution of code and to identify where actually the type is spent.

Running an executable with valgrind¶

valgrind --tool=callgrind --dump-instr=yes --collect-jumps=yes gsrvy (args)

Running an executable on a specific code zone¶

Mark the code zone using some macros

#include <valgrind/callgrind.h>

int main()
{
  foo1();
  CALLGRIND_START_INSTRUMENTATION;
  CALLGRIND_TOGGLE_COLLECT;
  bar1();
  CALLGRIND_TOGGLE_COLLECT;
  CALLGRIND_STOP_INSTRUMENTATION;
  foo2();
  foo3();
}

and run valgrind as follows

valgrind --tool=callgrind --dump-instr=yes --collect-jumps=yes --collect-atstart=no --instr-atstart=no gsrvy (args)

Advanced usage¶

If the callgrind.out.* file is empty call

callgrind_control -d [hint [PID/Name]]

where hint is an arbitrary string you can optionally specify to later be able to distinguish profile dumps.

Check whether instrumentation is turned on using

callgrind_control -b

You can also decide to switch instrumentation on using

callgrind_control -i on

(and switching off again by specifying “off” instead of “on”).

Result visualisation¶

Use the kcachegrind GUI to visualise the results.