C/C++ Profilers Comparison
With our C/C++ profiler you can profile directly on your embedded platform, like we did. You'll get a full call graph when profiling programs with dynamic libraries and multiple threads, all with CPU cycle precision.
| EProfiler | GProf | OProfile | Valgrind | Very Sleepy | gperftools | |
|---|---|---|---|---|---|---|
| Native on both Linux and Windows |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
| Portable to embedded systems |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
| Complete call graph |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
| Multithreaded programs |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
| Shared (Dynamic) library support |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
| Profiling only selected code |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
| No need for debug info in binaries |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
| Time resolution in CPU cycles |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
| No CPU simulation |
![]() |
![]() |
|
|
![]() |
![]() |
| Time estimation without profiler overhead |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |

