0

The java profiling tool async-profiler could generate flamegraph even if some method in the call stack is jit optimized.

How is that implemented? If one method is inlined or jit optimized, how could we fetch its origin info or call graph?

choxsword
  • 3,187
  • 18
  • 44
  • From the readme "*The general approach is receiving call stacks generated by perf_events and matching them up with call stacks generated by AsyncGetCallTrace, in order to produce an accurate profile of both Java and native code*". If you want more detail than that, you're basically hoping @apangin shows up. – Michael Mar 31 '23 at 11:58
  • @Michael Yes, I'd appreciate it if more details are provided. – choxsword Mar 31 '23 at 12:25
  • 2
    Why do you think that this is a special challenge? It’s not as if getting stack traces didn’t work with JIT compiled code. Hence, the other profilers also show JIT compiled methods in their output, just biased to safepoints. Considering that hot spots are the first to become JIT compiled and the focus of profiling tools are the hot spots, these tools would be even less useful if that didn’t work. – Holger Mar 31 '23 at 14:32

0 Answers0