doc/guide/src/ProfilingTheStack.adoc

   1 ProfilingTheStack
   2 =================
   3
   4 For all forms of <:Profiling:>, you can gather counts for all
   5 functions on the stack, not just the currently executing function.  To
   6 do so, compile your program with `-profile-stack true`.  For example,
   7 suppose that `list-rev.sml` contains the following.
   8
   9 [source,sml]
  10 ----
  11 sys::[./bin/InclGitFile.py mlton master doc/examples/profiling/list-rev.sml]
  12 ----
  13
  14 Compile with stack profiling and then run the program.
  15 ----
  16 % mlton -profile alloc -profile-stack true list-rev.sml
  17 % ./list-rev
  18 ----
  19
  20 Display the profiling data.
  21 ----
  22 % mlprof -show-line true list-rev mlmon.out
  23 6,030,136 bytes allocated (108,336 bytes by GC)
  24        function          cur  stack  GC
  25 ----------------------- ----- ----- ----
  26 append  list-rev.sml: 1 97.6% 97.6% 1.4%
  27 <gc>                     1.8%  0.0% 1.8%
  28 <main>                   0.4% 98.2% 1.8%
  29 rev  list-rev.sml: 6     0.2% 97.6% 1.8%
  30 ----
  31
  32 In the above table, we see that `rev`, defined on line 6 of
  33 `list-rev.sml`, is only responsible for 0.2% of the allocation, but is
  34 on the stack while 97.6% of the allocation is done by the user program
  35 and while 1.8% of the allocation is done by the garbage collector.
  36
  37 The run-time performance impact of `-profile-stack true` can be
  38 noticeable since there is some extra bookkeeping at every nontail call
  39 and return.