Add -falseprecision flag for ludicrously precise stats

The normal round-to-three-significant-figures behaviour, while morally
correct from a benchmarking perspective, makes it impossible to measure
variances between runs of less than 1%.  Occasionally you really do need
to measure that finely - for example, if tweaking the server's main
dispatch loop.

Signed-off-by: Adam Jackson <ajax@redhat.com>

0 files changed, 0 insertions, 0 deletions