Missing split option?

[-hh]

@Richard.
I like your benchmarking "soft spot" and I've learned
both from your stacks and considerations, appreciating
that until now.
But you really could add a smiley if you *interpret* a
1/10000 second difference of timing ;-)

@Phil.
The additional digits of the long seconds are probably
closer to random than to exact measurement. Anyway, as
the long seconds are 'produced' increasingly, they are
good for ordering events in a queue. But they are NOT
'exact' for measuring the duration of an event, TMHO.

Hermann

p.s. I think we should prefer the median for averaging
timing, or, even better, if the process measured is
important, show the whole distribution of timing values
(in categories/histogram).




--
View this message in context: http://runtime-revolution.278305.n4.nabble.com/Missing-split-option-tp4704799p4704829.html
Sent from the Revolution - User mailing list archive at Nabble.com.