% REPORTTIME=0 % sleep 1 sleep 1 0.00s user 0.00s system 0% cpu 1.008 total % REPORTTIME=1 % sleep 1 % REPORTTIME=0.99 % sleep 1 sleep 1 0.00s user 0.00s system 0% cpu 1.008 total % REPORTTIME=1.001 % sleep 1 % sleep 2 % REPORTTIME=1.1 % sleep 2 % REPORTTIME=0.99 % sleep 2 sleep 2 0.00s user 0.00s system 0% cpu 2.006 total % What's going on here? It seems that when $REPORTTIME >= 1, no time output is ever produced. The exact values here are unique to sleep, but similar behavior can be observed for other commands, too, albeit at different thresholds.
On 6/30/21, Marlon Richert <marlon.richert@gmail.com> wrote:
> % REPORTTIME=0
> % sleep 1
> sleep 1 0.00s user 0.00s system 0% cpu 1.008 total
> % REPORTTIME=1
> % sleep 1
> % REPORTTIME=0.99
> % sleep 1
> sleep 1 0.00s user 0.00s system 0% cpu 1.008 total
> % REPORTTIME=1.001
> % sleep 1
> % sleep 2
> % REPORTTIME=1.1
> % sleep 2
> % REPORTTIME=0.99
> % sleep 2
> sleep 2 0.00s user 0.00s system 0% cpu 2.006 total
> %
>
> What's going on here? It seems that when $REPORTTIME >= 1, no time
> output is ever produced.
>
> The exact values here are unique to sleep, but similar behavior can be
> observed for other commands, too, albeit at different thresholds.
REPORTTIME
If nonnegative, commands whose combined user and system execution
times (measured in seconds) are greater than this value have timing
statistics printed for them.
Note the difference between elapsed time (total) and cpu time (user +
system) in your output.
--
Mikael Magnusson
On Thu, Jul 1, 2021 at 2:13 AM Mikael Magnusson <mikachu@gmail.com> wrote:
> > What's going on here? It seems that when $REPORTTIME >= 1, no time
> > output is ever produced.
> >
> > The exact values here are unique to sleep, but similar behavior can be
> > observed for other commands, too, albeit at different thresholds.
>
> REPORTTIME
> If nonnegative, commands whose combined user and system execution
> times (measured in seconds) are greater than this value have timing
> statistics printed for them.
>
> Note the difference between elapsed time (total) and cpu time (user +
> system) in your output.
Ah, thanks. And I guess REPORTTIME is rounded down to an integer
value. That would explain it.