[fuse-devel] time benchmark question

[Tobias P. <tob...@gm...>]

hi,

i have a short question on measuring execution time of a `find`
invocation on a fuse file system mount. 

When I call time with `--format "%e %S %U"` I get 
22.20 0.20 1.0 for e, S, U respectively where :

e is : Elapsed real (wall clock) time used by the process, in seconds.
S is : Total number of CPU-seconds used by the system on behalf of the
       process in kernel mode
U is : Total number of CPU-seconds that the process used directly 
       (in user mode) in seconds.

Am I right to assume that the big gap between 'e' and (S+U) would
have to be accounted to CPU waiting for IO ? I didn't quite expect
such an outcome..

Sorry if this is not perfectly fuse ml related and thanks for
any input on this.

regards,
Tobi