bin@rhesus.primate.wisc.edu (Brain in Neutral) (04/09/88)
The csh "time" command produces lines of the form:
0.1u 0.4s 0:01 41% 18+17k 0+11io 21pf+8w
The csh manual explains the first 4 fields. What do the rest
mean?
---
Paul DuBois UUCP: {allegra,ihnp4,uunet}!uwvax!rhesus!dubois |
ARPA: dubois@rhesus.primate.wisc.edu --+--
|
alt.sex: just say rmgroup. |mouse@mcgill-vision.UUCP (der Mouse) (04/16/88)
In article <309@rhesus.primate.wisc.edu>, bin@rhesus.primate.wisc.edu (Brain in Neutral) writes: > The csh "time" command produces lines of the form: > 0.1u 0.4s 0:01 41% 18+17k 0+11io 21pf+8w > The csh manual explains the first 4 fields. What do the rest mean? From 4.2BSD csh source: %Uu %Ss %E %P%% %X+%Dk %I+%Oio %Fpf+%Ww %U user time, seconds (ru_utime field of rusage struct) %S system time, seconds (ru_stime field of ditto) %E elapsed time, min:sec or hr:min:sec %P percentage of #:## accounted for by #.#u + #.#s (ie, cpu utilization percentage) %X some sort of average (over time) of the text segment size %D ditto of the data+stack sizes %I "block input operations" (ru_inblock, see getrusage(2)) %O "block output operations" (ru_oublock) %F page faults (from disk) %W swaps (entire process swapped) %I and %O mean, according to the 4.3 man page for getrusage(), the following: ru_inblock the number of times the file system had to perform input. ru_outblock the number of times the file system had to perform ^ [sic] input. Under NOTES, a warning appears: The number ru_inblock and ru_outblock[sic] account only for real I/O; data supplied by the caching mechanism is charged only to the first process to read or write the data. (grammar mistakes theirs). And yes, <sys/resource.h> calls it ru_oublock even though half the man page calls it ru_outblock. der Mouse uucp: mouse@mcgill-vision.uucp arpa: mouse@larry.mcrcim.mcgill.edu
jerryp@cmx.npac.syr.edu (Jerry Peek) (04/18/88)
In article <309@rhesus.primate.wisc.edu> bin@rhesus.primate.wisc.edu (Brain in Neutral) writes: > > The csh "time" command produces lines of the form: > > 0.1u 0.4s 0:01 41% 18+17k 0+11io 21pf+8w There's a way (undocumented on a lot of systems) to change this output to any format you like. I have a file called "settime" that makes a more readable output. Here's a demo (note that the output has different numbers, but that's not the fault of the "settime" file): % date Mon Apr 18 10:48:56 EDT 1988 % set time=0 % date Mon Apr 18 10:49:06 EDT 1988 0.0u 0.0s 0:00 50% 8+0k 1+0io 0pf+0w % source ~/.lib/settime % date Mon Apr 18 10:49:25 EDT 1988 user=0.0 sec, system=0.0 sec, elapsed=0:00 min, cpu use=9%, 0 swaps, 2 disk reads, 1 disk writes. Data+stack+text pages: 24 avg., 3 max. Here's what's in the "settime" file. Sorry, but I couldn't see putting it into a shell archive. --Jerry Peek, Northeast Parallel Architectures Center, Syracuse, NY jerryp@amax.npac.syr.edu (315)423-1722 ----------- cut here and save in file named "settime" ----------- # file that sets C-shell "time" variable in my own format. # shows time of all processes and built-in commands. # Usage: % source ~/.lib/settime set time = (0 " user=%U sec, system=%S sec, elapsed=%E min, cpu use=%P, %W swaps,\ %I disk reads, %O disk writes. Data+stack+text pages: %K avg., %M max.")