🚀 go-pugleaf

In article <3nrfg3$baq@gw.digibd.com>, Robert Lipe <robertl@arnet.com> wrote:
>
>$ time perl < /dev/null
>
>real    0m0.15s
>user    0m0.02s
>sys     0m0.10s
>
>$ time perl < /dev/null
>
>real    0m0.08s
>user    0m0.02s
>sys     0m0.06s
>

Not wanting to be left out.  Here are my results, loading perl
100 times and /bin/sh once.

In order below, SGI chalange server 6 CPU and 256Meg ram, three
486DX2/66 each with 32M ram. SCO 3.2 on one the others running
BSD/OS 1.1 and 2.0 respectively.  The SCO box is consistantly half
the speed of the BSDI boxes.  But on average SCO loads in about 80uSec
once it has been cashed.  Pretty much identical with Robert's results.

pluto src $ uname -a
IRIX pluto 5.2 02282015 IP19 mips
pluto src $ perl -e 'print "perl;"x100;' | time sh
real    0m4.84s
user    0m0.86s
sys     0m3.88s

salt cfedde $ uname -a
salt salt 3.2 2 i386
salt cfedde $ perl -e 'print "perl;"x100;' | time sh
real    0m8.06s
user    0m2.23s
sys     0m4.96s

point chris $ uname -a
BSD/386 point.ezsrc.mrg.uswest.com 1.1 0 i386
point chris $ perl -e 'print "perl;"x100;' | time sh
        5.87 real         1.05 user         4.77 sys

loupe chris $ uname -a
BSD/OS loupe.ezsrc.mrg.uswest.com 2.0 BSDI BSD/OS 2.0 Kernel #0: Thu Feb
23 13:09:18 MST 1995
chris@loupe.ezsrc.mrg.uswest.com:/usr/src/sys/compile/LOUPE  i386
loupe chris $ perl -e 'print "perl;"x100;' | time sh
        5.51 real         1.14 user         5.11 sys