system & user time on RH/Debian

Cyril Chaboisseau (Cyril.Chaboisseau@obs.coe.int)
Mon, 26 Jan 1998 16:26:31 +0100


This is a multi-part message in MIME format.

--------------603051A113E
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

after performing some bench on my new dual PII 266Mhz running Debian
1.3, I noticed that the results looked strange to me :
some calculations under bc (2000 decimals of pi) returned a much bigger
value for the system time compared to the user time

since the kernel is recompiled from source, i first blamed it on time or
bc program, so i downloaded time-1.7 and bc-1.04, recompiled the all
thing and ran the program again :

./time ./bc pi.bc
4.37user 111.06system ... 100%CPU ...

while another machine with Red Hat 4.2 and 5.0, I've got the opposite :
30.45user 0.01system ...

my question is :
What part of the system/kernel is in charge of attributing sys/user time
and is Debian really wrong returning such results with a completely
recompiled kernel/time/bc programs

my config :
dual PII 266Mhz 440LX 256M sdram running 2.0.32+tulip patch
PCI cards : DPT3334UW, NCR53c810, S3, 100Mb/s DS21140 tulip

BTW, this system works really fine since 2/3 months

Thanks

-- 
Cyril Chaboisseau

--------------603051A113E Content-Type: text/plain; charset=us-ascii; name="pi.bc" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="pi.bc"

scale=1000 define a(n) { /* Fonction FACTORIEL */ auto i,j j=1 if ( n==0 ) return(1) for(i=2; i<=n; i++) j *= i return (j) }

l=0 m=1 for(k=0; m!=l; k++) { m = l l += ((a(4*k))*(1103+26390*k))/((a(k)^4)*(396^(4*k))) } 1/((2*sqrt(2)/9801)*l) quit

--------------603051A113E--