new code on the server

[Peter Van Epp]

	Its looking pretty good. Using the latest (from this morning) 
argus-3.0.0 and clients.rc.44 and using the same tcpdump file as input
5 machines match other than man records (which may or may not be an issue :-)).
As well there are some performance numbers which are quite interesting too.
This covers all cases Intel/Opteron 32/64 bit and Mac G4 (32 bit), G5 (64bit)
and IBM Power5 (64 bit) ppc processors so both endians and both word sizes:

-rw-r--r-- 1 vanepp users 1672399726 2007-05-17 09:53 eth23.tcp

FreeBSD 6.2 600meg P3

%time argus -r eth23.tcp -w eth23.intel32.argus
18.169u 21.101s 1:34.74 41.4%   277+9385k 12818+281io 10pf+0w

Mac G4 (32 bit) 

sh1012d:~ peter$ time ./argus -r eth23.tcp -w eth23.G4Mac32.argus  

real    1m10.111s
user    0m18.744s
sys     0m22.814s


IBM P510 quad Power5 PPC 64 bit:

vanepp at sniffer1:~> time argus -r eth23.tcp -w eth23.G5ppc64.argus 

real    0m7.492s
user    0m4.024s
sys     0m3.277s

Mac OS 10 quad core G5 ppc 64 bit 

test4:/var/log/argus vanepp$ time argus -r eth23.tcp -w eth23.Mac64.argus

real    0m9.760s
user    0m3.966s
sys     0m5.548s

Sun 4100 Opteron box:

[vanepp at burgess ~]$ time ./argus -r eth23.tcp -w eth23.Opteron64.argus

real    0m5.550s
user    0m2.566s
sys     0m2.739s

	All things being equal (which they may or may not be in things such 
as memory :-)) the Opteron (which is counter intuitive because it is not
network byte order) appears fastest of all 5 machines.
	ra output from ra3 -r xxx.argus -n on all machines only varies in the
man records (which may or may not be correct):

vanepp at sniffer1:~> diff eth23.G5ppc64.ra.out eth23.intel32.ra.out
154978a154979
> 07-05-17 09:45:59             man                549      0                     4065      1  1780416   154978         4065     82649308   CON
236955c236956
< 07-05-17 09:53:45             man                  0      0                     1114      1  2793055   236955         1114    147684976   STP
---
> 07-05-17 09:45:59             man                  0      0                     1114      1  1012639    81977         1114    123823144   STP

vanepp at sniffer1:~> diff eth23.G5ppc64.ra.out eth23.Mac64.ra.out
236955c236955
< 07-05-17 09:53:45             man                  0      0                     1114      1  2793055   236955         1114    147684976   STP
---
> 07-05-17 10:58:12             man                  0      0                     1114      1  2793055   236955         1114    124988008   STP

vanepp at sniffer1:~> diff eth23.G5ppc64.ra.out eth23.opteron64.ra.out
236955c236955
< 07-05-17 09:53:45             man                  0      0                     1114      1  2793055   236955         1114    147684976   STP
---
> 07-05-17 11:00:46             man                  0      0                     1114      1  2793055   236955         1114    147684976   STP

diff eth23.G5ppc64.ra.out eth23.G4Mac32.ra.out
209022a209023
> 07-05-17 16:02:01             man                571      0                     3812      1  2475888   209022         3812    111349812   CON
236955c236956
< 07-05-17 09:53:45             man                  0      0                     1114      1  2793055   236955         1114    147684976   STP
---
> 07-05-17 16:02:01             man                  0      0                     1114      1   317167    27933         1114    124988008   STP

Peter Van Epp / Operations and Technical Support 
Simon Fraser University, Burnaby, B.C. Canada