Nmap Development: nmap process eats 737MB and runs 5 days!

["Jesse W. Asher" <jasher1 () tampabay rr com>]

I've been trying to run a scan of a class C network using the following:

/usr/local/bin/nmap -v -sT -sU -sR -O -p 1-65535 -P0 -T4 -n -oA 
/usr/local/nmap/172.30.1.log --osscan_limit --min_hostgroup 100 
172.30.1.0/24
I started it on Sept. 28th and just had to kill it because it had gotten 
so big.  Here's the relevant line from top:
4003 root      15   0  737M 380M    84 D     1.1 75.8  9248m   1 nmap

The 5th and 6th fields show the memory related fields.  I'm running this 
on RedHat 9.
So, this begs the following questions:

1)  Why did it take so long to run?
2)  Why did it get so big memory wise?
3)  How can I do a complete scan of a subnet (complete meaning all 
possible TCP and UDP ports) and not have it take so long (and actually 
finish)?  I have to have pings turned off since they are disabled 
internally (long story and something that needs to be changed eventually).
Any assistance would be appreciated!  Especially since I've been trying 
to get this to work for three weeks!  ;-)