**Last Updated 05/13/2003

This is a side by side comparison of the various servers.  This is correlated
based on the newest server releases currently available:

Server Tested on: Pentium III 650 with 250M memory
OS: Redhat Linux (2.2.14-12 kernel, egcs-2.91.66 compiler)
Date of testing:  Oct 2nd, 2001
Gathered info by: UNIX commands (ps, ls, top)

All servers were started at the same time using similiar (as close as possible)
configurations (cache, disk options, optimization, compression).
This is with a brand new db (server defined) and immediately gathered on start.
All binaries were compiled with the same optimizations specified.
-------------------------------------------------------------------------------
Server Type     | Obj Size | DB Size  | Memory Footprint | CPU% | Binary Size
-------------------------------------------------------------------------------
TinyMUSH 3.0p2  | 2        | 25K      |  3.58M           | 0.5% | 0.56M
TinyMUX 1.6     | 2        | 25K      |  3.62M           | 0.5% | 0.47M
TinyMUSH 2.2.4u1| 2        | 25.5K    |  3.04M           | 0.2% | 0.41M
MUX 2.1.2.20    | 2        | 24K      |  4.48M           | 0.0% | 0.62M
PENN 1.7.4p10   | 3        | 1.1K     |  3.46M           | 0.3% | 0.64M
Rhost 3.2.4p12  | 2        | 25K      |  5.52M           | 0.0% | 1.58M
-------------------------------------------------------------------------------

Now we slammed the db by loading 50 total 30 char length attributes, which all
held a 4 char length string, on 2000 seperate objects.

The total user-defined attributes defined in all were 100,000.
The total objects defined were 2000 added to the default.
After load, immediate @dbck and @dump performed then the results gathered.
The CPU utilization was taken after a 5 minute wait after a fresh startup
of the game, to take into account any lingering cache/other that could effect
the results. 
-------------------------------------------------------------------------------
Server Type     | Obj Size | DB Size  | Memory Footprint | CPU% | Growth Rate
-------------------------------------------------------------------------------
TinyMUSH 3.0p2  | 2002     | 6.21M    | 15.58M           | 1.3  |  6.05K/obj
TinyMUX 1.6     | 2002     | 6.21M    | 15.14M           | 1.8  |  5.75K/obj
TinyMUSH 2.2.4u1| 2002     | 9.04M    | 44.13M           | 1.8  | 20.55K/obj
MUX 2.1.2.20    | 2002     | 7.48M    | 14.11M           | 1.3  |  4.80K/obj
PENN 1.7.4p10   | 2003     | 4.41M    | 12.28M           | 0.0  |  2.92K/obj
Rhost 3.2.4p12  | 2002     | 3.68M    | 11.92M           | 1.5  |  3.20K/obj
-------------------------------------------------------------------------------

Something of interest: MUX 1.6, TM3.0, and Tiny 2.2.4 had difficulties
loading the information in and required deleting the database, starting from
scratch and repeating the test until a successful load was accomplished.
Starting from scratch included rm -rf'ing the directory and doing everything
from point blank.  I wanted no calls of 'foul' or 'taint' or 'bias' in these
tests.  There were no issues loading the db with MUX 2.1, PENN 1.7.4p10, 
and Rhost 3.2.4.