Vmark fail after 1 min run on windows2003 AMD 64bits with 32 bits VM C2 flag

JDK version
=============
java version "1.5.0-beta"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0-beta-b23)
Java HotSpot(TM) Server VM (build 1.5.0-beta-b23, mixed mode)

Platform
=============
Windows_NT JTG-AMD2 5 02 586

Error message
=============
Please telnet with root/admin and change to d:\tmp directory to check it.

#An unexpected error has been detected by HotSpot Virtual Machine:
# Internal Error (4E4D4554484F440E43505004F6), pid=4068, tid=1568
# Java VM: Java HotSpot(TM) Server VM (1.5.0-beta-b23 mixed mode)
 --------------- T H R E A D ---------------
 Current thread (0x0d9786f8): JavaThread "Receiver-197" [_thread_in_Java, id=1568]
 Stack: [0x2d970000,0x2d9b0000)
[error occurred during error reporting, step 110, id 0xe0000000]

OS: Windows Server 2003 family Build 3790 Service Pack 1, v.1069
CPU:total 2(active 2) family 15, cmov, cx8, fxsr, mmx, sse, sse2
Memory: 4k page, physical 1047984k(544108k free), swap 2503324k(1870204k free)
###@###.### 2003-10-14


The error messages first shown up in "run.vmark.out" are:
java.net.BindException: Address already in use: connect

The error messages shows up after about 20 iterations.

From "netstat", it looks like that the ports used by vmark client in previous iterations are not released fast enough. Each vmark client needs 200 ports.
At the time when the error occurs, the ports used by the new vmark client
have conflict with the ports used by the older clients which already exit but the ports used by which are not released yet.
 
I added "sleep 10" after each test iteration so vmark server has more time
to reclaim ports. The error is gone with the 10 second delay.

###@###.### 2003-10-17