Since powershell is trying to do so much under the covers, it can  get silly at times.  When you have many thousand files in a directory, (which is a silly idea to begin with) powershell slows down to a crawl when the dir command bundled in ‘cmd.com’ still works great.  The easiest answer is to simply use that command instead, break you data into smaller chunks, and then powershell-ize your data after the fact.  If you don’t do this, you could really screw yourself when your powershell script appears to run slowly or stop completely. 

<pre>

cmd /c dir

</pre>

This will run the old dir command from cmd.com for you from powershell.  But what if cmd.com goes away, as Microsoft promises?  Either you must write a wrapper that you can replace with something other than powershell without any code change, or your must find that other command today.  I am thinking of a unix port like the real ‘ls’ (not powershell alias) or ‘find’.  Find suffers a similar issue in unix, but only if you give it parameters, if you just do a straight find ( find .  or find -print ), everything works as it just goes one record at a time.  Also, in powershell, what counts is how many files you have in one directory, or ask about from one command.