My experience with these issues (not on SQL Server though) is that overdone multi-tasking is often the cause of the problem. If there is similar/connected data/tables queried at (almost) the same time by many connections, the DBMS may have trouble keeping all the isolation at check. This is not that much of an issue of disk usage as to making some connections wait for things to be done by other ones. Synchronization is very expensive in terms of CPU usage.
The 100 connections is way too much in my opinion. (In my experience again) even 20 connections asked to be done by one machine may be overly optimistic.