I second Allen's questions.   And yes, the hyperthreading is the reason for the 4 cpu's in task manager.

When I first took this job I had a server do the same thing.  They were restarting it to get it going again.  You probably have a bottleneck somewhere.  It could be caused by hitting the knee in an exponential curb.  The database growth may have triggered this.

I fixed our problem by fixing a batch job that ran every hour.  It needed more than this, but it was enough to keep it going. 

Here are some potential, but not limited too probs you may be having

• Locking
• excessive and/or poor indexing
• too many objects in tempdb
• i/o bottleneck
• cpu bottleneck