SSAS -- to Partition or not?
Recently I had a question from a client whether there was really any performance improvement in SSAS cube processing when using partitioning strategies vs. just using a single partition.
While intuitively I "knew" that partitioning cube fact tables results in better parallelism and thus increased performance (when adequate CPU, memory and I/O resources can support it), I realized I really didn't have any empirical "proof" to backup my professional opinion. So--I setup a simple test.