When configuring a Databricks cluster for a machine learning project that demands high computational power, which cluster type is the most appropriate choice?