Dask can utilize pandas dataframes and make them run across different cores. Also is great for tasks that take up too much memory as discussed by the OP. This may replace the need for you to use DASK in alot of projects since you can work with 3.5x the RAM now.