When to use a hash partition

Oracle EE 12.1.0.2

I have two existing non-partitioned tables.  I'm just listing the relevant columns, there are more.

Table1 - 700 million rows

ID 

START_DT_GMT

DEVICE_ID

CLIENT_ID

PK = ID

Table2 - 2 billion rows

ID

TABLE1_ID

PK = ID

FK = TABLE1_ID with cascade delete enabled

I plan on converting Table1 to an interval partitioned table based on START_DT_GMT with an interval of 1 day.  We get currently get 500,000 rows a day and we keep no data older than 5 years.

We delete data based either CLIENT_ID+START_DT_GMT or by  DEVICE_ID. (Client data expires after a set period of time and individual devices are deleted).