Using chunk size and threading to improve the Historical Assignment hcm extract performance issue

Summary:

I have created an HCM Extract for assignment that retrieves historical data for employees. There is no date filter, it picks up past, future and present.There are 144567 employees that it retrieves in the baseline file and it is taking 1 hour 30 min to run even when we run for incremental loads using the changes only parameter.

To improve the performance I used the process configuration group and made the following changes -

Now it is completing in 60 min. What should be the most optimal value that i pass in chunk size and thread so that this process runs undr 30 min.