The release note says it might improve INSERT/UPDATE performance x2
But my testing case (PostgreSQL, FireDAC, INSERT 10000 records) - only shows a minor 5% improvement (dropped from 6 sec to 5.7 sec) - I mean, not bad, but far less than x2
Hence in what testing environment/set-up, can we accomplish the x2 performance boost?
The performance increases "up to" 2x faster. Performance always need to be measure, and each scenario is different.
The usual scenarios where the performance was improved is when the bottleneck is reading the object values and building the SQL statements. If most of your 6 seconds is taken by the database operations itself (FireDac), then there will be not much improvement in that area.