I had issues...
So my path to 10mm3/s included making the head screw barely finger tight, loosening the screws on the feed mechanism a bit, reducing the tension on the filament mechanism as much as I could, and setting the feed motor amperage at 1300. I saw no improvement at 1500 which I did try.
There is a bit of oozing as it changes layers but that's a separate problem. I had a horrible time with underextrusion and could barely print at 50mm/s.
The other changes were incremental but the largest jump was when I loosened the head screws.
A good test would be tightening those pieces and recording results with an adjustable torque driver so we can have precise points of measure to compare rather than just 'slightly loose' or 'barely finger tight.'