Hi @kailinatsafe and thanks for reaching out. I am not allowed to share the data public but have simulated something similar with a workspace:
- 20000 points "randomly" generated within a X and Y range - written as points to the filegeodatabase
- Buffered these points "randomly" with buffers in range 5 to 50 meters , replaced with a 2D box which is rotated "randomly" within -45-45 deg and written as polygons to the geodatabase
- chopped the boxes to max 2 vertices and written results as lines
All to simulate building centerpoints, outlines and polygons. With these data FME2022 is about 30% slower (using the GUI). With the actual data we are at approx 40%.
Best Pål Herman (zipped geodatabase attached)