Hello,
I am trying to optimize my code(python script) to speed up,
my data is unstructured data with:
# of cells 604 829
# of points 725 864 (float)
memory 67,4922 MB
fields points: PointIDs
fields cell: CellIDs Flag Tiff Scalars
which comes from a image (unifor rectiliniar grid) that i separate in sections of it by thresholds therefore becoming unstructured data.
after that i keep doing different functions (one after the other) while playing with the PointIDs and CellIDs fileds but i do not use anymore the Flag and Tiff scalars.
I thought to myself, that a pass array to remove the other fields will reduce the input of my functions and therefore make them faster (and less RAM demanding?), but at least in timings i see a huge drop on performance, eg. for 19 calls of my function, without it, it takes 25.5s where with it, it takes 261s.
the only thing i could think are:
- there is calls for
UpdatePipeline()so maybe this is making to re run the pass array filter? and therefore added to the time of my function? from my initial tests it looks like it is this the issue? - the pass array filter is not what i thought, and i am missing something
→ if it is 1. in general I should not use UpdatePipeline() but instead filter.UpdatePipeline() for the new filters created? is that so?