The pipeline now looks way more clean.
The performance didn't show any loss,
I think this is simply because most of the caching was simply unneeded as some cached values were so tiny that simply no overhead is implied by calculating them in real time
Conclusion
The pipeline now looks way more clean. The performance didn't show any loss, I think this is simply because most of the caching was simply unneeded as some cached values were so tiny that simply no overhead is implied by calculating them in real time
Performance notes:
Stresstest:
Many chunks