Cached Entries
Write a light-weighted and faster get_e_above_hull, with benchmark to the slow version.
Current Best:
Using batched parallel and cache : (batch_size = 50, total 2000 sample, with 2000/50=40 batch)
- 10 CPU: 0.01s/per structure
- 20 CPU: 0.006s/per structure
- 40 CPU: 0.003s/per structure
For details, see #20 (comment 3929)