Optimization
TODO: verify behavior on target decreasing and increasing. We know the first target decrease (cf #2 (closed)) but at what height is the first target increase? It's around height 56447.
| height | time | note | commit |
|---|---|---|---|
| 33 000 | 4m33s | cdeb0906 | |
| 33 000 | 3m28s | Debugging Dissolve removed |
cdeb0906 |
| 33 000 | 2m21s | Unnecessary llbracket call removed |
271633bc |
| 33 000 | 0m44s | Pre-compute the level of a block |
48c663ec |
| 33 000 | 0m37s | Avoid multiple score computations |
78ed5cfd |
| 33 000 | 0m35s | Pre-compute the score of a block |
e9e11cac |
| 33 000 | 0m09s | Get rid of whole chain copy.deepcopy
|
5f271280 |
| 60 000 | 0m19s | For the next lines debugging, compareCorrectness are set to False
|
43fe2bd4 |
| 801 124 | 7m03s | Enable debugging and compareCorrectness during last compression iteration |
b39d7e69 |
| 217 017 | 4m09s | Correct full chain execution (thanks to previous_score update) |
d1bb5df0 |
| 60 000 | 0m07s | Optimized everything except maybe U
|
a77b60bd |
| 801 124 | 79m5s | Unknown result and and long time | 5abfedd7 |
| 428 079 | 2m31s | Correct partial result | 54ce05cd |
To optimize:
-
theory -
pass already computed values -
use a better programming language (cf #1) -
floating computation may slow down? -
correctness optimize cf 54ce05cd entry above
Consider pypy3 -m cProfile -s tottime mining_in_logarithmic_space.py --load-from-headers for optimizing.
Edited by LOISON Benjamin