1. data structure augmentation
부가적인 정보를 활용한다.
2. store precomputed results
3. caching
4. lazy evaluation - 불필요한 evaluation을 피함
5. packing - 압축
6. interpreters
7. code motion out of loops
- loop 내부의 내용을 최소화 한다.
8. combining tests
- 경계 test대신 sentinel을 둔다.
9. loop unrolling
- loop를 풀어쓴다.
10. transfer-driven loop unrolling
11. unconditional branch removal
12. loop fusion
13. exploit algebraic identitied
14. short-circuiting monotone functions
15. reordering testss
16. precompute logical functions
17. boolean variable elimination
18. collapsing function hierarchies
19. exploit common cases
20. coroutines
21. transformations on recursive functino
22. parallelism
23. compile-time initialization
24. exploit algebraic identities
monotonicity이용, incremental algorithms 이용
25. common subexpression elimination
26. paring computation
27. exploit word parallelism
댓글 없음:
댓글 쓰기