Indexed summary. This entry is an agent-written synopsis of an article first published at thomasahle.com. Read the original for the full text.
This page explains an optimised method for evaluating degree-n polynomials that beats the well-known Horner scheme. While Horner's method requires n multiplications, a preprocessing step on the coefficients reduces that count to ⌊n/2⌋+1 for any monic polynomial, with one extra multiplication for the general case. The technique is relevant for approximating mathematical functions such as exp, sin, and cos, as well as for cryptographic and coding-theory applications where polynomial evaluation over finite fields is common.
Key points
- Horner's method evaluates a degree-n polynomial in n multiplications; the new approach achieves ⌊n/2⌋+1 after an offline preprocessing step.
- The preprocessing optimises the coefficient chain for the specific polynomial, so the speedup is free at evaluation time once completed.
- Applications include function approximation (transcendentals), hashing, and cryptography over finite fields.
- An interactive web tool accepts a polynomial and a field choice, then outputs the preprocessed evaluation steps.
- The method is classical in spirit but not widely known among practitioners.
Why it matters
For numerical libraries or cryptographic implementations that evaluate the same polynomial repeatedly, cutting multiplications by roughly half can directly reduce latency and energy use. The interactive tool makes the technique accessible to people who need a faster evaluator without deep study of the underlying algebra.
Source: Evaluating Polynomials Fast