Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userOptimization of the scalar complexity of Chudnovsky$^2$ multiplication algorithms in finite fields
113
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
We propose several constructions for the original multiplication algorithm of D.V. and G.V. Chudnovsky in order to improve its scalar complexity. We highlight the set of generic strategies who underlay the optimization of the scalar complexity, according to parameterizable criteria. As an example, we apply this analysis to the construction of type elliptic Chudnovsky$^2$ multiplication algorithms for small extensions. As a case study, we significantly improve the Baum-Shokrollahi construction for multiplication in $mathbb F_{256}/mathbb F_4$.
rate research
Read More
We propose a Recursive Polynomial Generic Construction (RPGC) of multiplication algorithms in any finite field $mathbb{F}_{q^n}$ based on the method of D.V. and G.V. Chudnovsky specialized on the projective line. They are usual polynomial interpolation algorithms in small extensions and the Karatsuba algorithm is seen as a particular case of this construction. Using an explicit family of such algorithms, we show that their bilinear complexity is quasi-linear with respect to the extension degree n, and we give a uniform bound for this complexity. We also prove that the construction of these algorithms is deterministic and can be done in polynomial time. We give an asymptotic bound for the complexity of their construction.
The Chudnovsky and Chudnovsky algorithm for the multiplication in extensions of finite fields provides a bilinear complexity which is uniformly linear whith respect to the degree of the extension. Recently, Randriambololona has generalized the method, allowing asymmetry in the interpolation procedure and leading to new upper bounds on the bilinear complexity. We describe the effective algorithm of this asymmetric method, without derivated evaluation. Finally, we give examples with the finite field $F_{16^{13}}$ using only rational places, $F_{4^{13}}$ using also places of degree two and $F_{2^{13}}$ using also places of degree four.
We indicate a strategy in order to construct bilinear multiplication algorithms of type Chudnovsky in large extensions of any finite field. In particular, by using the symmetric version of the generalization of Randriambololona specialized on the elliptic curves, we show that it is possible to construct such algorithms with low bilinear complexity. More precisely, if we only consider the Chudnovsky-type algorithms of type symmetric elliptic, we show that the symmetric bilinear complexity of these algorithms is in $O(n(2q)^{log_q^*(n)})$ where $n$ corresponds to the extension degree, and $log_q^*(n)$ is the iterated logarithm. Moreover, we show that the construction of such algorithms can be done in time polynomial in $n$. Finally, applying this method we present the effective construction, step by step, of such an algorithm of multiplication in the finite field $F_{3^{57}}$.
Thanks to a new construction of the so-called Chudnovsky-Chudnovsky multiplication algorithm, we design efficient algorithms for both the exponentiation and the multiplication in finite fields. They are tailored to hardware implementation and they allow computations to be parallelized while maintaining a low number of bilinear multiplications. We give an example with the finite field ${mathbb F}_{16^{13}}$.
We present three families of minimal border rank tensors: they come from highest weight vectors, smoothable algebras, or monomial algebras. We analyse them using Strassens laser method and obtain an upper bound $2.431$ on $omega$. We also explain how in certain monomial cases using the laser method directly is less profitable than first degenerating. Our results form possible paths in the search for valuable tensors for the laser method away from Coppersmith-Winograd tensors.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
