10.12.2012 Views

Prime Numbers

Prime Numbers

Prime Numbers

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

382 Chapter 7 ELLIPTIC CURVE ARITHMETIC<br />

algorithm whose complexity lies essentially between naive residue counting<br />

and the Shanks–Mestre algorithm. There is yet one more possible avenue of<br />

exploration: The DAGM of Exercise 2.42 might actually apply to truncated<br />

hypergeometric series (mod p) in some sense, which we say because the<br />

classical AGM—for real arguments—is a rapid means of evaluating such as<br />

the hypergeometric form above [Borwein and Borwein 1987].<br />

Incidentally, a profound application of the AGM notion has recently been<br />

used in elliptic-curve point counting; see the end of Section 7.5.2.<br />

7.27. Along the lines of Exercise 7.26, show that for a prime p ≡ 1 (mod 8),<br />

the elliptic curve E with governing cubic<br />

has order<br />

y 2 = x 3 + 3<br />

√ 2 x 2 + x<br />

<br />

#E = p +1− 2 (p−1)/4<br />

p−1 <br />

2<br />

p−1 mod ± p ,<br />

4<br />

where the mod± notation means that we take the signed residue nearest 0.<br />

Does this observation have any value for factoring of Fermat numbers? Here<br />

are some observations. We do know that any prime factor of a composite Fn<br />

is ≡ 1 (mod 8), and that 3/ √ 2 can be written modulo any Fermat number<br />

Fn > 5as3(23m/4 − 2m/4 ) −1 ,withm =2n ; moreover, this algebra works<br />

modulo any prime factor of Fn. In this connection see [Atkin and Morain<br />

1993a], who show how to construct advantageous curves when potential factors<br />

p are known to have certain congruence properties.<br />

7.28. Implement the ECM variant of [Peralta and Okamoto 1996], in which<br />

composite numbers n = pq2 with p prime, q odd, are attacked efficiently. Their<br />

result depends on an interesting probabilistic way to check whether x1 ≡ x2<br />

(mod p); namely, choose a random r and check whether the Jacobi symbol<br />

equality <br />

x1 + r x2 + r<br />

=<br />

n<br />

n<br />

holds, which check can be performed, remarkably, in ignorance of p.<br />

7.29. Here is a fascinating line of research in connection with Schoof<br />

point counting, Algorithm 7.5.6. First, investigate the time and space<br />

(memory) tradeoffs for the algorithm, as one decides upon one of the<br />

following representation options: (a) the rational point representations<br />

(N(x)/D(x),yM(x)/C(x)) as we displayed; (b) a projective description<br />

(X(x, y),Y(x, y),Z(x, y)) along the lines of Algorithm 7.2.3; or (c) an affine<br />

representation. Note that these options have the same basic asymptotic<br />

complexity, but we are talking here about implementation advantages, e.g.,<br />

the implied big-O constants.<br />

Such analyses have led to actual packages, not only for the “vanilla Schoof”<br />

Algorithm 7.5.6, but the sophisticated SEA variants. Some such packages are

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!