The curve was found by adapting Dujella’s 2002 use of Fermigier’s rank-\(\ge 8\) construction with rational 2-torsion. Dujella used products of three primes \(p_i\equiv1\pmod4\), giving four representations as sums of two squares. I instead used products of four such primes, giving eight representations and many possible Fermigier curves to test.
he best case came from \(5\cdot13\cdot29\cdot41\), using \((106,257)\), \((113,254)\), \((142,239)\), and (166,223)
jesper-petersen ·