The Riemann theta function is a theta function of complex variables that occurs in the construction
of quasi-periodic solutions of various
equations in mathematical physics (Deconinck et al. 2004). Any Abelian
function can be expressed as a ratio of homogeneous polynomials
in Riemann theta functions (Igusa 1972, Deconinck et al. 2004).
where
denotes the transpose. It is an entire
function of . Its modular transformation laws
involve nontrivial factors (NIST DLMF, Ch. 21). For integer column
vectors
and , the translation law is
Riemann (1857) first considered these functions associated with Riemann surfaces, and the most general form of the Riemann theta function defined above
was first considered by Wirtinger (1895).
Mumford (1983, 1984, 1991) gives an overview of the theory. Umemura's formula expresses polynomial roots in terms
of theta constants and Abelian
integrals on hyperelliptic curves (Umemura
1984). Algorithms for numerical evaluation are developed by Deconinck et al. (2004).
Elkies and Kieffer (2025) give a faster method whose cost, for fixed , is quasi-linear in the requested precision for Gaussian-scaled
theta values at exact reduced arguments. The scaling multiplies by .
The uniform complexity bound does not include reduction of arbitrary input arguments.
Their method also evaluates derivatives, with bounds
depending on the arguments.