We present an algorithm that computes the Hasse-Witt matrix of given hyperelliptic curve over Q at all primes of good reduction up to a given bound N. It is simpler and faster than the previous algorithm developed by the authors.
We present an efficient algorithm to compute the Hasse-Witt matrix of a hyperelliptic curve C/Q modulo all primes of good reduction up to a given bound N, based on the average polynomial-time algorithm recently introduced by Harvey. An implementation
for hyperelliptic curves of genus 2 and 3 is more than an order of magnitude faster than alternative methods for N = 2^26.
We develop a statistical method to measure the interaction cross-section of Dark Matter, exploiting the continuous minor merger events in which small substructures fall into galaxy clusters. We find that by taking the ratio of the distances between t
he galaxies and Dark Matter, and galaxies and gas in accreting sub-halos, we form a quantity that can be statistically averaged over a large sample of systems whilst removing any inherent line-of-sight projections. In order to interpret this ratio as a cross-section of Dark Matter we derive an analytical description of sub-halo infall which encompasses; the force of the main cluster potential, the drag on a gas sub-halo, a model for Dark Matter self-interactions and the resulting sub-halo drag, the force on the gas and galaxies due to the Dark Matter sub-halo potential, and finally the buoyancy on the gas and Dark Matter. We create mock observations from cosmological simulations of structure formation and find that collisionless Dark Matter becomes physically separated from X-ray gas by up to 20h^-1 kpc. Adding realistic levels of noise, we are able to predict achievable constraints from observational data. Current archival data should be able to detect a difference in the dynamical behaviour of Dark Matter and standard model particles at 6 sigma, and measure the total interaction cross-section sigma/m with 68% confidence limits of +/- 1cm2g^-1. We note that this method is not restricted by the limited number of major merging events and is easily extended to large samples of clusters from future surveys which could potentially push statistical errors to 0.1cm^2g^-1.
