Category: Sampling

  • On Hash Functions for the GPU

    This blog posts shows how the PCG2D hash function works geometrically. Further, we discuss a quantitative statistical test that may help to optimize parameters as compared to classical pass/fail test suites.

  • Sampling the disk from a square: The adoption method

    We introduce a new disk sampling algorithm that maps the square locally isometrically to the disk. Similar as to the classical rejection algorithm, this allows to push low discrepancy sampling strategies from the square to the disk with no warping. In contrast to rejection, it is amenable to SIMD implementations and our benchmark shows 2-3x…