Given a set of rectangular pieces and a rectangular container of fixed width and variable length, the two-dimensional strip packing problem (2D-SPP) aims at orthogonally placing all the pieces within the container such that the pieces do not overlap and the total length of the resulting layout is minimized. Up to now, most of the benchmark instances to the 2D-SPP reported in the literature are rather small. In this paper 360 large random-based instances with 1000 pieces each are introduced. Further essential problem-specific parameters such as the number of item types and the quotient of the container width and the mean item dimension are varied in a systematical way. Results for the 360 instances calculated by a recently developed genetic algorithm are also reported and may serve as reference solutions.