We present a framework for massively parallel simulations of climate impact models in agriculture and forestry: the parallel System for Integrating Impact Models and Sectors (pSIMS). This framework comprises a) tools for ingesting large amounts of data from various sources and standardizing them to a versatile and compact datatype; b) tools for translating this standard datatype into the custom formats required for point-based impact models in agriculture and forestry; c) a scalable parallel framework for performing large ensemble simulations on various computer systems, from small local clusters to supercomputers and even distributed grids and clouds; d) tools and data standards for reformatting outputs for easy analysis and visualization; and d) a methodology and tools for aggregating simulated measures to arbitrary spatial scales such as administrative districts (counties, states, nations) or relevant environmental demarcations such as watersheds and river-basins. We present the technical elements of this framework and the results of several example climate impact assessment and validation exercises that involved large parallel computations.