In this letter, we investigate the adoption and accelerationof a competing algorithm for low-power embedded systems.The algorithm, Efficient Large-Scale Stereo (ELAS) [12], isthe fastest CPU algorithm w.r.t. resolution on the Middleburydataset [13] and one of the most accurate non-global methods.ELAS is attractive since it very efficiently implements a slantedplane prior while its dense depth estimation is fully decompos-able over all pixels and, hence, suitable for parallel processing.Unfortunately, the intermediate step,i.e., estimation of coarsescene geometry through the triangulation of support points, is avery iterative, sequential and conditional process with an unpre-dictable memory access pattern; making it difficult to accelerateon an FPGA.