In recent decades, with the placement of LiDAR remote sensing instruments in orbit, we now have global coverage of the bare ground elevation on the Earth, Mars and beyond. Encoded in such planetary LiDAR data are interesting landscape features that promise to further our knowledge of planetary topography. However, discovery of such features comes with 3 major challenges: First, the volume of planetary LiDAR data can be massive, often comprising of hundreds of millions to billions of data points. This  calling for analytical algorithms with great efficiency. Second, interesting features can often repeat themselves in multiple scales in local regions, thus it is vital to enable multi-scale feature discovery. Third, planetary LiDAR data can be heterogeneous, and evaluation of the quality of the extracted features can often be hampered by a variety of interfering factors. In response to these challenges, we propose FARMYARD, a generic pipeline for Feature Discovery From Planetary LiDAR Data. To the best of our knowledge, this is the first time such a pipeline has been proposed, which provides a brand new methodology for comparative studies of planetary topography. Specifically, drawing on the parallel computing power of the Graphics Processing Unit (GPU), we propose a novel pseudo-on-pass sweep (POPS) framework for fast and memory-efficient feature extraction for massive planetary LiDAR data, a two-level division scheme for local regions with support for multi-scale features, and a Domain-Shifted Partition (DSP) scheme for feature evaluation that is robust against interfering factors. To showcase the utility of our FARMYARD pipeline, we deploy it to an ongoing real-world research project called PARKER, which seeks to find topographical signatures of life by discovering features that can potentially distinguish between the Earth and alien worlds with no known life activity. We also highlight the efficiency of our POPS framework with experiments on both synthetic and real data, which can be hundreds or even thousands of times faster than its CPU-based counterpart, including an MPI-based multi-core parallel solution.