We present r.lfp, a new GRASS addon for computing the Longest Flow Path (LFP) across a large number of watersheds, built for scalable high-performance hydrologic modeling. Based on the Memory-Efficient Longest Flow Path (MELFP) algorithm, r.lfp delivers substantial performance improvements by combining tail recursion with hybrid OpenMP parallelism and minimal memory usage, making it especially well suited for large digital elevation models (DEMs). r.lfp integrates seamlessly into the GRASS ecosystem, leveraging GRASS's native data structures and parallel processing framework. It supports both subwatershed-level and watershed-level LFP analysis modes, and offers reproducible results even across varied multi-threaded environments. The module's design emphasizes computational efficiency, reproducibility, and scriptability, enabling rapid analysis in research and operational workflows. In benchmark comparisons, r.lfp achieves orders-of-magnitude speedups over traditional implementations, without sacrificing accuracy. It also plays a key role in supporting advanced hydrologic modeling tasks, such as river network analysis and travel time estimation. As a contribution to the open-source geospatial community, r.lfp demonstrates how modern algorithmic design and memory-conscious parallelism can revitalize established GIS platforms like GRASS. Its development is part of a broader initiative to modernize and extend GRASS for continental-scale hydrologic modeling, with strong emphasis on scalability, interoperability, and scientific transparency. We highlight case studies on continental datasets to showcase r.lfp's performance, scalability, and integration potential with other GRASS modules and external tools.