Year: 2026
Type: preprint
Abstract: Three-dimensional (shape) similarity is a useful primitive for virtual screening and molecular machine learning, but current open implementations are limited in performance by a quadratic inner loop across pairs of query-and databasemolecule atoms. ROSHAMBO3 achieves a 78–154× speedup over the previous open SOTA ROSHAMBO2 by using query-amortized Gaussian fields that make the inner loop linear combined with system-scale optimization at the algorithmic, GPU, and host layers, screening up to 55M query–conformer pairs/s on NVIDIA GPUs, with secondary CPU (x86, ARM) and Apple Metal backends. We further describe the human expert–AI interaction loop that guided the exploration and implementation of ROSHAMBO3 and discuss implications for the scientific software commons. ROSHAMBO3 is open source at https://github.com/kimiatx-oss/roshambo3.
Source: ChemRxiv
Author: Imran S. Haque
Institution: Kimia Therapeutics
Language: English
Subfield: Drug Discovery
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7226027261
