Reproduce “FAME-QP: A GPU-Accelerated Quadratic-Programming Solver for Real-Time Model Predictive Control” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Santosh Mohan Rajkumar, Debdipta Goswami, Stephanie Stockar
FAME-QP accelerates constrained optimization in model predictive control by matching GPU execution to constraint structure. It handles diagonal rows elementwise and selects fused or factored operators, with graph replay and penalty probing reducing overhead. In MuJoCo humanoid simulation at horizon 20, mean solve time was 1.04 ms versus 3.25 ms for ReLU-QP, with similar balance deviations; this is a simulation result, not a hardware deployment.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.