Search papers, labs, and topics across Lattice.
2
0
2
Divergent ODE trajectories that typically cripple GPU vectorization in high-level frameworks run up to 700x faster when mapped directly to thread-level Rosenbrock-Wanner kernels with sparse matrix support.
Propagating mega-constellations is now 1500x faster thanks to a JAX-based SGP4 reimplementation, making large-scale collision avoidance tractable.
