GTSAM 4.3 Adds CUDA Backend for Nonlinear Optimization
fdellaert · x · 2026-08-22
GTSAM is introducing an experimental CUDA backend in version 4.3 to accelerate nonlinear factor-graph optimization, targeting robotics and computer vision applications.
Technical Details:
- Dual Optimization Paths: A general LM path (sparse Jacobian on host) and a specialized SfM path (projection linearization and Schur complement on GPU).
- Solver Layer: Shared layer providing dense Cholesky, cuDSS, and PCG.
- Performance: Benchmarks on a 10,000-pose graph show significant LM time reduction when combined with FAST-Sync and exact Lie-group Jacobians.
- API: The new gtsam::cuda::SparseLevenbergMarquardtOptimizer accepts standard factor graphs, keeping GPU allocations persistent across iterations.
More from Infra
- AgenticROS adds Organizations and Teams support for robot sharing — chrismatthieu · 2026-08-22
- Polymarket: 69% chance a U.S. state enacts data center moratorium by year-end — Polymarket · 2026-08-22
- New narrative for AI datacenters: Jobs, lower taxes, and better infrastructure — alexvoica · 2026-08-22
- Ramp launches Router LLM gateway to cut inference costs by 40% — round · 2026-08-22
- MCP Isn't Replacing APIs: It's Changing Who APIs Are Designed For — kush_patil · 2026-08-22
- Data Center Opposition Surged from 42 to 75 Percent in One Year — The Decoder · 2026-08-22