Research map / Learning theory and algorithms / Combinatorial and search optimization
FastDOG: Fast Discrete Optimization on GPU
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use the GPU-parallel FastDOG algorithms instead of BDD-CPU | BDD-based Lagrange decomposition · Benchmarks reported in Table 2 | runtime: 14 -> 13; 216 -> 110; 6 -> 0.2; 12 -> 0.4; 528 -> 54; 70 -> 14; 107 -> 9; 218 -> 13; 232 -> 39; 357 -> 137 | Helped |
| Use FastDOG instead of specialized CPU solvers | Structured prediction benchmarks · Graph matching hotel, house, worms; MRF color-seg, color-seg-n4, color-seg-n8, object-seg; cell tracking large | runtime: 3 -> 0.2; 3 -> 0.4; 214 -> 54; 155 -> 14; 9 -> 9; 30 -> 13; 3 -> 39; 90 -> 110 | Mixed |
| Use FastDOG instead of Gurobi | Structured prediction ILPs · Cell tracking large; MRF color-seg-n4; QAPLib large | runtime: 1584 -> 110; 980 -> 9; 6742 -> 6928 | Mixed |
| Use the proposed parallel dual update and perturbation rounding | FastDOG · Reported benchmark datasets | primal and dual solution quality | Mixed |
Closest papers by meaning
- DOGE-Train: Discrete Optimization on GPU with End-to-End Training (AAAI 2024 · similarity 0.88)
- From Sequential to Parallel: Reformulating Dynamic Programming as GPU Kernels for Large-Scale Stochastic Combinatorial Optimization (ICLR 2026 · similarity 0.87)
- B$^3$-PWL: GPU-Batched Branch-and-Bound for Piecewise-Linear Optimization with SOS2 Constraints (NeurIPS 2026 · similarity 0.87)
- FlashRNN: I/O-Aware Optimization of Traditional RNNs on modern hardware (ICLR 2025 · similarity 0.84)
- Beyond the Trial-and-Error Loop: Hybrid Projection and Automated Tuning for Distributed Training (NeurIPS 2026 · similarity 0.84)
- Accelerating Diffusion Models with Parallel Sampling: Inference at Sub-Linear Time Complexity (NeurIPS 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.