We present FAME-QP, a GPU-accelerated solver for dense quadratic programs arising in linear model predictive control. FAME-QP combines constraint-dependent operator representations with fused single- and two-kernel iterations replayed through CUDA graphs, automatically selecting the execution configuration through device calibration and problem-specific timing. It probes a cached operator bank to select the initial penalty and updates operators lazily when problem matrices change. Numerical benchmarks and simulated humanoid and quadrotor control experiments demonstrate reduced latency relative to competing GPU solvers while maintaining solution accuracy and control performance.
Figures & tables
Fig. 1: Cold-solve times on dense random QPs at matched accuracy, excluding setup. Lines show medians and bands min–max ranges across five instances, each timed by its median over repeated cold solves. All solvers use float64 except cuOSQP ( float32 ).
Solver / n
400
800
1200
1600
2000
FAME-QP
0.658±0.066
1.19±0.04
2.05±0.45
2.92±0.09
4.23±0.21
ReLU-QP
2.10±0.37
6.06±1.02
11.6±1.9
20.0±3.9
30.9±6.1
CuPIQP
23.7±7.0
58.9±11.3
178±42
248±63
449±120
CUDA MPC
22.7±0.1
59.9±0.5
82.8±0.8
103±0
138±3
π MPC
9.06±12.39
8.90±12.88
6.78±5.67
15.4±23.9
13.4±18.1
TABLE I: Online MPC step times [ms], mean ± standard deviation over 50 steps. Setup and initialization are excluded; fastest means are bold.
N ( n )
QP time [ms]
CoM RMS [mm]
FAME-QP
ReLU-QP
FAME-QP
ReLU-QP
10 (290)
0.650±0.091
1.47±0.05
18.7
18.7
15 (435)
0.763±0.116
2.26±0.12
19.1
19.2
20 (580)
1.04±0.20
3.25±0.13
20.5
19.5
TABLE II: Unitree G1 push recovery. QP times are mean ± std, excluding the first step.
Fig. 2: Unitree G1 push recovery in MuJoCo with FAME-QP : torso push, maximum lean, return, and final stance.
Fig. 3: Skydio X2 trajectory tracking in MuJoCo with FAME-QP .
Execution time per iteration [ μ s]
n
Unfused
Unfused
Fused
Fused
eager
graph
eager
graph
128
140
24.6
21.1
4.37
512
146
33.4
21.8
17.4
1028
145
72.9
54.4
53.4
Starting-penalty probing, common bank, ϵ=10−5
TABLE III: Execution and penalty ablations. Execution: median device time per iteration over 100 repetitions. Probing: mean iteration counts and mean per-instance median wall times over five test instances, with 20 cold solves per instance. Setup is excluded.
Quadratic programming optimizes quadratic objectives under constraints, supporting applications from robotics and control to finance and machine learning.
This paper introduces numerical optimizations for maximizing throughput on GPU when solving large batches (10,000 to over 100,000) of sequential quadratic programming (SQP) iterations, where all problems have the same structure. The optimizations are implemented in a toolbox WarpMPC for model-predictive control (MPC) in JAX and Warp. Based on the insight that all MPC problem instances in a batch share the same sparsity in time, cost, and constraints, we propose unrolling sparse linear factorizations and solves, which dominate alternating direction method of multipliers (ADMM) solver runtime. We avoid memory access bottlenecks and wasting computations via optimized memory layout, padding-reducing segmentation of the unrolled factorization, and dependency level scheduled backsolves, additionally accelerating sensitivity computation. We achieve throughputs of 8,000 to 250,000 SQP iterations per second on nonlinear cartpole, quadrotor, and humanoid robot benchmarks, outperforming baselines by 3× to 25×. We illustrate practical usefulness by synthesizing a dataset and training a neural network approximation of an MPC in under 5 minutes that stabilizes a nano quadrotor in hardware experiments.
Henrik Hose, Se Hwan Jeon, Charles Khazoom +2
Institute for Data Science in Mechanical Engineering (DSME), RWTH Aachen University, Germany · Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge MA, USA.
Quadratic programs (QPs) are fundamental building blocks of many robotic tasks such as inverse kinematics, whole-body control, and model predictive control. Extending these methods to large-scale robot learning and control requires solving many QPs concurrently and, for end-to-end learning, differentiating through their solutions. Efficient batched optimization is therefore essential to bringing established model-based methods into GPU-based robotics pipelines. We present cuPIQP, a batched and differentiable QP solver on GPUs based on a proximal interior-point method, with unified support for dense, sparse, and multistage problems. The solver exploits parallelism across independent QPs and their internal structure while preserving per-problem numerical updates and termination decisions. Benchmarks with up to 16384 QPs per batch demonstrate solve-time speedups of up to 18.8× over Moreau and 4.26× over eight-core PIQP. In an end-to-end learning task, cuPIQP accelerates training by 4.6× over QPAX and 7.8× over Moreau. By reducing the computational cost, cuPIQP provides efficient, reusable infrastructure for integrating model-based optimization into large-scale robot learning and control.
Fenglong Song, Roland Schwan, Yuwen Chen +2
Automatic Control Laboratory, EPFL, Switzerland. · Accessible and Accelerated Robotics Lab, Dartmouth College, USA. · Institute for Transport Planning and Systems, ETH Zürich, Switzerland.
Model Predictive Control (MPC) delivers constraint-aware control, but its reliance on online optimization limits its use on systems with fast dynamics, high-dimensional models, or long horizons. Existing GPU implementations typically treat the device as a linear-algebra accelerator, leaving the optimization loop dependent on repeated kernel launches and high-latency memory transfers. This paper introduces CUDA MPC, a GPU-native MPC framework that co-designs the optimization algorithm, execution model, and memory architecture for CUDA hardware. CUDA MPC pairs a parallel-in-horizon alternating direction method of multipliers (ADMM) splitting with a fused CUDA kernel that runs the entire iterative solve on the device. Intermediate optimization variables stay in low-latency, on-chip shared memory, and a localized atomic-flag protocol synchronizes only adjacent horizon blocks, minimizing host intervention, kernel-dispatch overhead, and global-memory traffic. Across six nonlinear robotics benchmarks spanning increasing state dimension and constraint density, CUDA MPC sustains real-time rates at horizons one to two orders of magnitude longer than CPU solvers: it solves an optimization-based collision-avoidance parking problem with 100 s of lookahead within a 0.1 s sampling interval, and is the only solver evaluated that achieves both real-time execution and collision-free coordination for a centralized 10-agent swarm, where acados and CasADi return no feasible solution and require 3.5 s and 4.5 s per solve. Against tensor-framework implementations of the same ADMM splitting, the fused kernel is up to 965× faster.
Babak Akbari, Melissa Greeff
Robora Lab, Department of Electrical and Computer Engineering Queen's University, Kingston, ON, Canada