# GPU-accelerated dynamic nonlinear optimization with ExaModels and MadNLP

Source: [GPU-accelerated dynamic nonlinear optimization with ExaModels and MadNLP](<https://www.sota2.com/research/paper/68fa5c4d1e209131ce9cfd7b>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-27T05:34:22.701Z

Authors: Fran\\c{c}ois Pacaud, Sungho Shin

Published: 2024-03-23T18:59:55.000Z

## Abstract

We investigate the potential of Graphics Processing Units (GPUs) to solve large-scale nonlinear programs with a dynamic structure. Using ExaModels, a GPU-accelerated automatic differentiation tool, and the interior-point solver MadNLP, we significantly reduce the time to solve dynamic nonlinear optimization problems. The sparse linear systems formulated in the interior-point method is solved on the GPU using a hybrid solver combining an iterative method with a sparse Cholesky factorization, which harness the newly released NVIDIA cuDSS solver. Our results on the classical distillation column instance show that despite a significant pre-processing time, the hybrid solver allows to reduce the time per iteration by a factor of 25 for the largest instance.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Quadrotor Navigation on Quadrotor Navigation Sector 180°](<https://www.sota2.com/research/sota/quadrotor-navigation-on-quadrotor-navigation-sector-180.md>): Quadrotor Navigation; Quadrotor Navigation Sector 180°; Solves/s: 4.7.

- [Quadrotor Navigation Sector 90° Performance](<https://www.sota2.com/research/sota/quadrotor-navigation-sector-90-performance.md>): Quadrotor Navigation; Quadrotor Navigation Sector 90°; Solution Rate (solves/s): 3.92.

- [Energy Management Optimization on Microgrid Energy Management N=96](<https://www.sota2.com/research/sota/energy-management-optimization-on-microgrid-energy-management-n-96.md>): Energy Management Optimization; Microgrid Energy Management N=96; Computation Time (ms): 15.3.

- [Energy Management Optimization on Microgrid Energy Management (N=192)](<https://www.sota2.com/research/sota/energy-management-optimization-on-microgrid-energy-management-n-192.md>): Energy Management Optimization; Microgrid Energy Management (N=192); Computation time (ms): 35.3.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2403.15913>)

[PDF](<https://arxiv.org/pdf/2403.15913.pdf>)