Rajkumar, Santosh Mohan, Goswami, Debdipta, Stockar, Stephanie · Nov 10, 2026
FAME-QP: A GPU-Accelerated Solver for Real-Time Model Predictive Control
Markdown · JSON · Original source
Authors Santosh Mohan Rajkumar, Debdipta Goswami and Stephanie Stockar present FAME-QP, a -accelerated solver for dense quadratic programs used in linear model predictive control. Model predictive control repeatedly solves an optimization problem to choose actions; the authors focus on reducing the needed for these calculations. FAME-QP uses constraint-dependent operator representations and fused iterations, replayed through CUDA graphs, and selects an execution configuration using device calibration and problem-specific timing. It also probes a cached operator bank to choose an initial penalty, then updates operators lazily when problem matrices change. The authors report that numerical and simulated humanoid and quadrotor control experiments show reduced relative to competing GPU solvers while maintaining solution accuracy and control performance. The abstract does not give the size of the latency reductions, so the reported comparison does not establish how much faster the solver is or how its performance would carry over beyond the described experiments.
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
eess.SY · Article
Source details
Original source date: Nov 10, 2026 (2026-11-10).
Date basis: Date basis not provided.
Added to Escape Meridian: 2026-11-11T23:17:11.124Z. First observed: 2026-11-11T05:03:50.603Z. Retrieved: 2026-11-11T23:16:35.230Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
The abstract reports reduced but gives no numerical values for the reductions. The described control experiments are simulated humanoid and quadrotor experiments; the abstract does not report results from physical systems.
Open question
How large were the reductions relative to the competing solvers in the reported and simulated control experiments?
Source-stated
The authors present FAME-QP as a GPU-accelerated solver for dense quadratic programs arising in linear model predictive control.
We present FAME-QP, a GPU-accelerated solver for dense quadratic programs arising in linear model predictive control.
Source-stated
FAME-QP selects its execution configuration using device calibration and problem-specific timing, and updates operators lazily when problem matrices change.
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.
Source-stated
The authors report reduced latency relative to competing GPU solvers, with solution accuracy and control performance maintained in numerical benchmarks and simulated humanoid and quadrotor control experiments.
Numerical benchmarks and simulated humanoid and quadrotor control experiments demonstrate reduced latency relative to competing GPU solvers while maintaining solution accuracy and control performance.
Source and contributor context
Rajkumar, Santosh Mohan, Goswami, Debdipta, Stockar, Stephanie · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
No suitable media was found by the supported page extraction.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Graphics processing unit
A chip that performs many calculations at once; the main hardware used to train and run AI models.
Benchmark
A defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Latency
The delay between sending a request and receiving the result.
TIME
TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.
arXiv
arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Editorial provenance
Version 1. Publication ID: c705e502-4059-40d7-87e1-a32c22f3b006.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-11T23:17:11.124Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: 3dce58b94e6155b7a5cfcb220a391c668256f231115714e9cc755e65035eafc7. Source text hash: b9fbcdbc9b455443023e8f03b3484e970cc4581fef116d591165609ea597ca40.
Revision and source-check history
Version 1 · 2026-11-11T23:17:11.124Z
FAME-QP: A GPU-Accelerated Solver for Real-Time Model Predictive Control
Authors Santosh Mohan Rajkumar, Debdipta Goswami and Stephanie Stockar present FAME-QP, a -accelerated solver for dense quadratic programs used in linear model predictive control. Model predictive control repeatedly solves an optimization problem to choose actions; the authors focus on reducing the needed for these calculations. FAME-QP uses constraint-dependent operator representations and fused iterations, replayed through CUDA graphs, and selects an execution configuration using device calibration and problem-specific timing. It also probes a cached operator bank to choose an initial penalty, then updates operators lazily when problem matrices change. The authors report that numerical and simulated humanoid and quadrotor control experiments show reduced relative to competing GPU solvers while maintaining solution accuracy and control performance. The abstract does not give the size of the latency reductions, so the reported comparison does not establish how much faster the solver is or how its performance would carry over beyond the described experiments.
Original source
eess.SY · Article
Source details
Original source date: Nov 10, 2026 (2026-11-10).
Date basis: Date basis not provided.
Added to Escape Meridian: 2026-11-11T23:17:11.124Z. First observed: 2026-11-11T05:03:50.603Z. Retrieved: 2026-11-11T23:16:35.230Z.
Limits
The abstract reports reduced but gives no numerical values for the reductions. The described control experiments are simulated humanoid and quadrotor experiments; the abstract does not report results from physical systems.
Open question
How large were the reductions relative to the competing solvers in the reported and simulated control experiments?
Source-stated
The authors present FAME-QP as a GPU-accelerated solver for dense quadratic programs arising in linear model predictive control.
We present FAME-QP, a GPU-accelerated solver for dense quadratic programs arising in linear model predictive control.
Source-stated
FAME-QP selects its execution configuration using device calibration and problem-specific timing, and updates operators lazily when problem matrices change.
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.
Source-stated
The authors report reduced latency relative to competing GPU solvers, with solution accuracy and control performance maintained in numerical benchmarks and simulated humanoid and quadrotor control experiments.
Numerical benchmarks and simulated humanoid and quadrotor control experiments demonstrate reduced latency relative to competing GPU solvers while maintaining solution accuracy and control performance.
Source and contributor context
Rajkumar, Santosh Mohan, Goswami, Debdipta, Stockar, Stephanie · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
No suitable media was found by the supported page extraction.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Graphics processing unit
A chip that performs many calculations at once; the main hardware used to train and run AI models.
Benchmark
A defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Latency
The delay between sending a request and receiving the result.
TIME
TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.
arXiv
arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Editorial provenance
Version 1. Publication ID: c705e502-4059-40d7-87e1-a32c22f3b006.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-11T23:17:11.124Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: 3dce58b94e6155b7a5cfcb220a391c668256f231115714e9cc755e65035eafc7. Source text hash: b9fbcdbc9b455443023e8f03b3484e970cc4581fef116d591165609ea597ca40.
passed · 2026-11-11T23:17:11.124Z · anthropic/claude-opus-5.5
Publication: c705e502-4059-40d7-87e1-a32c22f3b006
attribution — pass: The reading names the three listed authors and keeps the solver description and benchmark results attributed to them ('present', 'The authors report') rather than stating them as established fact.
Authors:Santosh Mohan Rajkumar, Debdipta Goswami, Stephanie Stockar
summary — pass: The reading leads with the actor and the development. It briefly explains model predictive control and why calculation time matters for real-time control, so a general reader can follow the core point: a faster GPU solver for robot control calculations. Implementation terms such as CUDA graphs and operator bank are kept exactly. They are secondary details and could use term-library tooltips. The title condenses the source title faithfully, and the question addresses a real gap in the abstract.
We present FAME-QP, a GPU-accelerated solver for dense quadratic programs arising in linear model predictive control.
date — pass: The originalPublished date, 2026-11-10, matches the v1 submission date on the arXiv page. The later feed timestamp of 2026-11-11 was correctly not used.
[Submitted on 10 Nov 2026]
support — pass: Every claim in the reading is supported by the abstract. The limitations are accurate and do not overreach: the abstract gives no latency figures, and the experiments are described as simulated. The reading also does not turn the benchmark results into a claim about real-world reliability.
Numerical benchmarks and simulated humanoid and quadrotor control experiments demonstrate reduced latency relative to competing GPU solvers while maintaining solution accuracy and control performance.
media — pass: The selected media is the verified arXiv abstract page for this paper, which is the direct original source. No other source media was found.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
A chip that performs many calculations at once; the main hardware used to train and run AI models.
Definition referenceTIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.
Explore TIMEA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceThe delay between sending a request and receiving the result.
Definition referencearXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivThe delay between sending a request and receiving the result.
Definition referenceThe delay between sending a request and receiving the result.
Definition referenceA chip that performs many calculations at once; the main hardware used to train and run AI models.
Definition referenceA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceA chip that performs many calculations at once; the main hardware used to train and run AI models.
Definition referenceTIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.
Explore TIMEA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceThe delay between sending a request and receiving the result.
Definition referencearXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivThe delay between sending a request and receiving the result.
Definition referenceThe delay between sending a request and receiving the result.
Definition referenceA chip that performs many calculations at once; the main hardware used to train and run AI models.
Definition referenceA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition reference