OpticalNeuromorphicEikonal_gradient_256: Pathfinding Benchmark on a 256x256 Grid

A 65,536-cell grid dataset for benchmarking GPU-accelerated navigation algorithms. It was created by Francisco Angulo de Lafuente to evaluate the Optical Neuromorphic Eikonal Solver, which reportedly achieves a 135x average speedup over CPU Dijkstra. The dataset models a 2D pathfinding problem with 20% obstacle density and a gradient speed field.

Use Cases

• Benchmarking pathfinding algorithm speed and accuracy based on the described 30-300x speedup and sub-1% error metrics.
• Evaluating navigation on structured grids with obstacles based on the 20% obstacle density and grid coordinates.
• Testing algorithms with variable movement costs based on the gradient speed field described.
• Comparing solver performance against reported near-optimal path lengths (1.025x optimal).

Strengths

• Grid size is explicitly defined as 256x256, containing 65,536 cells.
• Obstacle density is specified as 20%, providing a known challenge level.
• Benchmark performance metrics for the associated solver are provided, including a 135x average speedup and 0.64% mean absolute error.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Francisco Angulo de Lafuente, via OpenML and Hugging Face Datasets.

Collection Method

Synthetically generated for algorithm benchmarking, using a specified random seed (5).

Time Range

Temporal coverage is not specified.

Freshness

Last updated date is unknown.

Geography

Spatial coverage is not applicable; the dataset represents an abstract 2D grid.