Parallel Sorting Algorithm Performance Profiles for Machine Learning

Encompassing Caliper performance profiles for analyzing parallel sorting algorithms. It is intended for machine learning classification of algorithm classes using the performance data. The data was authored by Michael McKinsey and last updated in October 2025.

Use Cases

• Train a classifier to predict the parallel sorting algorithm class from Caliper performance profile data.
• Analyze performance characteristics within Caliper profiles to identify efficiency patterns across different parallel sorting implementations.
• Use the performance data from Caliper profiles to benchmark and compare the runtime behavior of various parallel sorting algorithms.

Strengths

• Data is specifically formatted as Caliper performance profiles, enabling detailed performance analysis.
• The dataset is associated with a tutorial notebook providing a concrete example of data analysis methodology.
• Last update was in October 2025, indicating recent maintenance.

Limitations

• The specific number of performance profiles, rows, columns, and file size is unknown.
• Data format requires specialized tools like Thicket for analysis, creating a barrier to entry.
• The scope of algorithms and hardware configurations covered is not specified.

Provenance

Source

Texas Data Repository Harvested Dataverse

Collection Method

Performance analysis research generating Caliper profiles.

Freshness

Last updated 2025-10-15.