Adiar 1.1.0: BDD and ZDD Performance Benchmark Data

Steffan Christ Sølvsten from Aarhus University created this dataset of benchmark experiments for the Adiar library version 1.1.0. The data compares the performance of Binary Decision Diagrams (BDDs) before and after adding Zero-suppressed Decision Diagram (ZDD) modularity, and also compares Adiar's ZDDs against the CUDD library. Experiments were run on a consumer laptop named 'Alice' and the Grendel cluster.

Use Cases

• Compare BDD performance before and after a major software modularity change based on the described 'Alice' experiments.
• Benchmark ZDD implementation performance against the established CUDD library based on the described 'Grendel' experiments.
• Analyze computational performance on combinatorial benchmarks like the N-Queens and Tic-Tac-Toe problems mentioned in the description.

Strengths

• Experiments are version-controlled, referencing specific Git commits (6db6c0cdc26ae290a187d50b3fbc74175b5b315f and 5a56cd743db9ee7bb24afc87be6ec76afea5fb9b) for reproducibility.
• Benchmarks include specific problem instances: N-Queens for N=13 and N=14, and Tic-Tac-Toe for N=22 and N=23.
• Experiments were conducted on two distinct hardware setups: a consumer laptop and a computation cluster, providing context for performance results.

Limitations

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

Provenance

Source

Steffan Christ Sølvsten, Aarhus University

Collection Method

Benchmark experiments run on specific hardware (Alice laptop, Grendel cluster) using code from a linked GitHub repository.