ENERGETIC Project: Accelerator vs. CPU Energy Savings Analysis

The ENERGETIC project data examines energy savings from using accelerators compared to CPU-only computations. The dataset, created by Michael Bane and last updated in May 2026, likely contains results from benchmarking different code classes on heterogeneous architectures. The project also investigated standardized methods for measuring energy consumption across different hardware platforms.

Use Cases

• Benchmarking energy consumption of different hardware accelerators based on the project's comparative analysis.
• Identifying code classes suitable for heterogeneous architectures based on the project's findings.
• Developing standardized energy measurement methodologies for portable performance analysis.
• Informing hardware selection for energy-constrained AI/ML workloads based on accelerator performance data.

Strengths

• Dataset size is 104.3 MB, indicating a medium-scale collection of results.
• Last update timestamp is 2026-05-28 17:27:06, suggesting recent maintenance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Michael Bane via figshare

Freshness

Last updated 2026-05-28 17:27:06