---
pretty_name: HPC-Bench Input Data
license: other
license_name: upstream-benchmark-licenses
tags:
  - hpc
  - benchmark
  - code-optimization
  - llm-evaluation
  - openmp
  - cuda
size_categories:
  - n<1K
---

# HPC-Bench Input Data

Input datasets for **HPC-Bench: A Comprehensive Benchmark for High Performance Computing Codes** (NeurIPS 2026 Evaluations and Datasets Track).

Code, benchmarks, and evaluation pipeline: https://github.com/Deep-Learning-Profiling-Tools/HPC-Bench

## Contents

One zip archive per (setting, benchmark), each containing the benchmark's `input_data/` directory with all five standardized input scales (mini, small, medium, large, extra-large).

| Path | Setting | Archives |
|:---|:---|---:|
| `input_data/EX1/<benchmark>.zip` | Serial CPU optimization | 92 |
| `input_data/EX2/<benchmark>.zip` | OpenMP CPU parallelization | 92 |
| `input_data/EX3/<benchmark>.zip` | CUDA GPU optimization | 23 |

`manifest.csv` lists every archive with its size and SHA-256 checksum. The EX3 `nn` benchmark generates its input at run time and has no archive.

## Download

Do not download these files by hand. From a clone of the GitHub repository, one command downloads the archives, verifies their checksums, and extracts each one into `EX*/<benchmark>/input_data/`:

```bash
python tools/fetch_input_data.py                      # everything (~38 GB zipped)
python tools/fetch_input_data.py --ex EX1             # one setting
python tools/fetch_input_data.py --benchmarks 2mm,bfs # selected benchmarks
```

## License

The input data are derived from the upstream benchmark suites (PolyBench, Rodinia, Parboil, MachSuite, MiBench, PARSEC, and Codee) and remain subject to their respective licenses.