SDRBench — NWChem

Two-electron repulsion integrals over Gaussian-type orbital basis sets, computed with libint (Valeev group, Virginia Tech), an integral engine of NWChemEx; 1D arrays.

Raw binary arrays, not an Arrow/Parquet dataset. datasets.load_dataset("sdrbench/nwchem") returns the table of files (files.jsonl: path, dtype, shape, sha256). Read the arrays with the sdrbench package or hf_hub_download + numpy as shown below.

This repository is an unmodified mirror of the NWChem data of SDRBench, the Scientific Data Reduction Benchmark. The originals are hosted by Argonne National Laboratory on Globus; every archive was unpacked and its files uploaded byte-for-byte (sha256 verified) (revision f657d8be5a). metadata/ holds the original SDRBench property and template files. Kept in sync automatically by szcompressor/sdrbench.

• Data provider: libint (https://github.com/evaleev/libint), Valeev research group, Virginia Tech
• Size: 19.4 GB in 2 variant(s)
• Format: raw little-endian binary, C order (slowest dimension first); dtype and shape per field below

Usage

pip install "sdrbench[sz3]"     # sdrbench + pysz (SZ3)

import numpy as np
import sdrbench
from pysz import sz, szConfig, szErrorBoundMode

ds = sdrbench.dataset("nwchem")                # default variant "default"
print(ds.fields)
x = ds["631"]                     # numpy array, dtype <f8, shape (102953248,)

# compress with SZ3 (pysz) at a 1e-3 value-range-relative error bound
conf = szConfig()
conf.errorBoundMode = szErrorBoundMode.REL
conf.relErrorBound = 1e-3
compressed, ratio = sz.compress(np.ascontiguousarray(x), conf)
y, _ = sz.decompress(compressed, x.dtype.type, x.shape)
max_err, psnr, nrmse = sz.verify(np.asarray(x), y)
print(f"ratio {ratio:.1f}x, PSNR {psnr:.1f} dB, max error {max_err:.3g}")

Files are fetched on first use (in parallel with ds.download("data/")) and cached, pinned to the revision of the installed sdrbench release; if Hugging Face is unreachable the package falls back to the original archive on Globus. Existing local SDRBench copies can be used with sdrbench.dataset("nwchem", root="/path").

Without the package, with huggingface_hub and numpy only:

from huggingface_hub import hf_hub_download
import numpy as np
p = hf_hub_download("sdrbench/nwchem", "nwchem/631-tst.bin.d64", repo_type="dataset", revision="f657d8be5aa99a0e0147f0f974c3e040fd0e8cbe")
x = np.fromfile(p, dtype="<f8").reshape((102953248,))

Variants and fields

default (default)

sdrbench.dataset("nwchem", "default") — folder nwchem/ — 6 files, 12.9 GB, from SDRBENCH-NWChem-dataset.tar.gz.

Three integral sets (631, acd, ccd), double precision, plus a reader (readbin.cpp).

Field	dtype	Shape (C order)	File
631	<f8	102953248	nwchem/631-tst.bin.d64
acd	<f8	801098891	nwchem/acd-tst.bin.d64
ccd	<f8	712996037	nwchem/ccd-tst.bin.d64

Other files: nwchem/SDRBENCH-NWChem-dataset-property.txt, nwchem/readbin.cpp, nwchem/template_data.txt

f32

sdrbench.dataset("nwchem", "f32") — folder nwchem/ — 3 files, 6.5 GB, from SDRBENCH-NWChem-dataset.tar.gz.

The same integrals as single-precision copies.

Field	dtype	Shape (C order)	File
631	<f4	102953248	nwchem/631-tst.bin.f32
acd	<f4	801098891	nwchem/acd-tst.bin.f32
ccd	<f4	712996037	nwchem/ccd-tst.bin.f32

Citation

Please cite SDRBench; sdrbench cite nwchem prints it:

@inproceedings{zhao2020sdrbench,
  title     = {{SDRBench}: Scientific Data Reduction Benchmark for Lossy Compressors},
  author    = {Zhao, Kai and Di, Sheng and Liang, Xin and Li, Sihuan and Tao, Dingwen and Chen, Zizhong and Cappello, Franck},
  booktitle = {2020 IEEE International Conference on Big Data (Big Data)},
  pages     = {2716--2724},
  year      = {2020}
}

License and terms

The data are distributed by the SDRBench team for research use under the terms of the original data providers listed above (see the SDRBench website); please acknowledge them as requested. Rights remain with the data providers.