Data

DCC Version 2

Release status: Coming soon (October 2026).

Dataset documentation and access procedures will be provided here when the release is available.

Publication

Piotr Luszczek, Daniel Burrill, William Bergeron, Vijay Gadepally, Matthew Hubbell, et al. “The MIT AI Systems Dataset: Environmental, Compute, and Workload Traces.” IEEE High Performance Extreme Computing Conference (HPEC), September 2026.

HPEC 2026 conference program

Data use agreement

Access to Version 2 will require a data use agreement. The agreement will be available through DocuSign. Signing is not yet available.

Data use agreement (not yet available)

DCC Version 1: Original release

The MIT Supercloud Dataset is available from the Amazon Open Data Registry. The January 2022 release is stored at the following S3 location:

s3://mit-supercloud-dataset/datacenter-challenge/202201/

Storage requirement: The dataset is approximately 2 TB. Ensure that sufficient storage is available before downloading the complete dataset.

Citation

Please cite the following paper when using or referencing this dataset:

Samsi, Siddharth, Weiss, Matthew, Bestor, David, et al. “The MIT Supercloud Dataset.” 2021 IEEE High Performance Extreme Computing Conference (HPEC). IEEE, 2021.

arXiv:2108.02037 | IEEE Xplore

Download instructions

The commands below use the AWS Command Line Interface. See the AWS CLI installation instructions before proceeding. Public access uses the --no-sign-request option.

1. List the January 2022 release

aws s3 ls s3://mit-supercloud-dataset/datacenter-challenge/202201/ --no-sign-request

2. Review the directory contents

The original January 2022 directory listing is reproduced below. CPU and GPU time series are stored in the corresponding folders and organized into subfolders.

                           PRE cpu/
                           PRE gpu/
2022-01-19 14:11:40         285 LICENSE
2022-01-19 14:11:40        7079 README.md
2022-01-20 12:20:09         339 labelled_job_stats.csv
2022-01-20 12:20:10       82332 labelled_jobids.csv
2022-01-20 12:20:10  2274303873 node-data.csv
2022-01-20 12:20:11   103165427 slurm-log.csv
2022-01-20 12:20:11         111 tres-mapping.txt

3. Download the data

The original command below recursively downloads the Datacenter Challenge collection into a local datacenter-challenge directory:

aws s3 cp s3://mit-supercloud-dataset/datacenter-challenge datacenter-challenge --recursive --no-sign-request

The 202201/ folder contains data released as of January 2022. Additional releases may appear in date-specific folders as data collection continues.

Documentation and contact

For collection methodology and full dataset details, refer to the publication above. Questions may be directed to mit-dcc@mit.edu.

Acknowledgement:
Research was sponsored by the United States Air Force Research Laboratory and the United States Air Force Artificial Intelligence Accelerator and was accomplished under Cooperative Agreement Number FA8750-19-2-1000. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the United States Air Force or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes not withstanding any copyright notation herein.