Resources

Main Server
Model : HPE ProLiant DL380 Gen9
Processor :
2x Intel(R) Xeon(R) CPU E5-2630 v4

(10 core, 2.20 GHz, 25MB, 85W)
Graphical Processing Unit : NVIDIA Tesla P100
3584 CUDA cores
Memory : 212GB
Storage : RAID 5 configuration
7.37TB 10000rpm SAS
Operating System : Ubuntu Server 16.04
Frameworks :
  • Tensorflow
  • Keras
Backup Server
Model : HPE ProLiant ML350 Gen9
Processor :
Intel(R) Xeon(R) CPU E5-2620 v4

(8 core, 2.10 GHz, 20MB, 85W)
Graphical Processing Unit : NVIDIA Tesla P4
2560 CUDA cores
Memory : 157GB
Storage : 25.57TB 7200rpm SAS
Operating System : Ubuntu Server 16.04
Cloud Computing
Cloud
GPUs
NVIDIA Tesla P40 (to be installed in main server)

HPC Usage Code of Conduct

  1. Users are required to fill out the server usage request form at the following link:
    🔗 https://s.id/request-server-AIRDC
  2. Each submission is valid for one specific job only (e.g., model training, dataset inference), not an entire project.
  3. Scheduling is conducted on business days by the AIRDC team, considering factors such as queue length, project category, and job size. The current queue status can be viewed here:
    🔗 https://s.id/server-queue-AIRDC
  4. Users will receive a confirmation email containing SSH access instructions one business day after scheduling is completed.
  5. The maximum usage duration for each job is 2 days (48 hours).
  6. If there is no response within 10 business days after the usage period ends, AIRDC will send a follow-up confirmation email. If the job is still ongoing, users must resubmit the form within 1 day to rejoin the queue.
    Without confirmation, all user data will be deleted from the server.

Server Specifications:

2 × Intel® Xeon® CPU E5-2630 v4 @ 2.20GHz
1 × NVIDIA Tesla P100, CUDA 10.1
212 GB RAM

Additional Notes:

  1. Server access is supported only via Command Line Interface (CLI) through SSH. Graphical User Interface (GUI) is not supported.
  2. Users are required to test their code beforehand on a local machine or Google Colab.
  3. The server environment is Docker-based (containerized), isolated, and customizable by the user. The maximum storage allocation per user is 150 GB.

Acknowledgement:

It is mandatory to include the attribution “NVIDIA – BINUS AI R&D Center” in every scientific publication (journal and conference) resulting from the use of this server.

Example:

Contact:

For further inquiries, please contact: