GPU Server Usage Policy

Institute GPU Server 192.168.3.126 (2 × NVIDIA A100)

IIIT Delhi | Version 2.0 | June 2026

1. Purpose and Scope

This policy governs the use of the shared department GPU server. It applies to every user who has been granted an account. The objective is fair, transparent and productive sharing of GPU resources.

2. Access and Accounts

  • Accounts are issued only for academic and research purposes.
  • Users are responsible for all activity under their account.
  • Account sharing is strictly prohibited.
  • GPU access must be through the approved quota management system.

3. Fair Use Budget

Policy Value
Default Weekly Quota 48 GPU Hours
Reset Every Monday 00:00
Carry Forward Not Allowed
Billing Wall Clock Time × Number of GPUs

4. User Tiers

User Weekly Quota Priority
Faculty 100 Hours Highest
PhD / Research Assistant 100 Hours High
M.Tech Project 50 Hours Medium
B.Tech / Coursework As Assigned Standard

5. GPU Scheduling Rules

  • Each job receives one or more dedicated GPUs.
  • No two jobs share a GPU.
  • Jobs wait in queue when GPUs are occupied.
  • Priority users are served first.
  • Over-budget jobs may run only on idle GPUs.
  • Over-budget jobs may be preempted after a 60-second grace period.
  • The newest over-budget job is evicted first.
  • Budgets reset every Monday.

6. Responsible Usage

  • Release GPUs immediately after use.
  • Request only the GPUs required.
  • Save checkpoints for long-running jobs.
  • Use tmux or screen for remote sessions.
  • Store files in your home directory.
  • Maintain backups of important data.

7. Transparency

GPU allocation, utilization and ownership are visible to all users. Only resource usage information is public. Your source code and data remain private.

8. Prohibited Activities

  • Bypassing the GPU quota system.
  • Obtaining unauthorized administrative privileges.
  • Cryptocurrency mining.
  • Commercial or non-academic workloads.
  • Keeping GPUs idle intentionally.
  • Sharing login credentials.
  • Illegal or prohibited activities.

9. Monitoring and Enforcement

GPU jobs are logged, monitored and audited. Misuse may result in reduced quota, temporary suspension or permanent loss of access.

10. Support and Exceptions

  • Additional GPU hours may be requested with Faculty approval.
  • User tiers are managed by administrators.
  • Contact: helpdesk@iiitd.ac.in
Agreement

By using the GPU Server you agree to comply with this policy.