Dear HPC Users, 👋
We are thrilled to share that our HPC ecosystem is growing! 🎉 🚀 We are currently welcoming large-scale users to our platform, which speaks volumes about the capability and reputation of our resources.
🔍 The Challenge
With this growth, we have experienced a surge in high-volume GPU requests. We understand that this has unfortunately led to queue interruptions and significantly longer wait times for users requesting smaller GPU allocations. ⏳
Our priority is to maintain a fair, efficient, and reliable environment for all researchers and developers in our community. ⚖️
🛠️ Temporary Policy Change
To restore balance and ensure everyone can progress with their work, we are implementing a temporary constraint on GPU allocations.
- The Rule: Effective this coming Saturday (23 May 2025), a maximum limit of 2 GPUs per job 🛑 will be enforced for all users.
- The Goal: This cap ensures that resources are freed up rapidly. Given our maximum job execution duration is 3 days, this change guarantees that a larger number of users will successfully receive their requested resources within a 3-day window at most. 🗓️✅
💡 Note: This is a short-term measure
We sincerely appreciate your understanding, patience, and cooperation as we navigate this growth and work to keep the cluster running smoothly for everyone.