Price list
Dear all,
Following the "Announcement on the Charge of HPC-AI-integrated Intelligent Computing Center" made on January 29, 2024, the HPC-AI-integrated Intelligent Computing Center has dedicated itself to providing high-performance computing resources at the campus level with the considerable co-construction and support from staff, faculty and students. In line with our mechanism of biannual charge adjustment, the University has fine-tuned the charging scheme, considering resource utilization, feedback from internal users, and practices at domestic universities. The adjustment is not only to better accommodate the demands of research, teaching and learning, but also to ensure the more effective and efficient use. Details are as follows:
1.The cost for accessing computing resources in the queue has not been altered, and is listed below:
| Resource Type | Shared (Official) | Exclusive (Official) | Emergency (Official) | Debug Pricing |
|---|---|---|---|---|
| CPU | CNY 0.02/core/hour | CNY 0.04/core/hour | CNY 0.06/core/hour | Free |
| GPU (A800) | CNY 1.60/GPU/hour | CNY 2.50/GPU/hour | CNY 4.00/GPU/hour | Free |
| GPU (A30/A40/Ascend) | CNY 0.50/GPU/hour | CNY 1.00/GPU/hour | CNY 1.50/GPU/hour | Free |
Notes:
- For GPU resources with MIG mode enabled, the GPU cost for a job will be calculated based on the requested memory capacity.
- The runtime of a job/container is calculated based on the actual start and end times. If the runtime is less than one hour, it will be calculated in seconds.
- "Shared" corresponds to low priority, "Exclusive" corresponds to medium priority, and "Emergency" corresponds to high priority.
- Based on the pricing, specify the relevant queues to use:
CPU Shared: (i64m512u, i64m512r, a128m512u, long_cpu, i96m3tu)
CPU Exclusive: (i64m512ue, i64m512re, a128m512ue, i96m3tue)
CPU Emergency: (emergency_cpu)
GPU (A800) Shared: (i64m1tga800u, long_gpu)
GPU (A800) Exclusive: (i64m1tga800ue)
GPU (A800) Emergency: (emergency_gpu)
GPU (A40) Shared: (i64m1tga40u)
GPU (A40) Exclusive: (i64m1tga40ue)
GPU (A40) Emergency: (emergency_gpua40)
2.The charging scheme for machine/bare-metal server leasing has been updated to a "per unit per day" billing model. Details are as follows:
| Resources Type | Pricing |
|---|---|
| CPU Server (Intel 64 cores) | CNY 62.00/unit/day |
| CPU Server (AMD 128 cores) | CNY 123.00/unit/day |
| Big CPU Memory Server | CNY 62.00/unit/day |
| GPU Server (8 A800) | CNY 480.00/unit/day |
| GPU Server (8 A30/A40/Ascend) | CNY 192.00/unit/day |
| Notes: | For leasing periods less than 1 day, the calculation will be based on 1 day. |
*Computing resource pricing effective date: September 1.
3.There is the new storage resource charging scheme:
As the current occupancy rate of storage resources has exceeded 70%, the Intelligent Computing Center will introduce the following new charge mechanism for storage resources to optimize the allocation of storage resources.
| Storage Type | Free Capacity | Pricing Beyond Free Capacity |
|---|---|---|
| HDD | 1TB/user | CNY 0.50/1TB/day |
| SSD | 200GB/user | CNY 0.50/100GB/day |
| Notes: | 1. Storage usage is collected from 0:00 to 1:00 daily. 2. Beyond the free capacity, HDD storage usage below 1TB is charged as 1TB, and SSD storage usage below 100GB is charged as 100 GB. |
* Storage resource pricing effective date: October 1.
* In the future, further measures of storage capacity management will be considered based on the usage of storage resources.
To strengthen the co-construction, sharing, and co-management of the HPC-AI-integrated Intelligent Computing Center, all users are advised to:
1.Optimize their job programs and closely monitor the CPU and GPU utilization during the execution of jobs, avoiding the waste of computing resources caused by low utilization rate.
2.Clean and migrate storage space in time to release unnecessary storage space occupation.
3.Jointly maintain the charging mechanism and pay the bill on time, since the Intelligent Computing Center regularly sends bills and payment reminders to users, so that they can timely catch up on the job running status and the expenditure details.
If you have any questions, or suggestions, or seek technical support, please feel free to contact us at hpc@hkust-gz.edu.cn.
Information Technology Department (ITD)
August 15, 2024