Spot GPU Cloud Servers
Product Introduction
The Spot GPU Cloud Server, compared with the GPU cloud host paid by time, has ultra-low discounts, aiming to provide customers with more convenient and low-cost GPU computing power resources in some scenarios.
The list price of the Spot GPU Cloud Server will change with the market price and inventory surplus. If the background resources are insufficient or the customer’s bid is lower than the current list price, the customer will be notified by email that the resources will be reclaimed after 30 minutes.
Applicable Scenarios
Spot GPU Cloud Servers are suitable for stateless applications that support resumable computing, such as image rendering, high-performance parallel computing, and big data tasks. The higher the distribution, scalability, and fault tolerance of the application, the more suitable it is to use Spot GPU Cloud Servers to save resource costs and improve throughput.
However, services with high stability requirements are not recommended to run on Spot GPU Cloud Servers. Insufficient system resources or failed bidding may cause the Spot instances to be released and tasks to be interrupted, potentially affecting normal business operations.
Life Cycle
After a Spot GPU Cloud Server is created, it can run stably for one hour. Within this one-hour period, it will not be affected by insufficient system resources or price adjustments and can be used smoothly.
If back-end resources are insufficient or the market price exceeds the customer’s bid, the system will notify the customer via email that the Spot GPU Cloud Server will be released after 30 minutes. Please make data backups in advance.
You can enable the snapshot service at the same time when creating a Spot GPU Cloud Server. If you still need to continue using it, you can restore the GPU cloud server from a snapshot.
Billing Instructions
Currently, only the cloud server itself supports the Spot model, meaning CPU, memory, and GPU cards enjoy ultra-low discounts, while associated EIPs and disks are still sold at regular prices.
If a Spot GPU Cloud Server has overdue payments, it will immediately enter the recovery list. Customers will be notified via email to renew payment immediately. If payment is not completed within 5 minutes, the resources will be shut down and recovered.
Spot GPU Cloud Servers do not support changes to the billing method or configuration upgrades.
Spot GPU Cloud Servers are not affected by customer sales discounts and cannot enjoy stacked discount benefits.
Supported Machine Configuration & Region
Region | Machine & Card Type | Configuration |
---|---|---|
North China 2A | GPU type cost-effective graphics card 5 | 16 core 32G GPU* 1 |
North China 2A | GPU type cost-effective graphics card 5 | 16 core 64G GPU* 1 |
North China 2A | GPU type cost-effective graphics card 5 | 32 core 64G GPU* 2 |
North China 2A | GPU type cost-effective graphics card 5 | 32 core 128G GPU* 2 |
North China 1B/Shanghai 2A | GPU type V100S | 10 core 32G GPU* 1 |
North China 1B/Shanghai 2A | GPU type V100S | 10 core 64G GPU* 1 |
North China 1B/Shanghai 2A | GPU type V100S | 10 core 96G GPU* 1 |
North China 1B/Shanghai 2A | GPU type V100S | 20 core 64G GPU* 2 |
North China 1B/Shanghai 2A | GPU type V100S | 20 core 96G GPU* 2 |
North China 1B/Shanghai 2A | GPU type V100S | 20 core 128G GPU* 2 |
North China 1B/Shanghai 2A | GPU type V100S | 20 core 160G GPU* 2 |
North China 1B/Shanghai 2A | GPU type V100S | 20 core 160G GPU* 2 |
North China 1B/Shanghai 2B | GPU type P40 | 4 core 8G GPU* 1 |
North China 1B/Shanghai 2B | GPU type P40 | 4 core 16G GPU* 1 |
North China 1B/Shanghai 2B | GPU type P40 | 8 core 16G GPU* 1 |
North China 1B/Shanghai 2B | GPU type P40 | 8 core 32G GPU* 1 |
North China 1B/Shanghai 2B | GPU type P40 | 8 core 64G GPU* 1 |
North China 1B/Shanghai 2B | GPU type P40 | 8 core 16G GPU* 2 |
North China 1B/Shanghai 2B | GPU type P40 | 8 core 32G GPU* 2 |
North China 1B/Shanghai 2B | GPU type P40 | 16 core 32G GPU* 2 |
North China 1B/Shanghai 2B | GPU type P40 | 16 core 64G GPU* 2 |
North China 1B | GPU type T4S | 4 core 8G GPU* 1 |
North China 1B | GPU type T4S | 4 core 16G GPU* 1 |
North China 1B | GPU type T4S | 4 core 32G GPU* 1 |
North China 1B | GPU type T4S | 8 core 16G GPU* 1 |
North China 1B | GPU type T4S | 8 core 32G GPU* 1 |
North China 1B | GPU type T4S | 8 core 64G GPU* 1 |
North China 1B | GPU type T4S | 8 core 16G GPU* 2 |
North China 1B | GPU type T4S | 8 core 32G GPU* 2 |
North China 1B | GPU type T4S | 8 core 64G GPU* 2 |
North China 1B | GPU type T4S | 8 core 128G GPU* 2 |