Operating Steps

1. Log in to the main interface of the BCC Management Console.
2. After successful login, select the “Product Service > Baidu Cloud Compute (BCC), you can enter the “Instance List” page.

3. Click “Create Instance”, select configuration information, you can set up the name of instance and the password of administrator, and then select the duration and quantity of purchase.

   Configuration Information	Instruction
   Mode of Payment	Pre-payment (yearly payment or monthly payment); post-payment (pay on demand)
   Current Regions	North China-Beijing, South China-Guangzhou, North China-Baoding, East China-Suzhou
   Availability zone	The availability zone means the same zone where electricity and network are independent of each other. The failure is isolated within an available zone.
   Type	GPU Instance
   GPU Type	NVIDIA TeslaV100, P40, P4, K40 and NVIDIA Deep Learning Development Card
   Quantity of GPU	1, 2, 4 and 8 blocks
   Operating System	Ubuntu 16.04 LTS amd64 (64bit)-CUDA9.2, BCC public image and others
   Type of Image	Provide Ubuntu16.04 GPU special image, common public image and custom image with integrated GPU driver and CUDA development environment.
   Local Disk	Offers the local disk with fixed capacity (currently without self-selection)
   Type of Network	The virtual private network belongs to the current server, if it is in the default case，the system defaults to a private network.

4. Click the “Next Step” to confirm the order and pay.
5. After successful purchase, view the page of “Server Instance List”.

Manage GPU Instances

You can start, stop and restart the GPU, modify the name of GPU, bind the public network IP and operate other features in the current account.

1. Log in to the main interface of the BCC Management Console.
2. After successful login, select “Product Service > Baidu Cloud Compute (BCC)” . The “Instance List” page is displayed.
3. View the details of current Baidu Cloud Compute(BBC).
4. (Optional) if you bind the public network, click the public network IP icon in the current GPU instance, and then select the Elastic IP to be bound in the Pop-up. If public network IP isn’t created before, please purchase first, see Creation of EIP Process.