Overview
Service Overview
GPU Server is a virtualized computing service that lets you freely allocate and use infrastructure resources such as CPU, GPU, and memory provided by the server, without having to purchase them individually, and allocate as much as needed at the required time. It is suitable for tasks that require fast computation speed, such as AI model experimentation, prediction, and inference in a cloud environment, and you can flexibly select and use resources with optimized performance based on the type and scale of the work. The GPU Server provides the following features.
Provided Features
- GPU Server Management: Through a web-based console, users can directly Self Service create, delete, and modify GPU Server provisioning, monitoring, and billing.
- Product offerings by GPU quantity: Depending on the project’s purpose and scale, you can freely select the number of H100/A100 GPUs to configure a virtual server.
- High‑Performance GPU Provision: We provide a high‑performance GPU server at physical‑server level using a pass‑through method.
- Storage Connection: Provides additional attached storage besides the OS disk. * You can connect and use Block Storage, File Storage, and Object Storage.
- Strong Security Application: Use the Security Group service to control inbound/outbound traffic exchanged with the external internet or other VPC(Virtual Private Cloud), securely protecting the server.
- Monitoring: You can view monitoring information such as the status of computing resources—including CPU, Memory, Disk, and GPU—through the Cloud Monitoring service.
- Network Configuration Management: The server’s subnet/IP can be easily changed from the values set at initial creation. * NAT IP provides a management feature that lets you enable or disable it as needed.
- Key Pair method: To ensure a secure OS access method, we provide a Key Pair method instead of ID/PW login.
- Image Management: You can create and manage Custom Images, and it provides sharing functionality between projects.
- ServiceWatch Service Integration Offering: You can monitor data through the ServiceWatch service.
Component
GPU Server provides GPUs, NVSwitch, and NVLink on top of virtualized computing resources.
- NVSwitch can only be enabled and used for instance types that allocate eight GPUs on a single GPU server.
Specifications by GPU Type
GPU (Graphic Processing Unit) performs the calculations needed to create images that compose the computer screen, and because it is specialized for parallel processing, it can handle large amounts of data quickly, processing large-scale parallel operations such as artificial intelligence (AI) and data analysis.
The following are the specifications of the GPU Types offered by the GPU Server service.
| Category | A100 Type | H100 Type | B300 Type |
|---|---|---|---|
| GPU Architecture | NVIDIA Ampere | NVIDIA Hopper | NVIDIA Blackwell Ultra |
| GPU Memory | 80 GiB | 80 GiB | 268 GiB |
| GPU Transistors | 54 billion 7N TSMC | 80 billion 4N TSMC | 208 billion 4NP TSMC |
| FP16 Tensor Core (Dense) | 312 TFLOPs | 989 TFLOPs | 2.25 PFLOPs |
| FP8 Tensor Core (Dense) | Not supported | 1,979 TFLOPs | 4.5 PFLOPs |
| FP4 Tensor Core (Dense) | Unsupported | Not supported | 13.5 PFLOPs |
| GPU Memory Bandwidth | 2,039 GB/s HBM2e | 3,352 GB/s HBM3 | 8 TB/s HBM3e |
| NVLink performance | NVLink 3 | NVLink 4 | NVLink 5 |
| NVLink Signaling Rate | 25 GB/s (x12) | 25 GB/s (x18) | 50 GB/s (x18) |
| NVSwitch GPU-to-GPU bandwidth | 600 GB/s | 900 GB/s | 1.8 TB/s |
| Total NVSwitch aggregate bandwidth | 4.8 TB/s | 7.2 TB/s | 14.4 TB/s |
Server type
The server types offered by the GPU Server are as follows. For detailed information about the server types provided by GPU Server, refer to GPU Server 서버 타입.
| Category | Server type | CPU vCore | Memory(GB) | Number of GPUs |
|---|---|---|---|---|
| GPU-A100-1 | g1v16a1 | 16 | 234 | 1 |
| GPU-A100-1 | g1v32a2 | 32 | 468 | 2 |
| GPU-A100-1 | g1v64a4 | 64 | 936 | 4 |
| GPU-A100-1 | g1v128a8 | 128 | 1,872 | 8 |
| GPU-H100-2 | g2v12h1 | 12 | 234 | 1 |
| GPU-H100-2 | g2v24h2 | 24 | 468 | 2 |
| GPU-H100-2 | g2v48h4 | 48 | 936 | 4 |
| GPU-H100-2 | g2v96h8 | 96 | 1,872 | 8 |
| GPU-B300-3 | g3v16b1 | 16 | 480 | 1 |
| GPU-B300-3 | g3v32b2 | 32 | 960 | 2 |
| GPU-B300-3 | g3v64b4 | 64 | 1,920 | 4 |
| GPU-B300-3 | g3v128b8 | 128 | 3,840 | 8 |
OS and GPU driver version
The operating systems (OS) supported by the GPU Server are as follows. Note that GPUs of type B300 are supported only from a specific GPU version onward, so please be careful when selecting images.
| OS | OS version | GPU driver version | Server type classification |
|---|---|---|---|
| Ubuntu | 24.04 | 580.126.20 | GPU-B300-3, GPU-H100-2, GPU-A100-1 |
| Ubuntu | 24.04 | 570.195.03 | GPU-H100-2, GPU-A100-1 |
| Ubuntu | 22.04 | 535.183.06 | GPU-H100-2, GPU-A100-1 |
| RHEL | 9.6 | 580.126.20 | GPU-B300-3, GPU-H100-2, GPU-A100-1 |
| RHEL | 8.10 | 580.126.20 | GPU-B300-3, GPU-H100-2, GPU-A100-1 |
| RHEL | 8.10 | 535.183.06 | GPU-H100-2, GPU-A100-1 |
Preceding Service
This is a service that must be installed in advance before creating this service. Please prepare by referring to the user guide provided in advance.
| Service Category | service | Detailed description |
|---|---|---|
| Networking | VPC | A service that provides an isolated virtual network in a cloud environment |
| Networking | Security Group | Virtual firewall that controls server traffic |