Bare Metal Service
Bare Metal Service is a cloud-based solution that offers dedicated bare metal instances with the high performance and security isolation of physical machines.
Superior Performance
Experience the power of physical machines without the overhead of virtualization. Bare Metal Service provides on-demand, high-specification instances, high-bandwidth storage, and low-latency network services, eliminating resource preemption and maximizing efficiency.
Quick Delivery
Deploy resources within minutes. Essential deliverables such as operating system setup, storage partitioning, and network services are automatically configured, with built-in monitoring and security capabilities to ensure a seamless setup.
Scalability
Easily scale instances up or down to meet changing workload demands, ensuring flexibility and cost-effectiveness.
Direct OS Installation
With Bare Metal Server, operating systems are installed directly onto physical hardware without a hypervisor, providing direct, full-access performance. Supported operating systems include CentOS, Linux, Windows, and more.
Enterprise-Grade Security
Benefit from enhanced isolation and security, reducing the risks typically associated with virtualization and safeguarding critical data with robust security measures.
AI Algorithm Training and Inference
Applications include image recognition, autonomous driving, and speech recognition, providing the computational power needed for advanced AI solutions.
Scientific Computing
Utilized in fields such as fluid dynamics, computational finance, molecular dynamics, and environmental analysis, offering powerful performance for complex simulations and data analysis.
Video Encoding and Decoding
Accelerated transcoding services for scenarios like live streaming and online education, ensuring smooth, high-quality media delivery.
Rendering
Rendering capabilities serve various scenarios, including cloud gaming and video structuring, providing seamless visual processing for intensive media applications.
H200
GPU Memory : 141GB
GPU Memory Bandwidth : 4.8TB/s
FP8 Tensor Core Performance : 4 PetaFLOPS
NVLink Bandwidth : 900G/s
Power : 700W
GB200
GPU Memory : 384GB GPU Memory Bandwidth : 16TB/s FP8 Tensor Core Performance : 20 PetaFLOPS FP4 Tensor Core Performance : 40 PetaFLOPS NVLink Bandwidth : 2×1.8TB/s Power : Up to 2700W
