Nenhum Resultado Encontrado

A sua Pesquisa não encontrou resultados

Virtual Machines and Bare Metal (GPU)

Users can process and analyze massive data sets more efficiently, making them ideal for complex machine learning (ML), artificial intelligence (AI) algorithms, and many industrial HPC applications.

Product Features

Open all Close all

Performance

  • Bare metal GPU servers, without the hypervisor overhead, deliver uncompromising and predictable performance.

Elastic and On-Demand Compute

  • True on-demand bare metal instances deliver elasticity for your workloads. Start, terminate, and monitor bare metal instances as needed, using the console, APIs, or SDKs.

Control

  • Bare metal instance isolation and dedicated access provide unparalleled visibility and control.

Secure and Flexible Network

    Virtual Cloud Networks (VCN)

  • Fully customizable private networks enable you to easily move your existing network topology to the cloud.
  • Network Isolation

  • An isolated virtual cloud network, combined with dedicated physical servers ensure strong isolation, even from Oracle.
  • Extension of On-Premises

  • Industry-standard encrypted IPsec Virtual Private Network (VPN) connection between your corporate data center and your VCN, or use FastConnect for a secure, unmetered connection to Oracle.

High Availability

    Fault Isolation

  • You can deploy applications in mulit-region, multi- Availability Domain (AD, or multi-fault domain configurations. ADs are are built with fully independent infrastructure such as buildings, power generators, cooling equipment, network, and servers, while fault domains partition the compute resources within an AD for further protection.
  • High Availability

  • Regions can contain multiple Availability Domains, connected by a high speed network backbone, And each AD contains three fault domains.
  • Low Latency

  • High speed network backbones enable highly reliable applications running across ADs with minimum impact to application performance.

Instance Types

For Hardware Accelerated Compute Workloads

These shapes provides a balance of cores, memory and network resources.

  • Remote block volumes deliver low latency, up to 25,000 IOPS per volume, and are configurable from 50 GB to 16 TB per volume
Shape Graphics Processor Central Processor GPU OCPU Memory (GB) Storage
VM.GPU2.1 NVIDIA Tesla P100 2.0 GHz Inte® Xeon® Platinum 8167M 1 12 104 Up to 1 PB of remote Block Volumes
BM.GPU2.2 NVIDIA Tesla P100 2.0 GHz Inte® Xeon® Platinum 8167M 2 28 192
VM.GPU3.1 NVIDIA Tesla V100 2.0 GHz Inte® Xeon® Platinum 8167M 1 6 90
VM.GPU3.2 NVIDIA Tesla V100 2.0 GHz Inte® Xeon® Platinum 8167M 2 12 180
VM.GPU3.4 NVIDIA Tesla V100 2.0 GHz Inte® Xeon® Platinum 8167M 4 24 360
BM.GPU3.8 NVIDIA Tesla V100 2.0 GHz Inte® Xeon® Platinum 8167M 8 52 768

Coming Soon: GPU Bare-Metal Instances with 8 GPUs, 2TB RAM, 25TB local NVMe storage, and 8x 100GB Network ports for cluster networking and RDMA.