Retired GPU-accelerated instance types

• Instance family overview:

  • Powered by the third-generation X-Dragon architecture and chip-level fast-path acceleration, this instance family delivers stable, predictable, ultra-high performance in compute, storage, and networking.

  • These instances use NVIDIA T4 GPUs to accelerate graphics and AI workloads. Combined with container technology, they can host over 60 virtual Android terminals, providing hardware-accelerated video transcoding for each display.

• Use cases:

  • Providing remote app services based on Android, such as always-on cloud applications, cloud gaming, cloud phones, and Android-based web crawlers.

• Compute:

  • vCPU-to-memory ratio of approximately 1:3.

  • Processor: 2.8 GHz Ampere^® Altra^® processor with a turbo frequency of 3.0 GHz. The native Arm compute platform provides efficient performance and excellent app compatibility for Android servers.

• Storage:

  • These are I/O optimized instances.

  • Supported disk types: elastic ephemeral disk, ESSDs, ESSD AutoPL disks, and Regional ESSDs. For more information, see Block Storage Overview.

• Network:

  • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

The following table describes the instance types and key metrics of the ebmgn6ia instance family.

The vgn6i instance family provides the following features:

• Compute:

  • Uses NVIDIA T4 GPU accelerators.

  • Instances include vGPUs created by partitioning a physical GPU.

    • vGPUs provide 1/4 or 1/2 of the computing power of a physical NVIDIA Tesla T4 GPU.

    • This provides 4 GB and 8 GB of GPU memory, respectively.

  • A vCPU-to-memory ratio of approximately 1:5.

  • Processor: 2.5 GHz Intel^® Xeon^® Platinum 8163 (Skylake).

• Storage:

  • I/O optimized instance.

  • Supports only SSD cloud disks and ultra disks.

• Network:

  • Supports IPv6.

  • Network performance improves with larger instance types.

• Use cases:

  • Real-time rendering for cloud gaming.

  • Real-time cloud rendering for augmented reality (AR) and virtual reality (VR).

  • Deep learning (DL) and machine learning (ML) inference, ideal for elastically scaling internet applications.

  • Deep learning training and educational environments.

  • Deep learning model experimentation.

The following table lists the instance types and specifications of the vgn6i family.

• Scenarios:

  • Deep learning.

  • Scientific computing, such as computational fluid dynamics, computational finance, genomics research, and environmental analysis.

  • High-performance computing (HPC), rendering, multimedia encoding and decoding, and other server-side GPU compute workloads.

• Compute:

  • Features NVIDIA P100 GPUs.

  • Offers multiple processor-to-memory ratios.

  • Processor: 2.5 GHz Intel ^® Xeon ^® E5-2682 v4 (Broadwell).

• Storage:

  • Equipped with high-performance NVMe SSD local disks.

  • I/O optimized instances.

  • Supported cloud disk types: SSD cloud disks and Ultra Disks.

• Network:

  • Supports IPv4 only.

  • Network performance scales with the instance type.

The gn5 family includes the following instance types:

• Scenarios: Ideal for server-side GPU computing workloads such as deep learning inference and multimedia encoding and decoding.

• Compute:

  • Powered by NVIDIA P4 GPUs.

  • A vCPU-to-memory ratio of 1:4.

  • Processor: 2.5 GHz Intel^® Xeon^® E5-2682 v4 (Broadwell).

• Storage:

  • I/O optimized instances.

  • Supported cloud disk types: SSD cloud disks and ultra cloud disks.

• Network:

  • Supports IPv4 and IPv6. For more information about IPv6 communication, see IPv6 communication.

  • Network performance scales with the instance type.

The gn5i instance family includes the following instance types and specifications:

The vgn5i instance family offers the following features:

• Compute:

  • Powered by NVIDIA P4 GPU accelerators.

  • Instances include vGPUs created through GPU slicing.

    • vGPUs provide 1/8, 1/4, 1/2, or the full computing power of a single NVIDIA Tesla P4 GPU.

    • vGPUs are available with 1 GB, 2 GB, 4 GB, or 8 GB of GPU memory.

  • A vCPU-to-memory ratio of 1:3.

  • Processor: 2.5 GHz Intel^® Xeon^® E5-2682 v4 (Broadwell).

• Storage:

  • I/O-optimized instance.

  • Supports only SSD cloud disks and Ultra Disks.

• Network:

  • Supports IPv6.

  • Network performance scales with the instance size.

• Use cases:

  • Real-time cloud rendering for gaming.

  • Real-time cloud rendering for augmented reality (AR) and virtual reality (VR).

  • AI inference for deep learning (DL) and machine learning (ML), ideal for elastically scaling internet applications with AI capabilities.

  • Deep learning training and educational environments.

  • Deep learning model experimentation.

The following table lists the specifications for the vgn5i instance family.

The sccgn6e instance family offers the following features:

• Includes all the features of an ECS Bare Metal Instance. For more information, see ECS Bare Metal Instance families.

• Compute:

  • GPU accelerator:

    • Innovative Volta architecture

    • 32 GB of HBM2 GPU memory

    • 5,120 CUDA Cores

    • 640 Tensor Cores

    • 900 GB/s GPU memory bandwidth

    • A single GPU supports six bidirectional NVLink links, each with a unidirectional bandwidth of 25 GB/s, for a total bidirectional bandwidth of 300 GB/s.

  • vCPU-to-memory ratio of 1:8

  • Processor: 2.5 GHz Intel^® Xeon^® Platinum 8163 (Skylake)

• Storage:

  • I/O optimized instance

  • Supports only ESSDs, ESSD AutoPL disks, standard SSDs, and ultra disks

  • Supports CPFS

• Network:

  • Supports IPv6

  • Supports VPCs

  • Supports RoCE v2 networks for low-latency RDMA.

• Use cases:

  • Ultra-large-scale machine learning cluster training

  • Large-scale high-performance scientific computing and simulation

  • Large-scale data analytics, batch computing, and video encoding

The following table lists the specifications of the sccgn6e instance family.

The sccgn6 instance family includes the following features:

• Includes all features of an ECS Bare Metal Instance. For more information, see ECS Bare Metal Instance families.

• Compute:

  • GPU accelerator: NVIDIA V100 (SXM2-based)

    • Innovative Volta architecture

    • 16 GB HBM2 GPU memory

    • 5,120 CUDA Cores

    • 640 Tensor Cores

    • 900 GB/s GPU memory bandwidth

    • Each GPU supports six bidirectional NVLink links, each providing 25 GB/s of bandwidth in a single direction for a total of 300 GB/s of bidirectional bandwidth.

  • A vCPU-to-memory ratio of 1:4

  • Processor: 2.5 GHz Intel^® Xeon^® Platinum 8163 (Skylake)

• Storage:

  • I/O optimized instance

  • Supports only ESSD cloud disk, ESSD AutoPL cloud disk, Standard SSD, and Ultra Disk.

  • Supports high-performance CPFS.

• Network:

  • Supports IPv6.

  • Supports VPCs.

  • Enables low-latency RDMA communication through RoCE v2 networks.

• Use cases:

  • Training for ultra-large-scale machine learning clusters

  • Large-scale high-performance scientific computing and simulation

  • Large-scale data analytics, batch computing, and video encoding

The following table describes the instance types and key metrics for the sccgn6 instance family.

The sccgn6ne instance family offers the following features:

• Provides all the features of an ECS Bare Metal Instance.

• Compute:

  • GPU accelerator: V100 (SXM2 form factor)

    • Innovative Volta architecture

    • 32 GB of HBM2 GPU memory

    • 5,120 CUDA Cores

    • 640 Tensor Cores

    • 900 GB/s GPU memory bandwidth

    • Six NVLink links at 25 GB/s each, for a total bandwidth of 300 GB/s.

  • A vCPU-to-memory ratio of 1:4

  • Processor: 2.5 GHz Intel^® Xeon^® Platinum 8163 (Skylake).

• Storage:

  • I/O optimized instance

  • Support for ESSDs, Standard SSDs, and Ultra Disks.

  • Support for the high-performance Cloud Paralleled File System (CPFS).

• Network:

  • Support for IPv6.

  • Support for Virtual Private Cloud (VPC).

  • Support for RoCE v2.

• Use cases:

  • Ultra-large-scale machine learning cluster training

  • Large-scale high-performance scientific and simulation computing

  • Large-scale data analytics, batch computing, and video encoding

The following table lists the specifications of the sccgn6ne instance family.

The gn4 instance family provides the following features:

• Powered by NVIDIA M40 GPUs

• Compute:

  • Various vCPU-to-memory ratios

  • Processor: 2.5 GHz Intel^® Xeon^® E5-2682 v4 (Broadwell)

• Storage:

  • I/O optimized instance

  • Supports only Standard SSD and Ultra Disk

• Network:

  • Larger instance types provide higher network performance.

• Use cases:

  • Deep learning

  • Scientific computing, including computational fluid dynamics, computational finance, genomics research, and environmental analysis

  • High-performance computing, rendering, multimedia encoding and decoding, and other server-side GPU computing workloads

The following table describes the instance types and specifications of the gn4 instance family.

The ga1 instance family provides the following features:

• Powered by AMD S7150 GPU cards.

• Features high-performance NVMe SSD local disks.

• Compute:

  • A vCPU-to-memory ratio of 1:2.5.

  • Processor: 2.5 GHz Intel^® Xeon^® E5-2682 v4 (Broadwell).

• Storage:

  • I/O optimized instances.

  • Supports only Standard SSD and Ultra Disk.

• Network:

  • Network performance scales with the instance type.

• Use cases:

  • Rendering and multimedia encoding/decoding.

  • Machine learning, high-performance computing, and high-performance databases.

  • Other server-side workloads requiring powerful parallel floating-point computing capabilities.

The following table lists the specifications of the ga1 instance types.