NCasT4_v3-series

The NCasT4_v3-series virtual machines are powered by Nvidia Tesla T4 GPUs and AMD EPYC 7V12(Rome) CPUs. The VMs feature up to 4 NVIDIA T4 GPUs with 16 GB of memory each, up to 64 non-multithreaded AMD EPYC 7V12 (Rome) processor cores(base frequency of 2.45 GHz, all-cores peak frequency of 3.1 GHz and single-core peak frequency of 3.3 GHz) and 440 GiB of system memory. These virtual machines are ideal for deploying AI services- such as real-time inferencing of user-generated requests, or for interactive graphics and visualization workloads using NVIDIA's GRID driver and virtual GPU technology. Standard GPU compute workloads based around CUDA, TensorRT, Caffe, ONNX and other frameworks, or GPU-accelerated graphical applications based on OpenGL and DirectX can be deployed economically, with close proximity to users, on the NCasT4_v3 series.


ACU: 230-260
Premium Storage: Supported
Premium Storage caching: Supported
Ultra Disks: Supported (Learn more about availability, usage, and performance)
Live Migration: Not Supported
Memory Preserving Updates: Not Supported
VM Generation Support: Generation 1 and 2
Accelerated Networking: Supported
Ephemeral OS Disks: Supported (In preview)
Nvidia NVLink Interconnect: Not Supported

Size vCPU Memory: GiB Temp storage (SSD) GiB GPU GPU memory: GiB Max data disks Max NICs / Expected network bandwidth (Mbps)
Standard_NC4as_T4_v3 4 28 180 1 16 8 2 / 8000
Standard_NC8as_T4_v3 8 56 360 1 16 16 4 / 8000
Standard_NC16as_T4_v3 16 110 360 1 16 32 8 / 8000
Standard_NC64as_T4_v3 64 440 2880 4 64 32 8 / 32000

Size table definitions

  • Storage capacity is shown in units of GiB or 1024^3 bytes. When you compare disks measured in GB (1000^3 bytes) to disks measured in GiB (1024^3) remember that capacity numbers given in GiB may appear smaller. For example, 1023 GiB = 1098.4 GB.

  • Disk throughput is measured in input/output operations per second (IOPS) and MBps where MBps = 10^6 bytes/sec.

  • Data disks can operate in cached or uncached modes. For cached data disk operation, the host cache mode is set to ReadOnly or ReadWrite. For uncached data disk operation, the host cache mode is set to None.

  • To learn how to get the best storage performance for your VMs, see Virtual machine and disk performance.

  • Expected network bandwidth is the maximum aggregated bandwidth allocated per VM type across all NICs, for all destinations. For more information, see Virtual machine network bandwidth.

    Upper limits aren't guaranteed. Limits offer guidance for selecting the right VM type for the intended application. Actual network performance will depend on several factors including network congestion, application loads, and network settings. For information on optimizing network throughput, see Optimize network throughput for Azure virtual machines. To achieve the expected network performance on Linux or Windows, you may need to select a specific version or optimize your VM. For more information, see Bandwidth/Throughput testing (NTTTCP).

Supported operating systems and drivers

To take advantage of the GPU capabilities of Azure NCasT4_v3-series VMs running Windows or Linux, Nvidia GPU drivers must be installed.

To install Nvidia GPU drivers manually, see N-series GPU driver setup for Windows for supported operating systems, drivers, installation, and verification steps.

The Azure Nvidia GPU driver extension will deploy CUDA drivers on the NCasT4_v3-series VMs. For graphics and visualization workloads manually install the GRID drivers supported by Azure.

Other sizes

Next steps

Learn more about how Azure compute units (ACU) can help you compare compute performance across Azure SKUs.