This article briefly describes the steps for running the Autodesk Maya application on a virtual machine (VM) that's deployed on Azure. It also presents the performance results of running Maya on Azure.
Maya is a professional toolset for 3D animation, modeling, simulation, and rendering. You can use Maya to create realistic characters, landscapes, and action sequences. Artists, modelers, and animators rely on this award-winning toolset to add lifelike 3D assets to animated and live-action films, TV shows, and video games.
Key features of Maya include:
- Bifrost for Maya, a plug-in that you can use to create physically accurate simulations in a single visual programming environment.
- USD in Maya, an extension for the Universal Scene Description (USD) framework. You can use this extension to load and edit large datasets quickly and to use native tools to work directly with data.
- Fast playback, which provides a fast way to review animations and leads to reduced playblasts with the Cached Playback system in Viewport 2.0. A playblast is a real-time preview of an animation in Maya.
- Unreal Live Link for Maya, a plug-in that you can use to stream real-time animation data from Maya to Unreal.
- A nondestructive, clip-based nonlinear time editor for making high-level animation edits.
- A graph editor. You can use this graphical representation of scene animation to create, view, and modify animation curves.
- Polygon modeling, a feature for creating 3D models that uses geometry that's based on vertices, edges, and faces.
- Nonuniform rational basis spline (NURBS) modeling. By using this feature, you can construct 3D models from geometric primitives and drawn curves.
- Character setup for creating sophisticated skeletons, inverse kinematics (IK) handles, and deformers for characters that deliver lifelike performances.
Maya was developed by Autodesk and offers a wide range of powerful tools for animation, simulation, and modeling. You can also use Maya for motion graphics, virtual reality, UV maps, low poly, and character creation. This 3D modeling software is popular in the video game industry. The Maya application can generate 3D assets for games but also for film, TV, and commercials. Besides offering a vast library of animation tools, Maya is customizable. You can use scripting languages like Maya Embedded Language (MEL) and Python to extend Maya functionality.
For information about a plug-in for using an Arnold renderer directly in Maya, see the Arnold for Maya user guide.
Why deploy Maya on Azure?
- Modern and diverse compute options to align with your workload's needs
- The flexibility of virtualization without the need to buy and maintain physical hardware
- Rapid provisioning
- Strong GPU acceleration, with increased performance as you add GPUs
Architecture
Download a Visio file of this architecture.
Components
Azure Virtual Machines is used to create Windows VMs. For information about deploying a VM and installing drivers, see Run a Windows VM on Azure.
Azure Virtual Network is used to create a private network infrastructure in the cloud.
Network security groups restrict access to VMs at the subnet level.
A public IP address provides users with access to Maya via the internet.
Azure Disk Storage provides a physical solid-state drive (SSD) for storage.
Deploy infrastructure and install Maya
Deploy Azure VMs. Before you install Maya, deploy your Azure VMs. Use a NVadsA10_v5-series VM to run Maya. You should use a Premium SSD managed disk and attach it to the VM.
Create and configure the supporting infrastructure. Configure a public IP address for inbound connectivity. Use network security groups to provide security for the subnet.
Install NVIDIA drivers. To take advantage of the GPU capabilities of NVadsA10_v5-series VMs, install NVIDIA GPU drivers. For information about deploying VMs and installing the drivers, see Run a Windows VM on Azure.
Compute sizing
| Size | vCPUs | Memory (GiB) | SSD (GiB) | GPU partition | GPU memory (GiB) | Maximum data disks |
|---|---|---|---|---|---|---|
| Standard_NV6ads_A10_v5 | 6 | 55 | 180 | 1/6 | 4 | 4 |
| Standard_NV12ads_A10_v5 | 12 | 110 | 360 | 1/3 | 8 | 4 |
| Standard_NV18ads_A10_v5 | 18 | 220 | 720 | 1/2 | 12 | 8 |
| Standard_NV36ads_A10_v5 | 36 | 440 | 720 | 1 | 24 | 16 |
| Standard_NV72ads_A10_v5 | 72 | 880 | 1400 | 2 | 48 | 32 |
Maya installation
Before you install Maya, you need to deploy and connect to a VM via Remote Desktop Protocol (RDP) and install the required NVIDIA drivers.
Important
An NVIDIA Fabric Manager installation is required for VMs that use NV Link or NV Switch.
You can install Maya from the Autodesk Maya portal. For a detailed installation procedure, see the Autodesk Maya help documentation.
Maya performance results
This performance analysis uses the Autodesk Maya 2023.1 trial version on Windows NVadsA10_v5-series VMs.
The following table provides details about the testing operating system:
| Version | Operating system | Architecture | Processor |
|---|---|---|---|
| Maya 2023.1 | Windows 10 Professional, 20H2 | x86-64 | AMD EPYC 74F3V (Milan) |
| Maya 2024 | Windows 11 Professional, 22H2 | x86-64 | AMD EPYC 74F3V (Milan) |
The tests use the following model:
| Model name | Image size | Image size in inches |
|---|---|---|
| Open scene | HD 1080 | 26.7 x 15.0 |
The following sections provide information about four metrics that measure the performance of Maya on an Azure VM.
Measurement 1: Open scene performance
The following table lists times for reading the model file with various GPU configurations. The time that it takes to load a test scene is the time that it takes to read a file.
Results on Maya 2023.1
| Number of GPUs | File read time (seconds) |
|---|---|
| 1/3 | 15 |
| 1/2 | 13.3 |
| 1 | 12.9 |
Results on Maya 2024
| Number of GPUs | File read time (seconds) |
|---|---|
| 1/3 | 13.1 |
| 1/2 | 15.4 |
| 1 | 12.1 |
On tests with 1/6 GPUs, Maya reports an error about not enough free memory remaining in the GPU.
The following image shows the loading view, which is the view that the independent software vendor (ISV) expects to see:
Measurement 2: Playback performance
The following table lists playback times and rates for various GPU configurations when a MEL script is used:
Results on Maya 2023.1
| Number of GPUs | Playback rate (frames/second) | Total playback time (seconds) |
|---|---|---|
| 1/3 | 23.89 | 17.79 |
| 1/2 | 24.02 | 17.69 |
| 1 | 24.02 | 17.69 |
Results on Maya 2024
| Number of GPUs | Playback rate (frames/second) | Total playback time (seconds) |
|---|---|---|
| 1/3 | 27.74 | 15.32 |
| 1/2 | 28.46 | 14.93 |
| 1 | 43.41 | 9.79 |
The model contains 425 frames. The playback time and rate are almost identical for all tested GPU configurations.
Measurement 3: Arnold renderer performance
The following table shows the rendering time for various GPU configurations. Results from tests with six vCPUs (cores) are used as the baseline for determining speedup.
Results on Maya 2023.1
| Number of GPUs | Rendering time (seconds) | Relative speedup |
|---|---|---|
| 6 vCPUs (cores) | 1,275 | NA |
| 1/3 | 138 | 9.24 |
| 1/2 | 96 | 13.28 |
| 1 | 27 | 47.22 |
The following chart shows the relative speed increases for the model as the number of GPUs increases:
Results on Maya 2024
| Number of GPUs | Rendering time (seconds) | Relative speedup |
|---|---|---|
| 6 vCPUs (cores) | 1,314 | NA |
| 1/3 | 115 | 11.43 |
| 1/2 | 91 | 14.44 |
| 1 | 35 | 37.54 |
The following chart shows the relative speed increases for the model as the number of GPUs increases:
The following image shows the rendered output:
Measurement 4: Caching performance
The following table lists performance data for caching via a Python script. When you use the Cached Playback feature, you can view changes that you make to the animation without having to create a new playblast.
Results on Maya 2023.1
| Number of GPUs | Filling playback (frames/second) | Cached playback (frames/second) | Fill time (seconds) |
|---|---|---|---|
| 1/3 | 6.19 | 21.31 | 0.88 |
| 1/2 | 17.09 | 24.11 | 0.67 |
| 1 | 18.86 | 24.11 | 0.60 |
Results on Maya 2024
| Number of GPUs | Filling playback (frames/second) | Cached playback (frames/second) | Fill time (seconds) |
|---|---|---|---|
| 1/3 | 14.38 | 45.72 | 0.91 |
| 1/2 | 14.70 | 49.40 | 0.75 |
| 1 | 18.34 | 52.80 | 0.62 |
Additional notes about tests
- Tests on the Maya application run successfully on NVadsA10_v5 VMs on the Azure cloud platform.
- In tests of the rendering process on an NVadsA10_v5 VM, Maya scales well from a partial to a full GPU configuration.
- Tests of the caching performance show that the fill time decreases as the configuration improves.
Azure cost
The following table lists the elapsed times in hours for running the model with various GPU configurations. To compute the total cost, multiply these times by the Azure VM hourly cost for an NVadsA10_v5 VM. For the current hourly cost, see Windows Virtual Machines Pricing and Linux Virtual Machines Pricing.
Maya 2023.1
| Number of GPUs or CPUs | Elapsed time (hours) |
|---|---|
| 1/3 GPUs | 0.038 |
| 1/2 GPUs | 0.027 |
| 1 GPU | 0.008 |
| 6 vCPUs | 0.354 |
Maya 2024
| Number of GPUs or CPUs | Elapsed time (hours) |
|---|---|
| 1/3 GPUs | 0.031 |
| 1/2 GPUs | 0.025 |
| 1 GPU | 0.009 |
| 6 vCPUs | 0.365 |
For these cost calculations, only the model rendering time is considered, measured in wall-clock time. The application installation time isn't considered. The calculations are indicative of your potential results, but actual values depend on the size of your model.
To estimate the cost of your configuration, use the Azure pricing calculator.
Summary
- You can successfully deploy and run Maya on NVadsA10_v5-series VMs.
- In tests of the rendering process on an NVadsA10_v5 VM, Maya scales well from a partial to a full GPU configuration.
Contributors
This article is maintained by Microsoft. It was originally written by the following contributors.
Principal authors:
- Hari Bagudu | Senior Manager
- Gauhar Junnarkar | Principal Program Manager
- Vinod Pamulapati | HPC Performance Engineer
Other contributors:
- Guy Bursell | Director Business Strategy
- Sachin Rastogi | Manager
To see nonpublic LinkedIn profiles, sign in to LinkedIn.
Next steps
- GPU-optimized virtual machine sizes
- Virtual machines in Azure
- Virtual networks and virtual machines in Azure
- Training path: Run high-performance computing (HPC) applications on Azure