7.4.0.0 - Arnold User Guide

26 March 2025

Arnold 7.4.0.0 is a feature release including improved support for light instances, a new run-once imager mode, an upgrade to MaterialX 1.38.10, and API updates.

Enhancements

Lights in procedural instances: When you instance a procedural, any lights created by the procedural are copied for each instance. For example, instances of ASS or USD assets with lights now render the lights properly. (ARNOLD-15657)
Imager run once mode: Imagers can now be set to run only once, at the end of a bucket or the end of a completed final frame. This can be useful for particularly expensive imagers and can be enabled via the ImagerProperties::run_once property. (ARNOLD-12191)
OpenColorIO 2.4.1: OpenColorIO has been updated to v2.4.1. (ARNOLD-15681)
Improved MaterialX support: MaterialX has been upgraded to v1.38.10-OpenPBR, this new version brings support for the OpenPBR MaterialX node, improved Lama support and new nodes such as the NPR nodes. See v1.38.10, v1.38.9 release notes for details. (ARNOLD-15659)
Update Autodesk Desktop Licensing installer Arnold now uses Autodesk Desktop Licensing installer version 15.1, which includes the following components: AdskLicensingService 11.16.6.5912, AdskLicensingAgent 5.20.12.4281, and AdskLicensingInstHelper 2.13.1.2514.

USD Enhancements

Distant Lights normalize support : Distant lights now support the normalize attribute, so you can now disable normalize (before it was always enabled and the normalize attribute was ignored). You can use the Arnold option usd_legacy_distant_light_normalize to roll back to the previous behaviour. (usd#2176)
USD 24.11: The USD procedural now uses USD 24.11.

API Additions

Replace AtImagerSchedule with AtImagerProperties: The imager_prepare API has a new AtImagerProperties structure that provides more control over how an Imager should run. (ARNOLD-15774)

Setting Imager bucket / full frame schedules will now work like so:
```
 imager_prepare
 {
    if (run_imager_full_frame)
       properties.schedule_full_frame = true;
    else
       properties.schedule_bucket = true;
 }
```

Improved return values for device selection API: AiDeviceSelect and AiDeviceAutoSelect now return the AtDeviceSelectErrorCode enum instead of an int to better describe the error status of the API call. (ARNOLD-14652)

AtDeviceSelectErrorCode
SUCCESS	Device selection was successful
FAILURE_NO_DEVICES_FOUND	No Arnold compatible devices found
FAILURE_NAME	No matches found for options.gpu_default_names regular expression
FAILURE_MEMORY	No matching device has free memory greater than options.gpu_default_min_memory_MB threshold
FAILURE_INVALID_ID	Invalid device id in selection
FAILURE_UNSUPPORTED	Unsupported device in selection
FAILURE_UNKNOWN	Generic error

Incompatible Changes

This version breaks binary compatibility. Shaders, procedurals, and other plugins compiled against older versions of Arnold are not compatible with this version of Arnold and must be updated and recompiled.

Autodesk Network Licensing version: Arnold 7.4 switches to ARNOL_2026 licensing. (ARNOLD-15400)
- Single-user subscriptions will work with Arnold 7.4. The plugin installers automatically register the new version. If you do not use a plugin, then after you download Arnold, open the Arnold License Manager and click Register.
- Existing network licenses will not work with Arnold 7.4. Log on to your Autodesk account to generate your new ARNOL_2026 licenses.
- If you do not run the installers, and you manually register Arnold, then use--pk C0PR1 --pv 2026.0.0.F with the AdskLicensingInstHelper command.
Removed sss_use_autobump option: Use the per-object autobump_visibility flag. (ARNOLD-15782)
Removed legacy_volume_albedo: This option is no longer available as the new volume albedo is more useful. (ARNOLD-15782)
Removed AiMsgSetCallback(), AiMsgAddCallback(), AiMsgResetCallback(): Use AiMsgRegisterCallback() and AiMsgDeregisterCallback() instead. (ARNOLD-15782)
Removed AiASSLoad(), AiASSWrite(), AiASSWriteWithMetadata(): Use AiSceneLoad() and AiSceneWrite() instead. (ARNOLD-15782)
Removed deprecated versions of AiOrenNayarBSDF() and AiMicrofacetSetThinFilm(): Use the version with the same function name but additional arguments. (ARNOLD-15782)
Removed AiLoadImage(): Use AiReadImage() instead. (ARNOLD-15782)
Deprecated enable_new_quad_light_sampler and enable_new_point_light_sampler options: The new samplers should always be better, so these options are marked as deprecated and will be removed in a future version of Arnold. (ARNOLD-15782)
Distant Lights normalize support : Distant lights now support the normalize attribute, so you can now disable normalize (before it was always enabled and the normalize attribute was ignored). You can use the Arnold option usd_legacy_distant_light_normalize to roll back to the previous behaviour. (usd#2176)

Bug Fixes

ARNOLD-15713 - [MaterialX] Mix shader does not work as surface shader
ARNOLD-15723 - No visible layering when layering a dielectric on top of a diffuse Lama shader. In MaterialX, layering of a Lama dielectric node on top of a Lama diffuse node was not generating the correct appearance of the diffuse surface coated with the dielectric. The OSL layering logic in Arnold has been improved to fix this issue.
usd#2201 - Hydra procedural should not modify the input usd stage with shutter range
usd#2197 - MaterialX textures not applied properly in hydra through mayaHydra

System Requirements

Windows 10 or later, with the Visual Studio 2019 redistributable.
Linux with at least glibc 2.17 and libstdc++ 4.8.5 (gcc 4.8.5). This is equivalent to RHEL/CentOS 7.
macOS 11 or later.
x86-64 CPUs need to support the SSE4.1 instruction set. Apple Mac models with M series chips are natively supported.
GPU rendering and OptiX denoising work on Windows and Linux only, and require an NVIDIA GPU with the Maxwell architecture or later.
- On Linux, we recommend 560.35.03 or higher drivers.
- On Windows, we recommend 560.76 or higher drivers.
Intel OIDN GPU support is limited to:
- MacOS
  - Apple native CPUs (M1 and newer)
- Windows
  - Intel Xe dedicated and integrated GPUs
  - NVIDIA GPUs using Turing or newer architectures
  - AMD GPUs with RDNA2 (Navi 21 only) and RDNA3 (Navi 3x) architectures