Fix unspecified device in Newton Renderer camera transforms

[nblauch]

Description

Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context.
List any dependencies that are required for this change.

Fixes # (issue)

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (existing functionality will not work without user modification)
• Documentation update

Screenshots

Please attach before and after screenshots of the change if applicable.

Checklist

• I have read and understood the contribution guidelines
• I have run the pre-commit checks with ./isaaclab.sh --format
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• I have updated the changelog and the corresponding version in the extension's config/extension.toml file
• I have added my name to the CONTRIBUTORS.md or my name already exists there

[greptile-apps]

Greptile Summary

This PR fixes a bug in RenderData.update() where wp.empty() and wp.launch() for the camera_transforms array were missing an explicit device argument, defaulting to Warp's default device instead of render_context.device. This could cause device-mismatch runtime errors when the simulation runs on a non-default device.

• Fix is correct: device=self.render_context.device is now properly passed to both wp.empty() (array allocation) and wp.launch() (kernel execution), ensuring the array lives on and the kernel runs on the renderer's target device.
• Dtype inconsistency introduced: wp.empty() now uses dtype=wp.transform, while the class-level type annotation (line 60), kernel parameter (line 140), and kernel body (line 143) all continue to use wp.transformf. These are equivalent types in Warp, so there is no functional issue, but the inconsistency is worth cleaning up.
• CHANGELOG not updated: The PR checklist shows the changelog entry as unchecked; the fix should be recorded in source/isaaclab_newton/docs/CHANGELOG.rst under the current version.
• Input tensor device: The positions and converted_orientations PyTorch tensors are passed to wp.launch without explicit verification that they reside on render_context.device. This is pre-existing behaviour, but is worth keeping in mind if multi-GPU scenarios are supported in the future.

Confidence Score: 4/5

• This PR is safe to merge; it correctly fixes a real device-mismatch bug with a targeted, minimal change
• The fix is clearly correct — missing device arguments on wp.empty and wp.launch would cause allocation and kernel execution on the wrong device. The only concern is a minor wp.transform / wp.transformf naming inconsistency introduced in the same hunk, which is functionally harmless since the two types are aliases in Warp. No tests are added, and the CHANGELOG was not updated.
• No files require special attention beyond the single changed file.

Important Files Changed

Filename	Overview
source/isaaclab_newton/isaaclab_newton/renderers/newton_warp_renderer.py	Correctly adds device=self.render_context.device to wp.empty() and wp.launch() in RenderData.update(), fixing device-unspecified allocations that could cause device mismatch errors; minor inconsistency between wp.transform (new code) and wp.transformf (class annotation, kernel signature, kernel body)

Sequence Diagram

sequenceDiagram
    participant Sensor
    participant RenderData
    participant WarpRuntime

    Sensor->>RenderData: update(positions, orientations, intrinsics)
    RenderData->>WarpRuntime: wp.empty((1, world_count), dtype=wp.transform,<br/>device=render_context.device)
    WarpRuntime-->>RenderData: camera_transforms [on render_context.device]
    RenderData->>WarpRuntime: wp.launch(_update_transforms, world_count,<br/>[positions, orientations, camera_transforms],<br/>device=render_context.device)
    WarpRuntime-->>RenderData: camera_transforms populated
    RenderData->>WarpRuntime: compute_pinhole_camera_rays(width, height, fov)
    WarpRuntime-->>RenderData: camera_rays

Loading

_{Last reviewed commit: bd52495}