ROCm-LS 25.11 release notes

ROCm-LS 25.11 release notes#

3 min read time

Applies to Linux

The release notes provide a summary of notable changes since the previous ROCm-LS release.

ROCm-LS release highlights#

The following are notable new features and improvements in ROCm-LS 25.11 since the release of 25.09. For detailed changes to individual components, see detailed component changes.

  • hipCIM 25.10.00:

    • hipCIM 25.10.00 exits Early Access (EA) and is now production-ready for life sciences imaging workloads on AMD GPUs.

    • Upstream parity: hipCIM 25.10.00 is based on upstream RAPIDS™ cuCIM v25.10.00 for feature and API parity, as well as easier migration of your existing applications.

    • Region property computations and utility functions: This release delivers significant performance and feature improvements for GPU-accelerated imaging on AMD ROCm, with a focus on region property computations and utility functions for common image workflows.

  • ROCm 7.0 support: ROCm-LS 25.11 adds support for ROCm 7.0 while continuing support for ROCm 6.4.

  • MONAI 1.0.0 is based on ROCm 6.4 and continues to be in EA.

ROCm-LS components#

The following table lists the versions of ROCm-LS components for ROCm-LS 25.11, including any version changes from 25.09 to 25.11. Click the GitHub icon to go to the component’s source code.

Category Component name Version Source code
Imaging hipCIM 25.04.00 ⇒ 25.10.00
AI/ML MONAI for AMD ROCm 1.0.0

Note

The MONAI version remains unchanged in this release.

Detailed component changes#

The following sections describe key changes to the ROCm-LS components:

hipCIM (25.10.00)#

This release is based on the upstream cuCIM v25.10.00 release and includes the following enhancements:

Added#

  • ROCm versions: Support for ROCm 7.0 with continued support for ROCm 6.4.

  • Utility functions: Montage and compare_images utility functions for quickly visualizing and validating image outputs in analysis workflows.

  • Geometry and morphology: GPU-accelerated implementations of convex_hull_image and convex_hull_object functions.

  • API extensions: Region properties API extended with regionprops_table and accelerated implementations across basic, intensity, moment-based, convexity, perimeter, and Euler characteristic properties.

Optimized#

  • End-to-end acceleration of regionprops and regionprops_table, reducing computation time and memory overhead on common segmentation and labeling tasks.

  • Improved performance for label2rgb, offering faster visualization of labeled images in typical post-processing pipelines.

Resolved issues#

  • Imaging integration: Improved compatibility with Pillow’s Image.fromarray when using deprecated mode parameters, preventing runtime issues in visualization workflows.

  • Morphology and binary operations: Refreshed binary morphology and binary_fill_holes components for correctness and stability.

Note

hipCIM targets functional parity with the most commonly used upstream APIs in the life sciences imaging. Minor behavioral differences might exist due to ROCm backend variations. Validate critical pipelines accordingly.

Contents