Fix OvPhysX 0.4 compatibility#5545
Merged
Merged
Conversation
Contributor
Greptile SummaryThis PR fixes several OvPhysX 0.4 API compatibility issues in IsaacLab: the
Confidence Score: 4/5The changes are well-scoped bug fixes with no cross-cutting regressions; the main area worth a closer look is the ovphysx constructor dispatch in the manager. The CPU-only tensor staging, write-view raw-pointer fix, and sensor class-name guard are all straightforward and correctly implemented. The only open questions are whether older ovphysx wheels require gpu_index even for CPU-mode instantiation, and whether the active_cuda_gpus string format matches the published 0.4 spec. Both are low-impact edge cases but warrant a quick confirmation before merge. source/isaaclab_ovphysx/isaaclab_ovphysx/physics/ovphysx_manager.py — the dual-path constructor dispatch and the active_cuda_gpus string representation deserve a final check against the ovphysx 0.4 API surface. Important Files Changed
Flowchart%%{init: {'theme': 'neutral'}}%%
flowchart TD
A[OvPhysxManager._warmup_and_load] --> B{ovphysx_device == gpu?}
B -- No --> C[physx_kwargs = device:cpu]
B -- Yes --> D{active_cuda_gpus in PhysX signature?}
D -- Yes, new 0.4 API --> E[kwargs += active_cuda_gpus + PhysXConfig]
D -- No, old API --> F[kwargs += gpu_index]
C --> G[ovphysx.PhysX instantiated]
E --> G
F --> G
subgraph Read path in ArticulationData
H[_read_binding_into_view] --> I{CPU_ONLY_TYPES AND view.device != cpu?}
I -- Yes --> J[binding.read to CPU scratch]
J --> K[wp.copy: CPU scratch to GPU view]
I -- No --> L[binding.read directly into view]
end
subgraph sensor_base.py
M[Register callbacks] --> N{physics_mgr_cls.__name__ == PhysxManager?}
N -- Yes --> O[Register prim-deletion callback]
N -- No --> P[Skip prim-deletion]
end
Reviews (1): Last reviewed commit: "Fix OvPhysX 0.4 compatibility" | Re-trigger Greptile |
Update the OvPhysX backend for the upcoming ovphysx 0.4 API while preserving the older constructor path. Fix CPU-only tensor binding reads into GPU buffers, use raw Warp buffers for write views, and add the OvPhysX preset to the cartpole camera task.
5318097 to
6f33d74
Compare
pbarejko
approved these changes
May 8, 2026
7 tasks
AntoineRichard
added a commit
that referenced
this pull request
May 19, 2026
# Description Implements `RigidObjectCollection` and `RigidObjectCollectionData` for the OVPhysX backend, completing the rigid-body asset surface alongside `RigidObject` (#5426) and `Articulation` (#5459). The collection manages N distinct rigid bodies per environment with `(env, body)` dual indexing. The asset creates **one native fused TensorBinding per tensor type** via the ovphysx 0.4.3 `create_tensor_binding(prim_paths=[glob_0, …, glob_{B-1}])` API, mirroring how PhysX's `RigidBodyView` aggregates multiple body prims into a single flat view. Each binding spans `num_instances * num_bodies` prims and returns body-major flat data `(body_0_env_0, body_0_env_1, …, body_1_env_0, …)`. The data class and asset writers use strided-view reshape helpers (`_reshape_view_to_data_2d/_3d` and `reshape_data_to_view_2d/_3d`, ported from the PhysX collection) to convert between body-major view layout and the instance-major `(N, B, D)` layout exposed to users — no Warp kernels added, no per-body Python fan-out at runtime. Fixes #5317 **Stacked on:** - #5459 — OVPhysX `Articulation` - #5426 — OVPhysX `RigidObject` **Carries a one-commit cherry-pick of #5545's `ovphysx_manager.py` portion** (required for ovphysx 0.4 `active_cuda_gpus` API). Drop the cherry-pick once #5545 lands. ## Type of change - New feature (non-breaking change which adds functionality) ## Screenshots N/A — backend infrastructure, no visible behaviour change. ## Checklist - [x] I have read and understood the [contribution guidelines](https://isaac-sim.github.io/IsaacLab/main/source/refs/contributing.html) - [x] I have run the [pre-commit checks](https://pre-commit.com/) with `./isaaclab.sh --format` - [x] I have made corresponding changes to the documentation - [x] My changes generate no new warnings - [x] I have added tests that prove my fix is effective or that my feature works - [x] I have updated the changelog (fragment file under `source/isaaclab_ovphysx/changelog.d/`) - [x] I have added my name to the `CONTRIBUTORS.md` or my name already exists there ## Test plan Tested in a Docker container built from `docker/Dockerfile.base` (IsaacSim `6.0.0-dev2`) with the ovphysx 0.4.3 wheel installed: - `./scripts/run_ovphysx.sh -m pytest source/isaaclab/test/assets/test_rigid_object_collection_iface.py -k ovphysx` → **636 passed** - `./scripts/run_ovphysx.sh -m pytest source/isaaclab_ovphysx/test/assets/test_rigid_object_collection.py` → **72 passed, 76 skipped (device-mode lock — a second invocation with `-k 'cpu'` covers CPU), 4 xfailed** (material-properties gap shared with `RigidObject` until the wheel exposes `RIGID_BODY_MATERIAL`) - `./isaaclab.sh -f` → clean
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
active_cuda_gpusand explicit DirectGPU Carbonite settings when supported, while preserving the oldergpu_indexconstructor path.ProxyArraywrappers.ovphysxphysics preset to the cartpole camera presets task.Validation
./isaaclab.sh -f./isaaclab.sh -p -m pytest source/isaaclab_ovphysx/test/assets/test_articulation_data.py source/isaaclab_ovphysx/test/assets/test_articulation.py./isaaclab.sh -p scripts/reinforcement_learning/rsl_rl/train.py --task Isaac-Cartpole-Direct-v0 --num_envs 64 --max_iterations 2 --headless presets=ovphysx./isaaclab.sh -p scripts/reinforcement_learning/rsl_rl/train.py --task Isaac-Ant-Direct-v0 --num_envs 64 --max_iterations 2 --headless presets=ovphysx./isaaclab.sh -p scripts/reinforcement_learning/rsl_rl/train.py --task Isaac-Humanoid-Direct-v0 --num_envs 64 --max_iterations 2 --headless presets=ovphysx./isaaclab.sh -p scripts/reinforcement_learning/rl_games/train.py --task=Isaac-Cartpole-Camera-Presets-Direct-v0 --num_envs=32 --max_iterations=2 --headless --enable_cameras presets=ovphysx,ovrtx_renderer,rgbDescription
This PR fixes several small IsaacLab-side issues needed for the OvPhysX backend to run the supported direct cartpole, ant, and humanoid tasks with the upcoming ovphysx 0.4 wheel. It also enables the cartpole camera presets task to select the
ovphysxphysics preset.The OvPhysX manager now detects the new constructor surface and passes explicit DirectGPU settings for GPU simulations. Older public wheels that still use
gpu_indexkeep the previous constructor path.Fixes # (not applicable)
Type of change
Screenshots
Not applicable.
Checklist
pre-commitchecks with./isaaclab.sh --formatsource/<pkg>/changelog.d/for every touched package (do not editCHANGELOG.rstor bumpextension.toml— CI handles that)CONTRIBUTORS.mdor my name already exists there