Repository navigation
Conversation
Contributor
There was a problem hiding this comment.
go2 replay realtime (arm64)
Details
| Benchmark suite | Current: 744dd85 | Previous: 243fecd | Ratio |
|---|---|---|---|
peak memory |
1761.289 MB |
1784.68 MB |
0.99 |
peak threads |
384 threads |
377 threads |
1.02 |
network (transport) |
2730.56 MB |
2728.316 MB |
1.00 |
disk write |
1.922 MB |
1.941 MB |
0.99 |
instructions |
138.955 G |
138.933 G |
1.00 |
This comment was automatically generated by workflow using github-action-benchmark.
Codecov Report❌ Patch coverage is @@ Coverage Diff @@
## main #5365 +/- ##
==========================================
+ Coverage 84.76% 84.80% +0.03%
==========================================
Files 1951 1958 +7
Lines 192276 193428 +1152
Branches 15683 15794 +111
==========================================
+ Hits 162990 164039 +1049
- Misses 25321 25433 +112
+ Partials 3965 3956 -9
Flags with carried forward coverage won't be shown. Click here to find out more.
... and 3 files with indirect coverage changes 🚀 New features to boost your workflow:
|
This branch has not been deployed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why do we need this?
Evaluate scene-aware trajectory optimization as an opt-in alternative to RoboPlan. This draft is for offline/simulation review only: soft scene collision is not a physical safety barrier.
What is this feature?
Select
world_backend=curoboandplanner.backend=curobotogether. The fixed-base xArm7 grasp URDF is used with a modeled open gripper, fitted link spheres and shared adjacency/config/SRDF exclusions. The shared xArm model config supplies the rigidlink7/xarm_gripper_base_linkpair directly, without a dedicated SRDF. Existing exclusions and SRDF support for other models are unchanged. Pink still solves pose goals. V2plan_cspaceuses the world snapshot as a scene cost, not as a TrajOpt acceptance constraint; self-collision, limits and native success remain required. Native trajectory times and velocities go through existing materialization/execution. RoboPlan remains default. No changes to mapper, hardware or executor.How to manually run this
From the checkout on Linux x86-64 with a supported NVIDIA GPU and driver:
This pulls
nvidia-curobo[cu12-torch]v0.9.0 from its pinned Git source with CUDA 12.8 PyTorch. See the manipulation guide for prerequisites. Selectworld_backend=curoboandplanner.backend=curoboin the simulated xArm grasp blueprint; inspect planned results without execution. Cartesian segments and the full grasp workflow are not supported by this draft.Verification
744dd854d: fresh locked install on Linux x86-64 with Python 3.12 and RTX 3090 (driver 590.48.01), using PyTorch 2.7.1+cu128. The standalonecuroboextra also installed and imported successfully on Python 3.11 with CUDA available. Separate base andplanninginstalls without development groups contained no cuRobo, Torch, or CUDA libraries; this does not change existing PyTorch selections in other extras or test groups.744dd854d: 139 passed, 3 deselected in the focused CPU suite, including the real cuRobo scene/config parser, xArm blueprint, RoboPlan and SRDF/attachment regressions, and docs branding. The two xArm7 grasp compositions retained byte-identical prepared URDFs, all 15 collision links, and all 28 effective exclusions each versus4ed8ad0dd.744dd854don RTX 3090: the native xArm7 home self-check passed. Empty-scene planning produced 55 timed samples; the same-endpoint detour improved sampled clearance from −24.7 mm to +9.8 mm (not a continuous safety guarantee). Unavoidable scene overlap remained accepted; non-exempt self-collision and an out-of-limit goal were rejected. Pink → native planning → timed materialization preserved 37 samples. No executor or hardware driver was started.4ed8ad0dd: 387 passed, 31 deselected in the broader CPU/default-backend suite. The independent RTX 3090 qualification at that head also passed home, free-space planning, detour, soft overlap, self-collision and limits rejection, and Pink-to-timed materialization. That run used the earlier SRDF-backed mount declaration.Which issue(s) does this PR close?
None.
Checklist