Request for Documentation on Jetson Thor Safety MCU (STM32) Architecture and Communication

Request for official CycloneDX SBOM for JetPack 6.2 / L4T 36.4.3 Minimal RootFS

JetPack 7 / L4T R39.2: Request for FFmpeg 8.0.1 NVIDIA source package, build recipe, and redistribution clarification

## Category


- [ ] Report an error in the documentation.
- [x] Request for something to be documented.
- [x] Suggestion to improve the documentation.
- [ ] Other (please explain)

## Description
Adds an FP64 Warp FEM cantilever topology optimization example. The beam is clamped on the left and carries a distributed downward load at the free end. NLopt MMA minimizes compliance under a volume constraint using Solid Isotropic Material with Penalization (SIMP), a linear density filter, implicit-adjoint element energies, and Warp automatic differentiation.

The example uses Warp FEM’s existing graph-enabled conjugate-gradient path with its diagonal preconditioner. Its command line exposes only the device, grid dimensions, maximum evaluations, and headless mode; the remaining optimizer, solver, material, and problem settings use documented example defaults.

Warp FEM topology optimization example

Request for official R35.6.0 UEFI fix for AGX Xavier A/B slot switching failure

## Bug description

NumPy versions before 2.4 fail when an object exposes a zero-sized array through `__array_interface__` with a NULL data pointer. NumPy incorrectly sends the object through its scalar conversion path, which causes empty CPU Warp arrays to fail during Torch conversion.

NumPy tracks this bug in [numpy/numpy#26037](https://github.com/numpy/numpy/issues/26037) and fixed it in [numpy/numpy#29338](https://github.com/numpy/numpy/pull/29338).

This is a compatibility-hardening request for Warp, not a Warp regression. NumPy 2.4 already fixes the underlying bug, but environments using older NumPy versions still hit the exception below.

### Reproduction

```python
import warp as wp

array = wp.zeros((1, 0), dtype=wp.vec2f, device="cpu")
wp.to_torch(array)
```

### Actual behavior

```text
TypeError: float() argument must be a string or a real number, not 'array'

The above exception was the direct cause of the following exception:

ValueError: setting an array element with a sequence
```

### Expected behavior

`wp.to_torch()` returns an empty `torch.float32` tensor with shape `(1, 0, 2)`.

## System information

- Warp: 1.18.0.dev2
- NumPy: 2.3.3
- PyTorch: 2.10.0+cu128
- Python: 3.12.3
- Operating system: Linux 6.8.0-138-generic, x86_64, glibc 2.39

The same failure reproduces with NumPy 1.26.4.


Search — feature-request

Issues