[FIX] Ensure tensors consistently move to backend device and add tests for GPU-specific behavior

[Leguark]

[FIX] Ensure tensors consistently move to backend device and add tests for GPU-specific behavior

Aligns tensor device in _repeat and _array helpers to backend settings. Adds regression tests to validate PyTorch GPU device handling and tensor movement in preprocessing and repeated operations.

[ENH] Add concurrency safety for mutable options and enhance octree-level handling

Introduces thread-safe handling of mutable option states to prevent collisions during concurrent computations. Adds parameterized octree-level support in tests and refactors option initialization for request-level isolation. Includes validations for stacked kernel dimension mismatches and updates relevant symbolic evaluator logic.

[Leguark]

• [ENH] Add finite-fault support to server schema, validation, and PyTorch backend #94
• [ENH] Wire finite-fault definitions into fault stack interpolation #93
• [ENH] Refactor project_points_onto_surface for robustness and nonlinear projection support #92
• [ENH] Introduce FiniteFault model with validation, serialization, and spline taper support #91
• [FIX] Ensure tensors consistently move to backend device and add tests for GPU-specific behavior #90 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.