cb.fabric-parity-gate

What problem does it solve?

Fabric parity gate ensures that the active CUDA/backend Fabric path matches the PyTorch Fabric reference for outputs, states, and gradients before performance closure, preventing subtle mismatches from slipping into production.

Core Features & Use Cases

• Verify parity across forward outputs, final state, input gradients, and parameter gradients on multiple backend paths.
• Build and validate a minimal parity matrix covering active routes, reset behavior, and materialization modes.
• Use during integration to certify readiness of fabrics/backend paths before rollout to production environments.

Quick Start

Run the Fabric parity gate at startup to verify outputs, states, and gradients across the active backend path before enabling performance measurements.