# MoE Activation Memory Footprint Audit

## Description
Calculates active versus total parameter memory footprint and expert weight swapping overhead on constrained GPU nodes.

## Category
`serving`

## Tags
`moe`, `memory`, `vram`, `footprint`, `expert-cache`

## When to Use
Size VRAM allocations for MoE models when only a subset of experts fit into high-speed memory.

## Operational Limitations
Deterministic memory sizing. Assumes steady-state layer distribution.

## Verification Script
Run the deterministic verification checker:
```bash
python3 scripts/check.py examples/synthetic.json
```
Exit codes:
- `0`: Check passed and thresholds satisfied.
- `1`: Threshold breach or operational violation detected.
- `2`: Schema error or invalid usage.