# Multi-LoRA Kernel Overhead Audit

## Description
Measures batched multi-adapter GPU kernel overhead (S-LoRA) versus single-model inference throughput.

## Category
`serving`

## Tags
`lora`, `s-lora`, `multi-tenant`, `adapters`, `kernel-overhead`

## When to Use
Audit multi-tenant LoRA serving engines serving dozens of concurrent fine-tuned adapters.

## Operational Limitations
Evaluates measured throughput compared to pure base model baseline.

## 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.
