"Use this quip-miner sub-skill for CPU, CUDA, Metal, Modal, and QPU
Scanned 9/9/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill mining-backends --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Mining Backends?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-mining-backends)More formats (shields.io, HTML) on the badges page.
---
name: mining-backends
description: "Use this quip-miner sub-skill for CPU, CUDA, Metal, Modal, and QPU
mining commands, backend dependencies, QPU budgets, unified streaming, and
PoW/mempool scheduling."
disable-model-invocation: true
metadata:
disco-role: operating
license: AGPL 3.0
---
# Mining Backends
Use this sub-skill when the user needs to run, tune, or troubleshoot quip-miner CPU/GPU/QPU mining backends or understand PoW plus mempool scheduling on those workers.
## Route By Task
- **Backend commands and tuning:** Read `references/backend-workflows.md` for CPU SA, CUDA, Metal, Modal, D-Wave QPU, gate-model QPU, optional dependencies, and safe command patterns.
- **Streaming and mempool architecture:** Read `references/unified-streaming-and-mempool.md` for the stream-driver subprocess, shared-memory ring, feeder specs, PoW/mempool preemption, and no-inline-sampling invariant.
- **Backend failures:** Read `references/troubleshooting.md` for CUDA allocation, Metal host mismatch, Modal unavailable, QPU credential/budget, topology binding, and mempool-owner problems.
- **Availability probe:** Run `scripts/backend_probe.py --json` or the root `scripts/quip_backend_probe.py --json` before claiming backend availability.
## Common Commands
```bash
quip-miner cpu --validator ws://127.0.0.1:9944 --num-cpus 4 --signer-key ~/.quip-miner/signing.json
quip-miner gpu --validator ws://127.0.0.1:9944 --gpu-backend local --signer-key ~/.quip-miner/signing.json
quip-miner gpu --validator ws://127.0.0.1:9944 --gpu-backend metal --signer-key ~/.quip-miner/signing.json
quip-miner gpu --validator ws://127.0.0.1:9944 --gpu-backend modal --signer-key ~/.quip-miner/signing.json
quip-miner qpu --validator ws://127.0.0.1:9944 --qpu-type dwave --daily-budget 30s --signer-key ~/.quip-miner/signing.json
```
Prefer config-driven production runs (`quip-miner --config config.toml`) once the backend inventory is known.
## Key Rules
- Every worker mines PoW continuously. Mempool jobs preempt PoW on the same workers when the group owns mempool; PoW resumes afterward.
- CPU/GPU mempool defaults on; QPU mempool defaults off and must be explicitly opted into in the vendor section.
- A live miner pulls topology from the chain. Passing `--topology` to live mining commands is an error.
- CUDA guidance requires actual CUDA runtime evidence; a CPU import is not enough.
- Metal has no CPU fallback. On non-Apple hosts, provide config/docs guidance only.
- Do not run paid QPU sampling or Modal cloud jobs unless the operator explicitly accepts credentials, network, cost, and runtime.
## Boundaries
- Route TOML schema and supervisor deployment to `../config-supervisor-deployment/SKILL.md`.
- Route wallet, bootstrap, descriptor, and solver registration to `../identity-wallet-bootstrap/SKILL.md`.
- Route telemetry and attempts archives to `../telemetry-attempt-archive/SKILL.md`.
- Route topology/proof replay and D-Wave h/J range details to `../topology-proof-validation/SKILL.md`.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!