WHATS NEW IN v1.6 (2026-07-13)
- v3 proofs: dummy prompts now carry the /nothink token (dev guidance) —
  avoids low-entropy thinking traces and the extra argon fallback = better
  proof throughput after block 8000. Update recommended for everyone.

aria-worker-tsc v1.6 (2026-07-12)
v1.6 : worker protege (code compile en binaire, non lisible) — meme moteur et memes preuves que v1.4. Miner fee 5% inchangee (annoncee au boot). Pool fee 5%. Rien d autre ne change cote minage.

v1.4 : MAJ OBLIGATOIRE avant le bloc 8000 (activation consensus v3 prompt binding). Images rebuild TensorCash v1.0.3. llama-backend desormais multi-arch (SM50->SM121 : PTX virtual + 86/89 reels + Blackwell 120a natif, CUDA 12.8). Miner fee 5% inchangee (annoncee au boot).
- MINER FEE 5% (announced): one 5-min round every 100 min mines to the package maintainer wallet (announced at worker startup). Pool fee 5% applies separately.
- includes v1.2 fix (LLAMA_CPP + STANDALONE_MODE)
════════════════════════════════════════════════════════════════════
  ARIAPOOL — TensorCash (TSC) community worker
  Proof-of-inference mining: your GPU runs the real Qwen3-8B model.
════════════════════════════════════════════════════════════════════

WHAT YOU NEED
  • Linux x86-64
  • An NVIDIA GPU with 24 GB VRAM  — RTX 5090 (Blackwell) or RTX 4090 (Ada).
    (16 GB is not enough: the model runs in bf16 ≈ 15 GB + working memory.)
  • Docker + the NVIDIA Container Toolkit. Install on Ubuntu/Debian:
        curl -fsSL https://get.docker.com | sh
        # NVIDIA toolkit:
        curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey \
          | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg
        curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list \
          | sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' \
          | sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
        sudo apt-get update && sudo apt-get install -y nvidia-container-toolkit
        sudo nvidia-ctk runtime configure --runtime=docker && sudo systemctl restart docker
  • ~40 GB free disk (images + model, downloaded once).

GET AN ADDRESS
   You need a TSC payout address (starts with tc1q…) — this is where you
   get paid. Use an address from your own TSC wallet or an exchange that
   supports TSC. Keep it safe: lose it = lose your coins.

START MINING
        ./run.sh                  # asks for your address, then launches
   First run downloads the model (~15 GB) and images (~8 GB). One time only.

WATCH IT
        docker compose logs -f miner-proxy      # worker connecting + mining
        docker compose logs -f llama-backend    # the GPU engine
   Your live stats (rate, shares, what you're owed, payouts):
        https://pool.ariabrain.com/tsc.html     (paste your address in "My stats")

HOW YOU GET PAID
   Every completed inference = 1 share, credited to your address.
   When the pool finds a block, 95 % of the reward is split across all miners
   by shares over the last hour (PPLNS). Pool fee: 5 %. Auto-payout once you
   reach 1 TSC, sent straight to your address.

STOP
        docker compose down

No ports to open. The worker connects OUT to the pool over wss:// (443).

Questions:  https://discord.gg/DUHmqnEr2E   ·   AriaPool — powered by ARIABRAIN
