DEPLOYMENT TASK: WSL2 CUDA LBM Daemon

Objective: Set up WSL2 with CUDA support and deploy the LBM-LLM real-time bridge.

Environment

Host: Windows 11 (Beast, 192.168.1.34)
GPU: RTX 4090
WSL: Already installed (Ubuntu)

Tasks

1. WSL2 CUDA Setup

# In WSL Ubuntu
wget https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-wsl-ubuntu.pin
sudo mv cuda-wsl-ubuntu.pin /etc/apt/preferences.d/cuda-repository-pin-600

wget https://developer.download.nvidia.com/compute/cuda/12.4.0/local_installers/cuda-repo-wsl-ubuntu-12-4-local_12.4.0-1_amd64.deb
sudo dpkg -i cuda-repo-wsl-ubuntu-12-4-local_12.4.0-1_amd64.deb
sudo cp /var/cuda-repo-wsl-ubuntu-12-4-local/cuda-*-keyring.gpg /usr/share/keyrings/

sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-4

2. Install Dependencies

sudo apt-get install -y build-essential libzmq3-dev libjson-c-dev

3. Compile LBM Daemon

Source file: D:\fractal-brain\beast-build\lbm_cuda_daemon.cu

Compile command:

cd /mnt/d/fractal-brain/beast-build
nvcc -o lbm_cuda_daemon lbm_cuda_daemon.cu -lzmq -ljson-c -O3 -arch=sm_89

4. Run Daemon

./lbm_cuda_daemon

Expected output:

Publishing on tcp://*:5555
Update rate: 100Hz
Metrics: coherence, h64, h32, asymmetry, vorticity, power_w

5. Verify Bridge

In Windows PowerShell:

cd D:\fractal-brain\beast-build
python lbm_llm_bridge.py

Should connect to tcp://localhost:5555 and receive LBM state.

Deliverables

Daemon running in WSL2, publishing to port 5555
Python bridge receiving data
Ollama integration working with physics-constrained inference
Log file: somatic_dataset.jsonl with [LBM State] + [Prompt] + [Output] triplets

Success Criteria

Kaelara can identify current vorticity/power without being told
Inference parameters (temperature, top_p) adjust based on LBM coherence
Real-time loop: LBM (100Hz) → Bridge → Ollama → Response

Start immediately. Report back when daemon is humming.

2.0 KiB Raw Blame History