TURBOCHARGE_NOW.sh

#!/bin/bash
# ╔══════════════════════════════════════════════════════════╗
# ║   🚀 ONE-CLICK TURBOCHARGE FOR QWEN2GOLEM 🚀            ║
# ║   Press this button to make EVERYTHING LIGHTNING FAST!   ║
# ╚══════════════════════════════════════════════════════════╝

echo ""
echo "  ⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡"
echo "  🔥 INITIATING ULTIMATE TURBOCHARGE SEQUENCE 🔥"
echo "  ⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡⚡"
echo ""
echo "  Target System: RTX 3050 6GB + i5 CPU + 16GB RAM"
echo "  Mission: ACHIEVE LIGHTNING SPEED WITHOUT COMPROMISE!"
echo ""

# Countdown for dramatic effect
echo "  Launching in..."
for i in 3 2 1; do
    echo "  $i..."
    sleep 1
done
echo "  🚀 BLAST OFF!"
echo ""

# Check if we need to install first
if [ ! -f "golem_optimizer.py" ] || [ ! -f "voice_optimizer.py" ]; then
    echo "⚠️  Optimization files not found. Running installer first..."
    if [ -f "install_optimizations.sh" ]; then
        chmod +x install_optimizations.sh
        ./install_optimizations.sh
    else
        echo "❌ Installation script not found! Please ensure all files are present."
        exit 1
    fi
fi

# Start Redis if not running
echo "🗄️ Checking Redis cache..."
if ! pgrep -x "redis-server" > /dev/null; then
    echo "   Starting Redis server (user mode)..."
    redis-server --daemonize yes || true
fi

# Clear GPU cache
echo "🎮 Preparing GPU..."
python -c "
import torch
if torch.cuda.is_available():
    torch.cuda.empty_cache()
    torch.cuda.synchronize()
    print('   ✅ GPU cache cleared')
else:
    print('   ⚠️ GPU not available')
"

# Refresh Gemini keys
echo "🔑 Refreshing API keys..."
if [ -f "refresh_gemini_keys.sh" ]; then
    ./refresh_gemini_keys.sh > /dev/null 2>&1 &
    echo "   ✅ Key refresh running in background"
fi

# Run the main optimizer
echo ""
echo "⚡ APPLYING OPTIMIZATIONS..."
python golem_optimizer.py

# Start the optimized server
echo ""
echo "🚀 STARTING TURBOCHARGED SERVER..."
echo ""

# Set environment variables for maximum performance
export CUDA_VISIBLE_DEVICES=0
export PYTORCH_CUDA_ALLOC_CONF=max_split_size_mb:512
export CUDA_LAUNCH_BLOCKING=0
export TORCH_CUDNN_V8_API_ENABLED=1
export TF32_ENABLE=1
export CUBLAS_WORKSPACE_CONFIG=:4096:8

# Check if gunicorn is available
if command -v gunicorn &> /dev/null; then
    echo "✅ Starting with Gunicorn (optimal performance)..."
    gunicorn home.chezy.golem_flask_server:app \
        --workers 4 \
        --worker-class gevent \
        --worker-connections 1000 \
        --bind 0.0.0.0:5000 \
        --timeout 30 \
        --keep-alive 5 \
        --max-requests 10000 \
        --max-requests-jitter 1000 \
        --log-level info
else
    echo "⚠️  Gunicorn not found, starting with Flask development server..."
    echo "   (Install gunicorn and gevent for better performance)"
    # Free port 5000 if busy
    if lsof -i :5000 -t >/dev/null 2>&1; then
        echo "   Port 5000 busy; stopping old process..."
        kill -9 $(lsof -i :5000 -t) || true
        sleep 1
    fi
    cd home/chezy/
    python golem_flask_server.py
fi