Add Docker-based build system with GPU-enabled builder and runtime containers

docker/runtime/Dockerfile

@@ -0,0 +1,56 @@
+# Dockerfile for Ollama37 Runtime Image
+#
+# This creates a minimal runtime-only image from pre-built artifacts.
+# The artifacts should be built first using the Makefile in ../
+#
+# Build context should be the project root (../..) so we can access docker/output/
+#
+# Usage:
+#   cd /path/to/ollama37
+#   make -C docker build           # Build the binary first
+#   make -C docker build-runtime   # Create the runtime image
+
+FROM nvidia/cuda:11.4.3-runtime-rockylinux8
+
+# Install minimal runtime dependencies
+# Note: Not running 'dnf update' to keep base image stable and build faster
+RUN dnf -y install \
+    libgomp \
+    libstdc++ \
+    && dnf clean all \
+    && rm -rf /var/cache/dnf
+
+# Create directory structure
+RUN mkdir -p /usr/local/bin /usr/local/lib/ollama
+
+# Copy the ollama binary from build output
+COPY docker/output/ollama /usr/local/bin/ollama
+
+# Copy all shared libraries from build output
+COPY docker/output/lib/ /usr/local/lib/ollama/
+
+# Prepend our library path to the existing LD_LIBRARY_PATH
+# Base image already has: /usr/local/nvidia/lib:/usr/local/nvidia/lib64
+ENV LD_LIBRARY_PATH="/usr/local/lib/ollama:${LD_LIBRARY_PATH}"
+
+# Base image already sets these, but we can override if needed:
+# NVIDIA_DRIVER_CAPABILITIES=compute,utility
+# NVIDIA_VISIBLE_DEVICES=all
+
+# Ollama server configuration
+ENV OLLAMA_HOST=0.0.0.0:11434
+
+# Expose the Ollama API port
+EXPOSE 11434
+
+# Create a data directory for models
+RUN mkdir -p /root/.ollama
+VOLUME ["/root/.ollama"]
+
+# Health check
+HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
+    CMD /usr/local/bin/ollama list || exit 1
+
+# Set entrypoint and default command
+ENTRYPOINT ["/usr/local/bin/ollama"]
+CMD ["serve"]