RAM Allocation Breakdown (GB)
Capacity: 32 GB
Model Weights
KV Cache Overhead
Docker Workloads
Free System Headroom
Memory Bus Saturation & Throughput
Bus: 38.4 GB/s
Inference Speed (tokens/s)
Background Bandwidth Loss
System Sizing & Bottleneck Diagnostics
Generated `docker-compose.yml` (Ollama Local Stack)
Production Ready
version: '3.8'
services:
ollama:
image: ollama/ollama:latest
container_name: local_ai_ollama
ports:
- "11434:11434"
environment:
- OLLAMA_NUM_PARALLEL=1
- OLLAMA_MAX_LOADED_MODELS=1
deploy:
resources:
limits:
memory: 26.50G
restart: unless-stopped