Skip to main content

​
Synopsis

llmfit system [OPTIONS]

​
Description

Detects and displays your system’s hardware specifications including RAM, CPU, GPU, and available backends for LLM inference. This command is useful for verifying what llmfit has detected about your system and troubleshooting detection issues.

​
Options

​
--json
boolean
default:"false"
Output system specs as JSON instead of formatted text.
​
--memory
string
Override GPU VRAM size (e.g., “32G”, “32000M”, “1.5T”). Useful when autodetection fails or for testing different configurations.

​
Usage Examples

​
Basic System Info

# Display system specifications
llmfit system

​
JSON Output

# Get system specs as JSON
llmfit system --json

​
Override GPU Memory

# Test with different VRAM sizes
llmfit system --memory 24G
llmfit system --memory 80G

​
Example Output

​
Human-Readable Format

╭─ System Hardware ──────────────────────────────────────────╮
│                                                            │
│  RAM:  64.0 GB total (58.2 GB available)                  │
│  CPU:  16 cores (Apple M2 Max)                            │
│  GPU:  Metal                                               │
│        - Apple M2 Max                                      │
│        - 64.0 GB VRAM                                      │
│        - Unified Memory                                    │
│                                                            │
╰────────────────────────────────────────────────────────────╯

​
JSON Format

{
  "system": {
    "total_ram_gb": 64.0,
    "available_ram_gb": 58.24,
    "cpu_cores": 16,
    "cpu_name": "Apple M2 Max",
    "has_gpu": true,
    "gpu_vram_gb": 64.0,
    "gpu_name": "Apple M2 Max",
    "gpu_count": 1,
    "unified_memory": true,
    "backend": "Metal",
    "gpus": [
      {
        "name": "Apple M2 Max",
        "vram_gb": 64.0,
        "backend": "Metal",
        "count": 1,
        "unified_memory": true
      }
    ]
  }
}

​
System Detection

llmfit automatically detects:
  • RAM: Total and available system memory via sysinfo crate
  • CPU: Core count and model name
  • GPU: NVIDIA (via nvidia-smi), AMD (via rocm-smi), Apple Silicon (via system_profiler)
  • Backend: CUDA, ROCm, Metal, or CPU-only
  • Unified Memory: Apple Silicon systems where VRAM = RAM

​
Platform Notes

​
NVIDIA

Requires nvidia-smi in PATH. Detects CUDA-capable GPUs and VRAM. 
# Verify nvidia-smi works
nvidia-smi

​
AMD

Requires rocm-smi in PATH. Detects ROCm-capable GPUs. 
# Verify rocm-smi works
rocm-smi

​
Apple Silicon

Uses system_profiler to detect Metal backend. VRAM equals system RAM due to unified memory architecture.

​
CPU-Only

If no GPU is detected, llmfit will use CPU inference mode for model recommendations.
  • llmfit - Launch TUI with detected specs
  • fit - Find models matching your system
  • plan - Plan hardware requirements for a model

Build docs developers (and LLMs) love

Get started for freeTalk to us