hardware-profile / README.md
Doommsky's picture
Upload README.md with huggingface_hub
68ec5c5 verified
|
Raw History Blame Contribute Delete
1.96 kB
---
license: apache-2.0
tags:
- hardware
- nvidia-rtx-4050
- amd-radeon-740m
- 16gb-ram
- ollama
- docker
- windows-11
---
# Hardware Profile — Doommsky
## GPU / VRAM
| Component | VRAM | Type |
|-----------|------|------|
| NVIDIA GeForce RTX 4050 Laptop GPU | 6 GB (6141 MiB) | Dedicated VRAM |
| AMD Radeon 740M Graphics | ~15 GB (shared system RAM) | Integrated / Shared RAM |
| Total System RAM | ~16 GB | DDR5 |
## GPU Details
### NVIDIA GeForce RTX 4050 Laptop GPU
- **VRAM:** 6 GB (6141 MiB)
- **Driver:** 573.05 (32.0.15.7305)
- **CUDA:** Supported
- **Temperature:** 56°C (idle)
- **Power Draw:** 16.69 W (idle)
- **Best for:** AI inference (7B-8B models with 4-bit quantization), gaming, CUDA workloads
### AMD Radeon 740M Graphics
- **VRAM:** Up to ~15 GB (shared system RAM, configurable in BIOS)
- **Driver:** 32.0.13058.2
- **Type:** Integrated GPU (shares system RAM)
- **Best for:** Display output, light compute, fallback when NVIDIA is busy
## Software Environment
| Software | Status |
|----------|--------|
| **OS** | Windows 11 |
| **Ollama** | Running on localhost:11434 |
| **Docker Desktop** | Available |
| **Node.js** | Available |
| **Python** | 3.14 |
| **Git** | Available |
| **CUDA** | Supported (NVIDIA driver 573.05) |
## AI/ML Capabilities
- **Can run:** 7B-8B parameter models with 4-bit quantization on GPU (RTX 4050)
- **Can run (CPU):** 13B+ models on CPU + shared RAM (AMD iGPU can assist)
- **Local inference:** Ollama on localhost:11434
- **Containerized ML:** Docker available
- **Hugging Face Hub:** Authenticated (hf CLI 2.4.0)
## Notes
- The AMD Radeon 740M shares system RAM, so VRAM is limited by total system memory (~16 GB)
- For best AI performance, use the NVIDIA RTX 4050 for GPU-accelerated inference
- The integrated AMD GPU can be used for display output while the NVIDIA GPU handles compute
- Ollama is configured to use the NVIDIA GPU for inference
## Last Updated
2026-11-17