Graphics Cards

Table of Contents

• Graphics Cards

  • Linux Support

  • Recommended VRAM Size

  • History

  • Bibliography

Linux Support

Here are various rankings for each major graphics cards vendor on Linux.

Stable drivers:

1. Intel

2. NVIDIA

3. AMD

Linux performance:

1. AMD

2. NVIDIA

3. Intel

Open source and native Linux support:

1. Intel

2. AMD

3. NVIDIA

Hybrid graphics support on laptops:

1. NVIDIA

2. AMD

Day 1 Linux support for new hardware:

1. Intel

2. NVIDIA

3. AMD

Recommended VRAM Size

Depending on the use case, these are the recommended minimum size of VRAM to have on the graphics card.

Gaming [1][2]:

Game Resolution	VRAM (GB)
720p	2
1080p	4
1440p	6
4K	8

Video editing [2][3]:

Video Resolution	VRAM (GB)
1080p	4
4K	6
6K	8
8K	10

Machine learning (ML) [4][5]:

ML Workload	VRAM (GB)
Minimum	8
Recommended	24
Maximum	48

• The amount of VRAM differs based on the type of ML model. Language models use the most amount of VRAM compared to image models which are more CPU-bound rather than GPU-bound. [5]

History

• Latest

Bibliography

1. “How Much VRAM Do You Need For Gaming?” GPU Mag. November 2, 2021. Accessed September 5, 2022. https://www.gpumag.com/how-much-vram-gaming/

2. “How much VRAM do you need? Professional and Gaming Workloads explored.” CG Director. May 9, 2022. Accessed September 5, 2022. https://www.cgdirector.com/how-much-vram-do-you-need/

3. “Hardware Recommendations for Premiere Pro.” Puget Systems. Accessed September 5, 2022. https://www.pugetsystems.com/recommended/Recommended-Systems-for-Adobe-Premiere-Pro-143/Hardware-Recommendations

4. “Hardware Recommendations for Machine Learning & AI.” Puget Systems. Accessed September 7, 2022. https://www.pugetsystems.com/recommended/Recommended-Systems-for-Machine-Learning-AI-174/Hardware-Recommendations

5. “Choosing the Best GPU for Deep Learning in 2020.” The Lambda Deep Learning Blog. February 18, 2022. Accessed September 7, 2022. https://lambdalabs.com/blog/choosing-a-gpu-for-deep-learning/