Executive TL;DR:
- Using Nvidia GPU’s VRAM as swap space is a viable option for systems with limited RAM.
- This feature is particularly useful for laptops with soldered memory and no upgrade path.
- However, it may not be the best solution for all users due to potential performance issues.
The Internet’s Verdict: 60% Positive, 40% Negative
Introduction to Nvidia VRAM as Swap Space
Some users find this feature useful, especially when they have a large amount of VRAM sitting idle. As one user noted,
Built for laptops with soldered memory and no upgrade path. If you have an RTX card sitting there with 8GB of VRAM and you’re getting swapped to SSD, this puts that VRAM to work.
Performance Concerns
However, others have raised concerns about the performance impact of using VRAM as swap space. For example,
Sequential throughput: ~1.3 GB/s [on a RTX 3070 Laptop] This RTX 3070 chip is on PCIe 4.0 x16 which should give 64GB/s. The 8GB of GDDR6 is 448GB/s. Swapping to an NVMe drive would be twice as fast, but with higher latency.
Use Cases and Limitations
Some users find this feature useful for specific use cases, such as development work. As one user noted,
Given my dev machine has 32GB of RAM and 32GB of VRAM that sits mostly idle when I’m not running AI models, this is not that bad of an idea.
However, others have raised concerns about the potential for backpressure and crashes when using VRAM as swap space, particularly with certain desktop environments.
Focus Keyword: Nvidia VRAM
