Code 504

The `code 504` error indicates that the GPU on the rendering server does not meet the requirements for LarkXR cloud rendering. This is commonly caused by an unsupported GPU model, incorrect driver configuration, or insufficient hardware encoding channel capacity.

## Cause 1: The GPU Does Not Meet Configuration Requirements

The rendering server's graphics card may not satisfy the required model support, driver version, or concurrency specifications for LarkXR.

### Solution

Verify that the GPU meets all LarkXR hardware requirements, including supported GPU models, driver versions, and concurrent session limits. Refer to the official hardware requirements documentation for details:

[LarkXR GPU Requirements](https://www.pingxingyun.com/online/sr3_2.html?id=384)

## Cause 2: Insufficient Hardware Encoding Channels on GeForce GPUs

GeForce-series GPUs have a limited number of hardware encoding channels due to NVIDIA's own restrictions. Remote management tools such as Sunlogin or other similar software also consume encoding channels, further reducing the number of available concurrent sessions. Depending on the GPU model, GeForce cards typically support only 2 to 3 concurrent hardware encoding sessions. NVIDIA drivers version 531.61 and later support up to five concurrent sessions.

### Solution

Avoid using GeForce-series GPUs for LarkXR rendering deployments. Use a professional-grade GPU (such as an NVIDIA Quadro or Tesla series), or use a LarkXR hardware-software integrated machine that is pre-configured for cloud rendering workloads.

Reference:
[NVIDIA Video Encode and Decode GPU Support Matrix](https://developer.nvidia.com/video-encode-and-decode-gpu-support-matrix-new)

## Notes

- GeForce GPUs are not recommended for production cloud rendering environments due to their hardware encoding channel limitations.
- If remote management tools are in use, account for the encoding channels they consume when calculating available concurrency.
- Always verify driver version compatibility before deployment.