Dear Topaz Labs Support Team,
I know this is out of scope of your product but any advice would be apricated have user got VM to work before?
I am encountering issues while using Topaz Video AI on an AWS virtual machine, and I hope you can assist me.
Setup Details:
- AWS Instance Type: g5.xlarge
- Operating System: Windows Server (version as applicable)
- GPU: NVIDIA A10G with 24 GB VRAM
- NVIDIA Driver Version: 566.03
- CUDA Version: 12.7
- Topaz Video AI Version: [Please specify the version you’re using]
Issue Description:
I am attempting to upscale a 4K video to 8K using Topaz Video AI on this AWS instance. The application recognizes the GPU (confirmed via nvidia-smi), and the model loads successfully. However, when I start processing:
- The task remains stuck on “Queued” and does not progress.
- Frames do not process; there’s no movement or progress indicator.
- GPU utilization stays at 0% during the attempted processing.
- Attempting to process using the CPU yields the same result.
Steps I’ve Taken to Troubleshoot:
- Installed the latest NVIDIA drivers compatible with the A10G GPU.
- Installed CUDA Toolkit version 12.7.
- Verified GPU recognition and status using
nvidia-smi. - Confirmed that the GPU is selected as the AI Processor in Topaz Video AI settings.
- Attempted to run Topaz Video AI as an administrator (option not available).
- Ensured that the instance has sufficient resources (16 GB RAM, ample storage space).
- Tried processing different video files and formats, including lower resolutions.
- Disabled additional features like face refinement and stabilization.
- Reviewed application logs for errors (can provide logs if needed).
Additional Information:
- I noticed that the system requirements for Topaz Video AI specify an NVIDIA GTX 900 series or higher with at least 6 GB VRAM.
- Since the NVIDIA A10G is a data center GPU, I’m wondering if there are compatibility issues.
- I’m accessing the AWS instance via Remote Desktop Protocol (RDP).
Questions:
- Is Topaz Video AI compatible with the NVIDIA A10G GPU?
- Are there known issues when running the software on data center GPUs or in virtualized/cloud environments?
- Are there any recommended configurations or workarounds to get the software functioning in this setup?
Any guidance or assistance you can provide would be greatly appreciated.
