Source code shows that buffer-manager-thread-count option is broken

[sunkenQ]

Description
I used the libtorch backend from the official tritonserver image, and I want to set buffer-manager-thread-count to speed up the operation of input and output tensors on cuda share memory. But I see the source code that it doesn't work, so how do I set it via backend config?

https://github.com/triton-inference-server/server/blob/1bc36c3cc286e6823339b5184c9ed058e85ed911/src/command_line_parser.h#L155C1-L158C43

Triton Information
24.10

Are you using the Triton container or did you build it yourself?

Using triton container