[Doc] Update mooncake nvlink transport doc for PD disaggregation (#11321)

Signed-off-by: Shangming Cai <csmthu@gmail.com>
This commit is contained in:
Shangming Cai
2025-10-08 15:59:29 +08:00
committed by GitHub
parent edefab0c64
commit 0a7c4bded7

View File

@@ -121,6 +121,14 @@ python -m sglang.launch_server \
PD Disaggregation with Mooncake supports the following environment variables for fine-grained control over system behavior.
#### NVLink Transport Configuration
To enable NVLink transport for KV cache transfers with the mooncake backend (recommended for NVL72 deployments), set the following environment variables. Note that auxiliary data transfer will still use TCP as a temporary workaround.
```bash
export SGLANG_MOONCAKE_CUSTOM_MEM_POOL=True
export MC_FORCE_MNNVL=True
```
#### Prefill Server Configuration
| Variable | Description | Default |
|:--------:|:-----------:|:--------: