[Docs] Remove references to removed env vars (#59530)
Signed-off-by: Sundri Lai <laiyanting.neu@gmail.com>
Sundri Lai committed
Oct 1, 2026 at 01:26 UTC
bcee730b1a9d25f0fd283a0ef6c19133ebeebf4f
2 files changed
+4
-8
docs/features/engram.md
+3
-6
@@ -83,10 +83,7 @@ directly from the Triton kernel through unified virtual addressing (UVA),
83
prefetching on a side CUDA stream so transfers overlap with the compute of
84
preceding layers. The freed GPU memory goes to the KV cache.
85
86
-`cpu_offload` defaults to `true` (on): it follows the
87
-`VLLM_PLE_CPU_OFFLOAD` environment variable, which also defaults to `true`.
88
-An explicit `--engram-config` value takes precedence over the environment
89
-variable.
86
+`cpu_offload` defaults to `true` (on)
87
88
To keep the tables resident on GPU instead (e.g. when host memory is
89
scarce):
@@ -96,8 +93,8 @@ vllm serve deepseek-ai/DeepSeek-V4.1-Flash \
93
--engram-config.cpu_offload false
94
```
95
99
-or equivalently `VLLM_PLE_CPU_OFFLOAD=0`. CPU offload requires a GPU with UVA
100
-support; vLLM fails fast when it is unavailable.
96
+CPU offload requires a GPU with UVA support;
97
+vLLM fails fast when it is unavailable.
98
99
## Data-parallel topologies (DeepSeek V4.1)
100
docs/serving/online_serving/renderer.md
+1
-2
@@ -51,8 +51,7 @@ and provision transport limits and memory accordingly.
51
For example, start a standard inference server with scale-out endpoints enabled:
52
53
```bash
54
-VLLM_ENABLE_SCALE_OUT_ENDPOINTS=1 \
55
- vllm serve meta-llama/Llama-3.1-8B-Instruct
54
+vllm serve meta-llama/Llama-3.1-8B-Instruct --enable-scale-out
55
```
56
57
Send a Responses request and use `jq` to extract the `token_ids` field: