EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[Platform] Enable ARM-only CPU binding with NUMA-balanced A3 policy and update docs/tests (#6686)

Browse Source 
### What this PR does / why we need it?

- Keeps enable_cpu_binding default on, but skips binding on non‑ARM CPUs
inside bind_cpus, with a clear log.
- Uses a table-driven binding policy: A3 uses NUMA‑balanced binding;
other device types use NUMA‑affinity binding.
- Updates docs to reflect the exact behavior and adds/updates unit tests
for the new logic.

### Does this PR introduce _any_ user-facing change?

- Yes. CPU binding is now enabled by default via additional_config, and
documented in the user guide.
- CPU binding behavior differs by device type (A3 vs. others).

### How was this patch tested?

Added/updated unit tests:

test_cpu_binding.py
1.   test_binding_mode_table covers A2 vs A3 binding mode mapping.
2. test_build_cpu_pools_fallback_to_numa_balanced covers fallback when
affinity info is missing.
3. TestBindingSwitch.test_is_arm_cpu covers ARM/x86/unknown arch
detection.
4.   test_bind_cpus_skip_non_arm covers non‑ARM skip path in bind_cpus.

test_worker_v1.py
1. Updated mocks for enable_cpu_binding default True to align with new
config default.

- vLLM version: v0.14.1
- vLLM main: d7de043

---------

Signed-off-by: chenchuw886 <chenchuw@huawei.com>
Co-authored-by: chenchuw886 <chenchuw@huawei.com>
This commit is contained in:
Frank Chen
2026-02-25 11:15:14 +08:00
committed by GitHub
parent ac9a7d1301
commit 3da2ba22eb
6 changed files with 80 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
docs/source/user_guide/configuration/additional_config.md
View File

@@ -39,7 +39,7 @@ The following table lists additional configuration options available in vLLM Asc
| `multistream_overlap_shared_expert` | bool | `False` | Whether to enable multi-stream shared expert. This option only takes effect on MoE models with shared experts. |
| `multistream_overlap_gate` | bool | `False` | Whether to enable multi-stream overlap gate. This option only takes effect on MoE models with shared experts. |
| `recompute_scheduler_enable` | bool | `False` | Whether to enable recompute scheduler. |
| `enable_cpu_binding` | bool | `False` | Whether to enable CPU Binding. |
| `enable_cpu_binding` | bool | `True` | Whether to enable CPU binding. Only takes effect on ARM CPUs; when enabled, A3 uses NUMA-balanced binding strategy and other device types use NUMA-affinity's. |
| `SLO_limits_for_dynamic_batch` | int | `-1` | SLO limits for dynamic batch. This is new scheduler to support dynamic batch feature |
| `enable_npugraph_ex` | bool | `False` | Whether to enable npugraph_ex graph mode. |
| `pa_shape_list` | list | `[]` | The custom shape list of page attention ops. |