[Refactor] Refactor DeepEP dispatcher (#22822)

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Cheng Wan <54331508+ch-wan@users.noreply.github.com>
This commit is contained in:
Артем Савкин
2026-05-18 04:36:42 +03:00
committed by GitHub
co-authored by gemini-code-assist[bot] Cheng Wan
parent 5147de26e4
commit a080358cac
30 changed files with 302 additions and 182 deletions
@@ -1472,6 +1472,12 @@ Please consult the documentation below and [server_args.py](https://github.com/s
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.02)"}}><code>auto</code></td>
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}><code>normal</code>, <code>low_latency</code>, <code>auto</code></td>
</tr>
<tr>
<td style={{padding: "9px 12px", fontWeight: 500, backgroundColor: "rgba(255,255,255,0.02)"}}><code>--deepep-dispatcher-output-dtype</code></td>
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}>Select DeepEP dispather output dtype, could be <code>bf16</code>, <code>fp8</code>, <code>int8</code> (only Ascend A2/A3 NPU), <code>nvfp4</code> or <code>auto</code>. Default is <code>auto</code>, which follows a priority order (server argument → deprecated env var → input_global_scale check → dispatcher_output_dtype from quant_config → flashinfer/cutlass backend → NPU BF16 default → GPU FP8 default).</td>
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.02)"}}><code>auto</code></td>
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}><code>bf16</code>, <code>fp8</code>, <code>int8</code>, <code>nvfp4</code>, <code>auto</code></td>
</tr>
<tr>
<td style={{padding: "9px 12px", fontWeight: 500, backgroundColor: "rgba(255,255,255,0.02)"}}><code>--ep-num-redundant-experts</code></td>
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}>Allocate this number of redundant experts in expert parallel.</td>