For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).
Python function
dispatch_start
dispatch_start()
max.nn.ep.dispatch_start(config, buffers, tokens, ids, *, input_scales=None)
Launches each device’s dispatch; dispatch_finish() waits for it.
-
Parameters:
-
- config (EPConfig) – The expert-parallel configuration.
- buffers (EPBuffers) – The exchange’s buffers.
- tokens (Tensor) – Each device’s tokens, as for
dispatch(). - ids (Tensor) – Each device’s top-k expert ids, as for
dispatch(). - input_scales (Tensor | None) – As for
dispatch().
-
Raises:
-
ValueError – If
check_supported()rejectsconfig, or ifconfigquantizes NVFP4 with per-token global scales, which onlydispatch()sends. -
Return type:
-
None