IMPORTANT: To view this page as Markdown, append `.md` to the URL (e.g. /get-started.md). For the complete documentation index, see llms.txt.
Skip to main content
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).

Mojo function

ep_token_recv

def ep_token_recv[n_nodes: Int]() -> Bool

Returns whether the dispatch receives each token once per rank.

Within one node every rank stores straight into its peers' receive buffers, so a token is copied to a rank once and listed under each of its experts there. Across nodes SHMEM puts publish each expert through its own channel, so each routed pair keeps its own message slot.

Parameters:

  • ​n_nodes (Int): Number of physical nodes.

Returns:

Bool: The token_recv parameter of the dispatch kernels.

Was this page helpful?