For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Mojo module

grouped_matmul

Implements grouped (batched expert) GEMM for SM90 (Hopper) GPUs.

Provides the grouped_matmul_sm90 entry point, which dispatches a variable-length batched matmul across MoE expert weight matrices using TMA-based warp-specialized pipelining via HopperMatmulSM90Kernel.

Functions

• ​default_config_sm90: Returns the default SM90 matmul config for the given WGMMA shape.
• ​grouped_matmul_sm90: Performs grouped GEMM for MoE routing on SM90 (Hopper) GPUs.

Was this page helpful?

Thank you! We'll create more content like this.

Thank you for helping us improve!

😔 What went wrong?

Some code doesn’t work

It includes inaccurate information

It's missing information I need

It was difficult to understand

Other

Submit