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 package

linalg

Provides CPU and GPU implementations of linear algebra functions.

Packages

• ​arch: Provides architecture specific utility functions.
• ​matmul: Provides the backend implementation for matmuls.

Modules

• ​accumulate:
• ​bmm:
• ​fp4_quantization:
• ​fp4_utils:
• ​fp8_quantization:
• ​fp8_utils: Re-exports FP8 quantization utilities from comm.fp8_utils.
• ​gemv:
• ​grouped_matmul:
• ​grouped_matmul_block_scaled_dispatch: General dispatch for grouped block-scaled matmul.
• ​grouped_matmul_sm100:
• ​grouped_matmul_sm100_1d1d:
• ​grouped_matmul_sm100_blockwise_fp8:
• ​grouped_matmul_tile_scheduler:
• ​lora:
• ​matrix_band_part: The module implements matrix band part functions.
• ​mxfp4_dequant: MXFP4 dequantization kernel for H100 (SM90).
• ​mxfp4_matmul_sm90: MXFP4 matmul on H100 (SM90) via dequant-to-FP8 + FP8 GEMM.
• ​packing:
• ​qr_factorization:
• ​structuring:
• ​transpose: The module implements Transpose functions.
• ​utils:
• ​utils_gpu:

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