IMPORTANT: To view this page as Markdown, append `.md` to the URL (e.g. /docs/manual/basics.md). For the complete Mojo documentation index, see llms.txt.
Skip to main content
Version: Nightly
For the complete Mojo documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /docs/manual/basics.md).

mma_apple

Apple Silicon MMA implementation for matrix multiply-accumulate operations.

This module provides two simdgroup_matrix MMA shapes:

  • 16x16x16 (_mma_apple): Apple M5 only (Metal 4.0 / AIR 2.8.0), float and integer-widening.
  • 8x8 (_mma_apple_8x8): all Apple GPU generations (M1-M5), float-only.

Supported operations:

  • Float multiply-accumulate (16x16): {F16, BF16, F32, E4M3, E5M2} inputs, F32 accumulator
  • Integer widening multiply-accumulate (16x16): {I8, U8} inputs, I32/U32 accumulator
  • Float multiply-accumulate (8x8): {F16, BF16, F32} inputs, F32 accumulator

Functions