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 package

cpu

Provides the CPU backend implementations for matmuls.

Modules

• ​apple_accelerate: Wraps the Apple Accelerate cblas_sgemm routine to provide single-precision matmul, GEMV, and batched matmul kernels for Apple CPUs.
• ​default: Provides the generic CPU matmul microkernel used as the fallback path.
• ​i8mm: Implements the ARM I8MM (8-bit integer matrix multiply) CPU microkernel.
• ​impl: Implements tiled CPU matrix multiplication kernels and their dispatcher.
• ​neon: Implements the ARM NEON lane-FMA microkernel for CPU matmul.
• ​vnni: Implements the integer CPU matmul microkernel (x86 VNNI or ARM NEON).