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

enqueue_apple_grouped_matmul

def enqueue_apple_grouped_matmul(c: TileTensor[Engine=c.Engine, address_space=c.address_space, linear_idx_type=c.linear_idx_type], a: TileTensor[Engine=a.Engine, address_space=a.address_space, linear_idx_type=a.linear_idx_type], b: TileTensor[Engine=b.Engine, address_space=b.address_space, linear_idx_type=b.linear_idx_type], a_offsets: TileTensor[.uint32, Engine=a_offsets.Engine, address_space=a_offsets.address_space, linear_idx_type=a_offsets.linear_idx_type], expert_ids: TileTensor[.int32, Engine=expert_ids.Engine, address_space=expert_ids.address_space, linear_idx_type=expert_ids.linear_idx_type], max_num_tokens_per_expert: Int, num_active_experts: Int, ctx: DeviceContext)

Enqueues the Apple M5 grouped matmul C = A @ b[expert]^T.

Routes on the largest group: up to APPLE_GROUPED_GEMV_MAX_TOKENS tokens per expert takes the GEMV (decode), anything larger the MMA GEMM (prefill). b is [num_experts, N, K] with static N and K; the operand types are those of enqueue_apple_grouped_gemv.

Raises:

If the attached GPU is not Apple M5 (compute_capability == 5).