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

launch

def launch[Body: RowBody, //, axis: Int, simd_width: Int, supports_tiled: Bool = False, associative: Bool = False](body: Body, shape: Coord)

Top-level CPU scaffolder. Picks the tier from axis and supports_tiled, parallelizes over outputs (or output tiles), and invokes the body once per output (or tile).

Tier choice:

Inner axis (axis == rank - 1) — one worker per row, vectorize within the row at simd_width. emit_tile_width = 1.
Non-inner axis with supports_tiled — one worker per slice; each walks the innermost-non-axis dim in simd_width tiles, with a scalar tail for the remainder. emit_tile_width = simd_width.
Non-inner axis without supports_tiled — one worker per output, scalar reduce. emit_tile_width = 1.

Parameters:

axis (Int): Axis being reduced.
simd_width (Int): SIMD width for tile dispatch. Bodies compute it via rowwise.pick_simd_width[...] and pass it directly; the scaffolder no longer derives it from a dtype.
supports_tiled (Bool): Whether the body can run in the tiled tier (emit_tile_width > 1). Single-output reductions (reduce_sum, …) set this True; multi-output kernels (softmax, layernorm) set it False.
associative (Bool): Opt in the inner-axis cooperative tier's wide (_CPU_ILP_ACCUMULATORS * simd_width) same-dtype accumulator for rows >= _CPU_WIDE_ACC_MIN_ROW, breaking the serial accumulate dependency chain. Additive reductions (reduce_sum, reduce_mean) set this True; everything else keeps the native width.

Args:

body (Body): The per-row computation. Receives a Context and the row's Coord; uses reduce / pjoin / once / simd to compose the algorithm.
shape (Coord): Tensor shape.

Raises:

If a worker raises.