Bendtorch.attn_tree: def of bendtorch@1.3.0.0
Import: import bendtorch@1.3.0.0/main.bend as Bendtorch
Lines 2916-2924 of bendtorch@1.3.0.0/main.bend: https://jend.leeeet.dev/read.txt?path=bendtorch@1.3.0.0/main.bend&start_line=2916&end_line=2924

import Base

# whether the units (a sequence's key/value group each, fslices) run in the tree: from 8 units on, or from 2 when each is
# big. With the tiny Llama's sequences (t = 16, w = 32) the tree cost more than it gave at 2 sequences (0.48 against
# 0.37 ms on 16 threads) and gained from 8 on (1.32 against 1.52 ms; 1.5x at 32, 1.7x at 128); with SmolLM2's (t = 128,
# w = 576, about half a second of a step each) the two sequences one after another kept a single thread busy for a
# third of the step (NOTES, 2026-10-21), and its 2 sequences x 3 groups now give 6 units. t w counts in U32 (a Nat
# comparison walks its numbers one unit at a time). The answer never changes the result: the laws attn_par_fwd and
# attn_par_bwd show the tree gives the units in order, as one after another.
def attn_tree(+b: Nat, +t: Nat, +w: Nat) -> Bool:
  Bool.or(Nat.is_le(8n, b), Bool.and(Nat.is_le(2n, b), U32.is_le(4096, (U32.from_nat(t) * U32.from_nat(w) : U32))))

Uses (9):
  https://jend.leeeet.dev/definition.txt?path=Base/base.bend&name=Bool
  https://jend.leeeet.dev/definition.txt?path=Base/base.bend&name=Bool.and
  https://jend.leeeet.dev/definition.txt?path=Base/base.bend&name=Bool.or
  https://jend.leeeet.dev/definition.txt?path=Base/base.bend&name=Nat
  https://jend.leeeet.dev/definition.txt?path=Base/base.bend&name=Nat.is_le
  https://jend.leeeet.dev/definition.txt?path=Base/base.bend&name=U32
  https://jend.leeeet.dev/definition.txt?path=Base/base.bend&name=U32.from_nat
  https://jend.leeeet.dev/definition.txt?path=Base/base.bend&name=U32.is_le
  https://jend.leeeet.dev/definition.txt?path=Base/base.bend&name=U32.mul