def run_stepped

bendtorch-core@0.1.0.0 · src/tensor/matmul_parallel_proof.bend:86

import Base
import bendtorch-core@0.1.0.0/src/tensor/matmul.bend as Matrix
import bendtorch-core@0.1.0.0/src/tensor/matrix_index_proof.bend as MatrixIndex
import bendtorch-core@0.1.0.0/src/tensor/elementwise_parallel_proof.bend as Parallel

86# The offsets of the next term are the stepped ones.
87def run_stepped(+rest: Nat, +dl: Nat, +dr: Nat, -A: Array<F32>, -B: Array<F32>, +inner: Nat, +columns: Nat,
88  +row: Nat, +column: Nat, +cursor: Nat, -value: F32, -l: Array<F32>, -r: Array<F32>,
89  run: Parallel.Fact<DotRun<1n+rest, dl, dr, A, B, inner, columns, row, column, 1n+cursor, value, Matrix.product_run(rest, 1n, columns, Nat.add(Nat.mul(row, inner), 1n+cursor), Nat.add(Nat.mul(1n+cursor, columns), column), value, Array.get(F32, l, Matrix.left_index(inner, row, 1n+cursor)), Array.get(F32, r, Matrix.right_index(columns, column, 1n+cursor)))>>) ->
90  Parallel.Fact<DotRun<1n+rest, dl, dr, A, B, inner, columns, row, column, 1n+cursor, value, Matrix.product_run(rest, 1n, columns, Nat.add(1n, Nat.add(Nat.mul(row, inner), cursor)), Nat.add(columns, Nat.add(Nat.mul(cursor, columns), column)), value, Array.get(F32, l, U32.from_nat(Nat.add(1n, Nat.add(Nat.mul(row, inner), cursor)))), Array.get(F32, r, U32.from_nat(Nat.add(columns, Nat.add(Nat.mul(cursor, columns), column)))))>>:
91  %Equal.sym(Nat, Nat.add(1n, Nat.add(Nat.mul(row, inner), cursor)), Nat.add(Nat.mul(row, inner), 1n+cursor), MatrixIndex.stepped_left(inner, row, cursor)) :
92    Parallel.Fact<DotRun<1n+rest, dl, dr, A, B, inner, columns, row, column, 1n+cursor, value, Matrix.product_run(rest, 1n, columns, _, Nat.add(columns, Nat.add(Nat.mul(cursor, columns), column)), value, Array.get(F32, l, U32.from_nat(_)), Array.get(F32, r, U32.from_nat(Nat.add(columns, Nat.add(Nat.mul(cursor, columns), column)))))>>
93  %Equal.sym(Nat, Nat.add(columns, Nat.add(Nat.mul(cursor, columns), column)), Nat.add(Nat.mul(1n+cursor, columns), column), MatrixIndex.stepped_right(columns, column, cursor)) :
94    Parallel.Fact<DotRun<1n+rest, dl, dr, A, B, inner, columns, row, column, 1n+cursor, value, Matrix.product_run(rest, 1n, columns, Nat.add(Nat.mul(row, inner), 1n+cursor), _, value, Array.get(F32, l, U32.from_nat(Nat.add(Nat.mul(row, inner), 1n+cursor))), Array.get(F32, r, U32.from_nat(_)))>>
95  run