ikarus/benchmarks.larceny/src/ntakl.sml

42 lines
1.1 KiB
Standard ML

(* TAKL -- The TAKeuchi function using lists as counters. *)
fun list_n n =
let fun loop (n, a) =
if n = 0
then a
else loop (n-1, n::a)
in loop (n, [])
end
val l18 = list_n 18
val l12 = list_n 12
val l6 = list_n 6
(* Part of the fun of this benchmark is seeing how well the compiler
can understand this ridiculous code, which dates back to the original
Common Lisp. So it probably isn't a good idea to improve upon it. *)
fun shorterp (x, y) =
not (y = [])
andalso (x = [] orelse shorterp (tl x, tl y))
(* But SML/NJ runs this benchmark about 15 times as fast when the
code above is rewritten as follows. So I did it. Pooh. *)
fun shorterp (_, []) = false
| shorterp ([], _) = true
| shorterp (_::x, _::y) = shorterp (x, y)
fun mas (x, y, z) =
if not (shorterp (y, x))
then z
else mas (mas (tl x, y, z),
mas (tl y, z, x),
mas (tl z, x, y))
fun takl_benchmark (n) =
run_benchmark ("takl", n, fn () => mas (l18, l12, l6),
fn (x) => length(x) = 7)
fun main () = takl_benchmark (takl_iters)