ocamlformat

Replace the pure-OCaml FNV-1 implementation in CCInt64 with C stubs
that call the same cc_fnv_hash_int64 core as CCInt.hash:

- hash: int64 -> int via caml_cc_hash_int64 ([@unboxed] + [@@noalloc])
- hash_to_int64: int64 -> int64 via new caml_cc_hash_int64_to_int64
  entry points (masks result to non-negative, needs caml_copy_int64
  for bytecode so cannot be [@@noalloc])

Add 18 tests in t_int64.ml covering:
- Non-negativity of hash and hash_to_int64
- Consistency between hash and hash_to_int64
- Different inputs produce different hashes
- Determinism
- QuickCheck: hash is non-negative for random inputs

benchs/dune

@@ -10,7 +10,6 @@
   qcheck
   oseq
   batteries
-  base
   sek)
  (flags :standard -warn-error -3-5 -w -60 -safe-string -color always)
  (optional)

benchs/run_benchs.ml

@@ -1824,4 +1824,32 @@ module Str = struct
            ])
 end
 
+module Hash = struct
+  let hash_ocaml (n : int) : int =
+    let offset_basis = 0xcbf29ce484222325L in
+    let prime = 0x100000001b3L in
+    let h = ref offset_basis in
+    for k = 0 to 7 do
+      (h := Int64.(mul !h prime));
+      h := Int64.(logxor !h (of_int ((n lsr (k * 8)) land 0xff)))
+    done;
+    Int64.to_int !h land max_int
+
+  let bench_hash n =
+    let run_ocaml () =
+      for i = 0 to n - 1 do
+        opaque_ignore (hash_ocaml i)
+      done
+    and run_c_stub () =
+      for i = 0 to n - 1 do
+        opaque_ignore (CCInt.hash i)
+      done
+    in
+    B.throughputN 3 ~repeat
+      [ "ocaml_fnv", run_ocaml, (); "c_stub", run_c_stub, () ]
+
+  let () =
+    B.Tree.register ("hash" @>>> [ "int" @>> app_ints bench_hash [ 1_000 ] ])
+end
+
 let () = try B.Tree.run_global () with Arg.Help msg -> print_endline msg

src/core/CCHash.ml

@@ -7,7 +7,7 @@ type 'a t = 'a -> hash
 type 'a iter = ('a -> unit) -> unit
 type 'a gen = unit -> 'a option
 
-(* FNV hashing
+(* FNV-1 hashing (multiply then XOR)
    https://en.wikipedia.org/wiki/Fowler%E2%80%93Noll%E2%80%93Vo_hash_function
 *)
 let fnv_offset_basis = 0xcbf29ce484222325L

src/core/CCInt.ml

@@ -4,20 +4,11 @@ include Int
 
 type 'a iter = ('a -> unit) -> unit
 
-(* use FNV:
+(* use FNV-1:
    https://en.wikipedia.org/wiki/Fowler%E2%80%93Noll%E2%80%93Vo_hash_function *)
-let hash (n : int) : int =
+external hash : (int[@untagged]) -> (int[@untagged])
-  let offset_basis = 0xcbf29ce484222325L in
+  = "caml_cc_hash_int_byte" "caml_cc_hash_int"
-  let prime = 0x100000001b3L in
+[@@noalloc]
-
-  let h = ref offset_basis in
-  for k = 0 to 7 do
-    (h := Int64.(mul !h prime));
-    (* h := h xor (k-th byte of n) *)
-    h := Int64.(logxor !h (of_int ((n lsr (k * 8)) land 0xff)))
-  done;
-  (* truncate back to int and remove sign *)
-  Int64.to_int !h land max_int
 
 let range i j yield =
   let rec up i j yield =

src/core/CCInt64.ml

@@ -11,21 +11,14 @@ let max : t -> t -> t = Stdlib.max
 
 let sign i = compare i zero
 
-(* use FNV:
+(* use FNV-1:
    https://en.wikipedia.org/wiki/Fowler%E2%80%93Noll%E2%80%93Vo_hash_function *)
-let hash_to_int64 (n : t) =
+external hash_to_int64 : (int64[@unboxed]) -> (int64[@unboxed])
-  let offset_basis = 0xcbf29ce484222325L in
+  = "caml_cc_hash_int64_to_int64_byte" "caml_cc_hash_int64_to_int64"
-  let prime = 0x100000001b3L in
 
-  let h = ref offset_basis in
+external hash : (int64[@unboxed]) -> (int[@untagged])
-  for k = 0 to 7 do
+  = "caml_cc_hash_int64_byte" "caml_cc_hash_int64"
-    h := mul !h prime;
+[@@noalloc]
-    (* h := h xor (k-th byte of n) *)
-    h := logxor !h (logand (shift_right n (k * 8)) 0xffL)
-  done;
-  logand !h max_int
-
-let[@inline] hash (n : t) : int = to_int (hash_to_int64 n) land Stdlib.max_int
 
 (* see {!CCInt.popcount} for more details *)
 let[@inline] popcount (b : t) : int =

src/core/cc_stubs.c

@@ -0,0 +1,67 @@
+#include <caml/alloc.h>
+#include <caml/mlvalues.h>
+#include <stdint.h>
+
+/* FNV-1 hash for a 64-bit integer.
+   https://en.wikipedia.org/wiki/Fowler%E2%80%93Noll%E2%80%93Vo_hash_function
+
+   FNV-1 order: multiply then XOR (as opposed to FNV-1a which XORs first).
+   Uses the standard 64-bit FNV parameters:
+     offset_basis = 0xcbf29ce484222325
+     prime        = 0x00000100000001b3
+
+   Core routine: operates on all 8 bytes of an int64_t. */
+
+static inline int64_t cc_fnv_hash_int64(int64_t n) {
+  uint64_t un = (uint64_t)n;
+  uint64_t h = UINT64_C(0xcbf29ce484222325);
+  const uint64_t prime = UINT64_C(0x100000001b3);
+  for (int k = 0; k < 8; k++) {
+    h *= prime;
+    h ^= (un >> (k * 8)) & 0xff;
+  }
+  return (int64_t)h;
+}
+
+/* --- CCInt.hash entry points (int -> int) --- */
+
+/* Mask to the OCaml int range (63 bits on 64-bit, 31 on 32-bit)
+   before hashing, so negative OCaml ints hash the same as
+   the unsigned representation seen by OCaml's [lsr]. */
+#define OCAML_INT_MASK ((UINT64_C(1) << (8 * sizeof(value) - 1)) - 1)
+
+/* native: untagged int in, untagged int out */
+CAMLprim intnat caml_cc_hash_int(intnat n) {
+  int64_t projected = (int64_t)((uint64_t)n & OCAML_INT_MASK);
+  return (intnat)((uint64_t)cc_fnv_hash_int64(projected) & Max_long);
+}
+
+/* bytecode: boxed value in, boxed value out */
+CAMLprim value caml_cc_hash_int_byte(value v_n) {
+  return Val_long(caml_cc_hash_int(Long_val(v_n)));
+}
+
+/* --- int64 hash entry points (int64 -> int) --- */
+
+/* native: unboxed int64 in, untagged int out */
+CAMLprim intnat caml_cc_hash_int64(int64_t n) {
+  return (intnat)((uint64_t)cc_fnv_hash_int64(n) & Max_long);
+}
+
+/* bytecode: boxed int64 value in, boxed value out */
+CAMLprim value caml_cc_hash_int64_byte(value v_n) {
+  return Val_long(caml_cc_hash_int64(Int64_val(v_n)));
+}
+
+/* --- int64 -> int64 entry points (for hash_to_int64) --- */
+
+/* native: unboxed int64 in, unboxed int64 out.
+   Masks to non-negative int64 (matches OCaml's Int64.max_int). */
+CAMLprim int64_t caml_cc_hash_int64_to_int64(int64_t n) {
+  return cc_fnv_hash_int64(n) & INT64_MAX;
+}
+
+/* bytecode: boxed int64 in, boxed int64 out */
+CAMLprim value caml_cc_hash_int64_to_int64_byte(value v_n) {
+  return caml_copy_int64(cc_fnv_hash_int64(Int64_val(v_n)) & INT64_MAX);
+}

src/core/dune

@@ -6,6 +6,10 @@
   (action
    (run %{project_root}/src/core/cpp/cpp.exe %{input-file})))
  (flags :standard -nolabels -open CCMonomorphic)
+ (foreign_stubs
+  (language c)
+  (flags :standard -std=c99 -O2)
+  (names cc_stubs))
  (libraries either containers.monomorphic containers.domain))
 
 (ocamllex

tests/core/t_int64.ml

@@ -101,3 +101,42 @@ eq' 63 (popcount max_int);;
 eq' 1 (popcount min_int);;
 eq' 10 (popcount 0b1110010110110001010L);;
 eq' 5 (popcount 0b1101110000000000L)
+
+(* hash tests *)
+let ( >= ) = Stdlib.( >= )
+let ( = ) = Stdlib.( = )
+let ( <> ) = Stdlib.( <> );;
+
+(* hash is non-negative *)
+t @@ fun () -> hash 0L >= 0;;
+t @@ fun () -> hash 1L >= 0;;
+t @@ fun () -> hash (-1L) >= 0;;
+t @@ fun () -> hash max_int >= 0;;
+t @@ fun () -> hash min_int >= 0;;
+
+(* hash_to_int64 is non-negative *)
+t @@ fun () -> CCInt64.compare (hash_to_int64 0L) 0L >= 0;;
+t @@ fun () -> CCInt64.compare (hash_to_int64 (-1L)) 0L >= 0;;
+t @@ fun () -> CCInt64.compare (hash_to_int64 min_int) 0L >= 0;;
+
+(* hash is consistent with hash_to_int64 *)
+t @@ fun () -> hash 42L = Stdlib.(Int64.to_int (hash_to_int64 42L) land max_int)
+;;
+
+t @@ fun () ->
+hash (-1L) = Stdlib.(Int64.to_int (hash_to_int64 (-1L)) land max_int)
+;;
+
+(* different inputs produce different hashes *)
+t @@ fun () -> hash 0L <> hash 1L;;
+t @@ fun () -> hash 1L <> hash 2L;;
+t @@ fun () -> hash 1L <> hash (-1L);;
+t @@ fun () -> hash_to_int64 0L <> hash_to_int64 1L;;
+
+(* deterministic *)
+t @@ fun () -> hash 123L = hash 123L;;
+t @@ fun () -> hash_to_int64 123L = hash_to_int64 123L;;
+
+(* quickcheck: hash is always non-negative *)
+q Q.(map Int64.of_int int) (fun n -> hash n >= 0);;
+q Q.(map Int64.of_int int) (fun n -> CCInt64.compare (hash_to_int64 n) 0L >= 0)