Speedup movegen with VBMI2

Passed STC
LLR: 2.96 (-2.94,2.94) <0.00,2.00>
Total: 166720 W: 43191 L: 42701 D: 80828
Ptnml(0-2): 348, 18567, 45069, 18999, 377
https://tests.stockfishchess.org/tests/view/686ae98dfe0f2fe354c0c867

Refactor movegen to emit to a vector with 16-bit elements, which enables a
speedup with AVX512-VBMI2 when writing moves to the move list.

Very crude timing measurements of perft via timing ./stockfish "go perft 7"
demonstrates approximately 17% perft speedup:

Summary
  ./Stockfish-dev/src/stockfish 'go perft 7' ran
    1.17 ± 0.04 times faster than ./Stockfish-base/src/stockfish 'go perft 7'

Estimated overall nps increase of 0.4% via speedtest: 33605229 -> 33749825
(many thanks JonathanHallstrom).

The corresponding arch is avx512icl as it is a good baseline for consumer
avx-512 feature set support; Intel Ice Lake was the first consumer AVX-512 CPU
and it is a decent subset of what AMD Zen 4 supports.

closes https://github.com/official-stockfish/Stockfish/pull/6153

No functional change

src/movepick.cpp

@@ -123,7 +123,7 @@ MovePicker::MovePicker(const Position& p, Move ttm, int th, const CapturePieceTo
 // Captures are ordered by Most Valuable Victim (MVV), preferring captures
 // with a good history. Quiets moves are ordered using the history tables.
 template<GenType Type>
-void MovePicker::score() {
+ExtMove* MovePicker::score(MoveList<Type>& ml) {
 
     static_assert(Type == CAPTURES || Type == QUIETS || Type == EVASIONS, "Wrong type");
 
@@ -138,8 +138,12 @@ void MovePicker::score() {
         threatByLesser[QUEEN] = pos.attacks_by<ROOK>(~us) | threatByLesser[ROOK];
     }
 
-    for (auto& m : *this)
+    ExtMove* it = cur;
+    for (auto move : ml)
     {
+        ExtMove& m = *it++;
+        m          = move;
+
         const Square    from          = m.from_sq();
         const Square    to            = m.to_sq();
         const Piece     pc            = pos.moved_piece(m);
@@ -189,6 +193,7 @@ void MovePicker::score() {
             }
         }
     }
+    return it;
 }
 
 // Returns the next move satisfying a predicate function.
@@ -222,14 +227,16 @@ top:
 
     case CAPTURE_INIT :
     case PROBCUT_INIT :
-    case QCAPTURE_INIT :
+    case QCAPTURE_INIT : {
+        MoveList<CAPTURES> ml(pos);
+
         cur = endBadCaptures = moves;
-        endCur = endCaptures = generate<CAPTURES>(pos, cur);
+        endCur = endCaptures = score<CAPTURES>(ml);
 
-        score<CAPTURES>();
         partial_insertion_sort(cur, endCur, std::numeric_limits<int>::min());
         ++stage;
         goto top;
+    }
 
     case GOOD_CAPTURE :
         if (select([&]() {
@@ -246,9 +253,10 @@ top:
     case QUIET_INIT :
         if (!skipQuiets)
         {
-            endCur = endGenerated = generate<QUIETS>(pos, cur);
+            MoveList<QUIETS> ml(pos);
+
+            endCur = endGenerated = score<QUIETS>(ml);
 
-            score<QUIETS>();
             partial_insertion_sort(cur, endCur, -3560 * depth);
         }
 
@@ -283,14 +291,16 @@ top:
 
         return Move::none();
 
-    case EVASION_INIT :
+    case EVASION_INIT : {
+        MoveList<EVASIONS> ml(pos);
+
         cur    = moves;
-        endCur = endGenerated = generate<EVASIONS>(pos, cur);
+        endCur = endGenerated = score<EVASIONS>(ml);
 
-        score<EVASIONS>();
         partial_insertion_sort(cur, endCur, std::numeric_limits<int>::min());
         ++stage;
         [[fallthrough]];
+    }
 
     case EVASION :
     case QCAPTURE :