add slightly faster RMQ

Implementations/content/data-structures/Static Range Queries (9.1)/FasterRMQ.h

@@ -0,0 +1,24 @@
+/**
+ * Description: Faster 1D range minimum query. 
+ * Source: KACTL
+ * Verification: 
+	* https://judge.yosupo.jp/submission/126814
+ * Memory: O(N\log N)
+ * Time: O(1)
+ */
+
+tcT, size_t SZ> struct RMQ { // floor(log_2(x))
+	static constexpr int level(int x) { return 31-__builtin_clz(x); }
+	array<array<T,SZ>, level(SZ)+1> jmp;
+	T cmb(T a, T b) { return min(a, b); }
+	void init(const V<T>& v) { assert(sz(v) <= SZ);
+		copy(all(v), begin(jmp[0]));
+		for (int j = 1; 1<<j <= sz(v); ++j) {
+			F0R(i,sz(v)-(1<<j)+1) jmp[j][i] = cmb(jmp[j-1][i],
+				jmp[j-1][i+(1<<(j-1))]);
+		}
+	}
+	T query(int l, int r) {
+		assert(l <= r); int d = level(r-l+1);
+		return cmb(jmp[d][l],jmp[d][r-(1<<d)+1]); }
+};

Implementations/content/data-structures/Static Range Queries (9.1)/RMQ (9.1).h

@@ -10,7 +10,7 @@
  * Time: O(1)
  */
 
-tcT> struct RMQ {
+tcT> struct RMQ { // floor(log_2(x))
 	int level(int x) { return 31-__builtin_clz(x); }
 	V<T> v; V<vi> jmp;
 	int cmb(int a, int b) {