Use list::splice to move lru nodes

Summary:
By using `list::splice`, we can avoid making a call to create and delete nodes. By doing so, only internal pointers are reassigned.

`list::splice` documentation: https://www.boost.org/doc/libs/1_67_0/doc/html/boost/intrusive/list.html#idp55504432-bb

Reviewed By: yfeldblum

Differential Revision: D22369091

fbshipit-source-id: efde0f520f81506b96272b320ae9d65ce1cabcb2

folly/container/EvictingCacheMap.h

@@ -256,8 +256,7 @@ class EvictingCacheMap {
     if (it == index_.end()) {
       return end();
     }
-    lru_.erase(lru_.iterator_to(*it));
-    lru_.push_front(*it);
+    lru_.splice(lru_.begin(), lru_, lru_.iterator_to(*it));
     return iterator(lru_.iterator_to(*it));
   }
 
@@ -343,8 +342,7 @@ class EvictingCacheMap {
     if (it != index_.end()) {
       it->pr.second = std::move(value);
       if (promote) {
-        lru_.erase(lru_.iterator_to(*it));
-        lru_.push_front(*it);
+        lru_.splice(lru_.begin(), lru_, lru_.iterator_to(*it));
       }
     } else {
       auto node = new Node(key, std::move(value));

folly/container/test/EvictingCacheMapBench.cpp

+/*
+ * Copyright (c) Facebook, Inc. and its affiliates.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <folly/container/EvictingCacheMap.h>
+
+#include <folly/Benchmark.h>
+
+using namespace folly;
+
+// This function creates a cache of size `numElts` and scans through it `n`
+// times in order to cause the most churn in the LRU structure as possible.
+// This is meant to exercise the performance of the lookup path in the cache,
+// most notably promotions within the LRU.
+void scanCache(uint32_t n, size_t numElts) {
+  BenchmarkSuspender suspender;
+
+  EvictingCacheMap<size_t, size_t> m(numElts);
+  for (size_t i = 0; i < numElts; ++i) {
+    m.insert(i, i);
+  }
+
+  suspender.dismissing([&]() {
+    for (uint32_t i = 0; i < n; ++i) {
+      for (size_t j = 0; j < numElts; ++j) {
+        m.get(j);
+      }
+    }
+  });
+}
+
+BENCHMARK_PARAM(scanCache, 1000) // 1K
+BENCHMARK_PARAM(scanCache, 10000) // 10K
+BENCHMARK_PARAM(scanCache, 100000) // 100K
+BENCHMARK_PARAM(scanCache, 1000000) // 1M
+
+int main() {
+  runBenchmarks();
+}