bucket_counting.patch

commit c978cc44712cc3637d961981a5c9f8932c9d42a6
Author: Lőrinc <pap.lorinc@gmail.com>
Date:   Thu Apr 30 13:47:56 2026 +0200

    temp: log CCoinsMap bucket distribution
    
    Log CCoinsMap bucket counts before cache flush and sync operations.
    
    This temporary measurement commit records size, bucket count, load factor, maximum bucket size, and bucket-size histogram bins for comparing outpoint hashers during reindex-chainstate.

diff --git a/src/coins.cpp b/src/coins.cpp
index c403e006c8..717ae03127 100644
--- a/src/coins.cpp
+++ b/src/coins.cpp
@@ -10,10 +10,53 @@
 #include <util/log.h>
 #include <util/trace.h>
 
+#include <array>
+
 TRACEPOINT_SEMAPHORE(utxocache, add);
 TRACEPOINT_SEMAPHORE(utxocache, spent);
 TRACEPOINT_SEMAPHORE(utxocache, uncache);
 
+namespace {
+
+void LogCoinsMapBucketDistribution(const CCoinsMap& map, const char* context)
+{
+    if (map.empty()) return;
+
+    std::array<size_t, 8> buckets{};
+    size_t max_bucket_size{0};
+    for (size_t bucket{0}; bucket < map.bucket_count(); ++bucket) {
+        const size_t bucket_size{map.bucket_size(bucket)};
+        if (bucket_size > max_bucket_size) max_bucket_size = bucket_size;
+
+        if (bucket_size <= 4) {
+            ++buckets[bucket_size];
+        } else if (bucket_size <= 8) {
+            ++buckets[5];
+        } else if (bucket_size <= 16) {
+            ++buckets[6];
+        } else {
+            ++buckets[7];
+        }
+    }
+
+    LogInfo("CCoinsMap bucket distribution (%s): size=%zu bucket_count=%zu load_factor=%.6f max_bucket_size=%zu buckets{0=%zu,1=%zu,2=%zu,3=%zu,4=%zu,5-8=%zu,9-16=%zu,17+=%zu}",
+            context,
+            map.size(),
+            map.bucket_count(),
+            static_cast<double>(map.load_factor()),
+            max_bucket_size,
+            buckets[0],
+            buckets[1],
+            buckets[2],
+            buckets[3],
+            buckets[4],
+            buckets[5],
+            buckets[6],
+            buckets[7]);
+}
+
+} // namespace
+
 CoinsViewEmpty& CoinsViewEmpty::Get()
 {
     static CoinsViewEmpty instance;
@@ -259,6 +302,7 @@ void CCoinsViewCache::BatchWrite(CoinsViewCacheCursor& cursor, const uint256& in
 
 void CCoinsViewCache::Flush(bool reallocate_cache)
 {
+    LogCoinsMapBucketDistribution(cacheCoins, reallocate_cache ? "Flush reallocate=true" : "Flush reallocate=false");
     auto cursor{CoinsViewCacheCursor(m_dirty_count, m_sentinel, cacheCoins, /*will_erase=*/true)};
     base->BatchWrite(cursor, m_block_hash);
     Assume(m_dirty_count == 0);
@@ -271,6 +315,7 @@ void CCoinsViewCache::Flush(bool reallocate_cache)
 
 void CCoinsViewCache::Sync()
 {
+    LogCoinsMapBucketDistribution(cacheCoins, "Sync");
     auto cursor{CoinsViewCacheCursor(m_dirty_count, m_sentinel, cacheCoins, /*will_erase=*/false)};
     base->BatchWrite(cursor, m_block_hash);
     Assume(m_dirty_count == 0);