LBRYCommunity/lbrycrd
1
0
Fork 0
Code Issues 29 Pull requests 1 Projects Releases 32 Packages Wiki Activity

Add a score index to the mempool.

Browse source 
The score index is meant to represent the order of priority for being included in a block for miners.  Initially this is set to the transactions modified (by any feeDelta) fee rate.  Index improvements and unit tests by sdaftuar.
This commit is contained in:
Alex Morcos 2015-10-26 14:06:06 -04:00
parent c49d5bc9e6
commit f3fe83673e
4 changed files with 106 additions and 12 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
src/rpcblockchain.cpp
View file

@ -190,6 +190,7 @@ UniValue mempoolToJSON(bool fVerbose = false)
UniValue info(UniValue::VOBJ);
info.push_back(Pair("size", (int)e.GetTxSize()));
info.push_back(Pair("fee", ValueFromAmount(e.GetFee())));
info.push_back(Pair("modifiedfee", ValueFromAmount(e.GetModifiedFee())));
info.push_back(Pair("time", e.GetTime()));
info.push_back(Pair("height", (int)e.GetHeight()));
info.push_back(Pair("startingpriority", e.GetPriority(e.GetHeight())));
@ -247,6 +248,7 @@ UniValue getrawmempool(const UniValue& params, bool fHelp)
" \"transactionid\" : { (json object)\n"
" \"size\" : n, (numeric) transaction size in bytes\n"
" \"fee\" : n, (numeric) transaction fee in " + CURRENCY_UNIT + "\n"
" \"modifiedfee\" : n, (numeric) transaction fee with fee deltas used for mining priority\n"
" \"time\" : n, (numeric) local time transaction entered pool in seconds since 1 Jan 1970 GMT\n"
" \"height\" : n, (numeric) block height when transaction entered pool\n"
" \"startingpriority\" : n, (numeric) priority when transaction entered pool\n"

52
src/test/mempool_tests.cpp
View file

@ -102,12 +102,13 @@ BOOST_AUTO_TEST_CASE(MempoolRemoveTest)
removed.clear();
}
template<int index>
void CheckSort(CTxMemPool &pool, std::vector<std::string> &sortedOrder)
{
BOOST_CHECK_EQUAL(pool.size(), sortedOrder.size());
CTxMemPool::indexed_transaction_set::nth_index<1>::type::iterator it = pool.mapTx.get<1>().begin();
typename CTxMemPool::indexed_transaction_set::nth_index<index>::type::iterator it = pool.mapTx.get<index>().begin();
int count=0;
for (; it != pool.mapTx.get<1>().end(); ++it, ++count) {
for (; it != pool.mapTx.get<index>().end(); ++it, ++count) {
BOOST_CHECK_EQUAL(it->GetTx().GetHash().ToString(), sortedOrder[count]);
}
}
@ -163,7 +164,7 @@ BOOST_AUTO_TEST_CASE(MempoolIndexingTest)
sortedOrder[2] = tx1.GetHash().ToString(); // 10000
sortedOrder[3] = tx4.GetHash().ToString(); // 15000
sortedOrder[4] = tx2.GetHash().ToString(); // 20000
CheckSort(pool, sortedOrder);
CheckSort<1>(pool, sortedOrder);
/* low fee but with high fee child */
/* tx6 -> tx7 -> tx8, tx9 -> tx10 */
@ -175,7 +176,7 @@ BOOST_AUTO_TEST_CASE(MempoolIndexingTest)
BOOST_CHECK_EQUAL(pool.size(), 6);
// Check that at this point, tx6 is sorted low
sortedOrder.insert(sortedOrder.begin(), tx6.GetHash().ToString());
CheckSort(pool, sortedOrder);
CheckSort<1>(pool, sortedOrder);
CTxMemPool::setEntries setAncestors;
setAncestors.insert(pool.mapTx.find(tx6.GetHash()));
@ -201,7 +202,7 @@ BOOST_AUTO_TEST_CASE(MempoolIndexingTest)
sortedOrder.erase(sortedOrder.begin());
sortedOrder.push_back(tx6.GetHash().ToString());
sortedOrder.push_back(tx7.GetHash().ToString());
CheckSort(pool, sortedOrder);
CheckSort<1>(pool, sortedOrder);
/* low fee child of tx7 */
CMutableTransaction tx8 = CMutableTransaction();
@ -216,7 +217,7 @@ BOOST_AUTO_TEST_CASE(MempoolIndexingTest)
// Now tx8 should be sorted low, but tx6/tx both high
sortedOrder.insert(sortedOrder.begin(), tx8.GetHash().ToString());
CheckSort(pool, sortedOrder);
CheckSort<1>(pool, sortedOrder);
/* low fee child of tx7 */
CMutableTransaction tx9 = CMutableTransaction();
@ -231,7 +232,7 @@ BOOST_AUTO_TEST_CASE(MempoolIndexingTest)
// tx9 should be sorted low
BOOST_CHECK_EQUAL(pool.size(), 9);
sortedOrder.insert(sortedOrder.begin(), tx9.GetHash().ToString());
CheckSort(pool, sortedOrder);
CheckSort<1>(pool, sortedOrder);
std::vector<std::string> snapshotOrder = sortedOrder;
@ -273,7 +274,7 @@ BOOST_AUTO_TEST_CASE(MempoolIndexingTest)
sortedOrder.insert(sortedOrder.begin()+5, tx9.GetHash().ToString());
sortedOrder.insert(sortedOrder.begin()+6, tx8.GetHash().ToString());
sortedOrder.insert(sortedOrder.begin()+7, tx10.GetHash().ToString()); // tx10 is just before tx6
CheckSort(pool, sortedOrder);
CheckSort<1>(pool, sortedOrder);
// there should be 10 transactions in the mempool
BOOST_CHECK_EQUAL(pool.size(), 10);
@ -281,9 +282,42 @@ BOOST_AUTO_TEST_CASE(MempoolIndexingTest)
// Now try removing tx10 and verify the sort order returns to normal
std::list<CTransaction> removed;
pool.remove(pool.mapTx.find(tx10.GetHash())->GetTx(), removed, true);
CheckSort(pool, snapshotOrder);
CheckSort<1>(pool, snapshotOrder);
pool.remove(pool.mapTx.find(tx9.GetHash())->GetTx(), removed, true);
pool.remove(pool.mapTx.find(tx8.GetHash())->GetTx(), removed, true);
/* Now check the sort on the mining score index.
* Final order should be:
*
* tx7 (2M)
* tx2 (20k)
* tx4 (15000)
* tx1/tx5 (10000)
* tx3/6 (0)
* (Ties resolved by hash)
*/
sortedOrder.clear();
sortedOrder.push_back(tx7.GetHash().ToString());
sortedOrder.push_back(tx2.GetHash().ToString());
sortedOrder.push_back(tx4.GetHash().ToString());
if (tx1.GetHash() < tx5.GetHash()) {
sortedOrder.push_back(tx5.GetHash().ToString());
sortedOrder.push_back(tx1.GetHash().ToString());
} else {
sortedOrder.push_back(tx1.GetHash().ToString());
sortedOrder.push_back(tx5.GetHash().ToString());
}
if (tx3.GetHash() < tx6.GetHash()) {
sortedOrder.push_back(tx6.GetHash().ToString());
sortedOrder.push_back(tx3.GetHash().ToString());
} else {
sortedOrder.push_back(tx3.GetHash().ToString());
sortedOrder.push_back(tx6.GetHash().ToString());
}
CheckSort<3>(pool, sortedOrder);
}
BOOST_AUTO_TEST_CASE(MempoolSizeLimitTest)
{
CTxMemPool pool(CFeeRate(1000));

24
src/txmempool.cpp
View file

@ -36,6 +36,8 @@ CTxMemPoolEntry::CTxMemPoolEntry(const CTransaction& _tx, const CAmount& _nFee,
nFeesWithDescendants = nFee;
CAmount nValueIn = tx.GetValueOut()+nFee;
assert(inChainInputValue <= nValueIn);
feeDelta = 0;
}
CTxMemPoolEntry::CTxMemPoolEntry(const CTxMemPoolEntry& other)
@ -53,6 +55,11 @@ CTxMemPoolEntry::GetPriority(unsigned int currentHeight) const
return dResult;
}
void CTxMemPoolEntry::UpdateFeeDelta(int64_t newFeeDelta)
{
feeDelta = newFeeDelta;
}
// Update the given tx for any in-mempool descendants.
// Assumes that setMemPoolChildren is correct for the given tx and all
// descendants.
@ -392,6 +399,15 @@ bool CTxMemPool::addUnchecked(const uint256& hash, const CTxMemPoolEntry &entry,
}
UpdateAncestorsOf(true, newit, setAncestors);
// Update transaction's score for any feeDelta created by PrioritiseTransaction
std::map<uint256, std::pair<double, CAmount> >::const_iterator pos = mapDeltas.find(hash);
if (pos != mapDeltas.end()) {
const std::pair<double, CAmount> &deltas = pos->second;
if (deltas.second) {
mapTx.modify(newit, update_fee_delta(deltas.second));
}
}
nTransactionsUpdated++;
totalTxSize += entry.GetTxSize();
minerPolicyEstimator->processTransaction(entry, fCurrentEstimate);
@ -769,6 +785,10 @@ void CTxMemPool::PrioritiseTransaction(const uint256 hash, const string strHash,
std::pair<double, CAmount> &deltas = mapDeltas[hash];
deltas.first += dPriorityDelta;
deltas.second += nFeeDelta;
txiter it = mapTx.find(hash);
if (it != mapTx.end()) {
mapTx.modify(it, update_fee_delta(deltas.second));
}
}
LogPrintf("PrioritiseTransaction: %s priority += %f, fee += %d\n", strHash, dPriorityDelta, FormatMoney(nFeeDelta));
}
@ -818,8 +838,8 @@ bool CCoinsViewMemPool::HaveCoins(const uint256 &txid) const {
size_t CTxMemPool::DynamicMemoryUsage() const {
LOCK(cs);
// Estimate the overhead of mapTx to be 9 pointers + an allocation, as no exact formula for boost::multi_index_contained is implemented.
return memusage::MallocUsage(sizeof(CTxMemPoolEntry) + 9 * sizeof(void*)) * mapTx.size() + memusage::DynamicUsage(mapNextTx) + memusage::DynamicUsage(mapDeltas) + memusage::DynamicUsage(mapLinks) + cachedInnerUsage;
// Estimate the overhead of mapTx to be 12 pointers + an allocation, as no exact formula for boost::multi_index_contained is implemented.
return memusage::MallocUsage(sizeof(CTxMemPoolEntry) + 12 * sizeof(void*)) * mapTx.size() + memusage::DynamicUsage(mapNextTx) + memusage::DynamicUsage(mapDeltas) + memusage::DynamicUsage(mapLinks) + cachedInnerUsage;
}
void CTxMemPool::RemoveStaged(setEntries &stage) {

40
src/txmempool.h
View file

@ -69,6 +69,7 @@ private:
CAmount inChainInputValue; //! Sum of all txin values that are already in blockchain
bool spendsCoinbase; //! keep track of transactions that spend a coinbase
unsigned int sigOpCount; //! Legacy sig ops plus P2SH sig op count
int64_t feeDelta; //! Used for determining the priority of the transaction for mining in a block
// Information about descendants of this transaction that are in the
// mempool; if we remove this transaction we must remove all of these
@ -98,10 +99,13 @@ public:
unsigned int GetHeight() const { return entryHeight; }
bool WasClearAtEntry() const { return hadNoDependencies; }
unsigned int GetSigOpCount() const { return sigOpCount; }
int64_t GetModifiedFee() const { return nFee + feeDelta; }
size_t DynamicMemoryUsage() const { return nUsageSize; }
// Adjusts the descendant state, if this entry is not dirty.
void UpdateState(int64_t modifySize, CAmount modifyFee, int64_t modifyCount);
// Updates the fee delta used for mining priority score
void UpdateFeeDelta(int64_t feeDelta);
/** We can set the entry to be dirty if doing the full calculation of in-
* mempool descendants will be too expensive, which can potentially happen
@ -139,6 +143,16 @@ struct set_dirty
{ e.SetDirty(); }
};
struct update_fee_delta
{
update_fee_delta(int64_t _feeDelta) : feeDelta(_feeDelta) { }
void operator() (CTxMemPoolEntry &e) { e.UpdateFeeDelta(feeDelta); }
private:
int64_t feeDelta;
};
// extracts a TxMemPoolEntry's transaction hash
struct mempoolentry_txid
{
@ -186,6 +200,24 @@ public:
}
};
/** \class CompareTxMemPoolEntryByScore
*
* Sort by score of entry ((fee+delta)/size) in descending order
*/
class CompareTxMemPoolEntryByScore
{
public:
bool operator()(const CTxMemPoolEntry& a, const CTxMemPoolEntry& b)
{
double f1 = (double)a.GetModifiedFee() * b.GetTxSize();
double f2 = (double)b.GetModifiedFee() * a.GetTxSize();
if (f1 == f2) {
return b.GetTx().GetHash() < a.GetTx().GetHash();
}
return f1 > f2;
}
};
class CompareTxMemPoolEntryByEntryTime
{
public:
@ -223,10 +255,11 @@ public:
*
* CTxMemPool::mapTx, and CTxMemPoolEntry bookkeeping:
*
* mapTx is a boost::multi_index that sorts the mempool on 3 criteria:
* mapTx is a boost::multi_index that sorts the mempool on 4 criteria:
* - transaction hash
* - feerate [we use max(feerate of tx, feerate of tx with all descendants)]
* - time in mempool
* - mining score (feerate modified by any fee deltas from PrioritiseTransaction)
*
* Note: the term "descendant" refers to in-mempool transactions that depend on
* this one, while "ancestor" refers to in-mempool transactions that a given
@ -323,6 +356,11 @@ public:
boost::multi_index::ordered_non_unique<
boost::multi_index::identity<CTxMemPoolEntry>,
CompareTxMemPoolEntryByEntryTime
>,
// sorted by score (for mining prioritization)
boost::multi_index::ordered_unique<
boost::multi_index::identity<CTxMemPoolEntry>,
CompareTxMemPoolEntryByScore
>
>
> indexed_transaction_set;