2015-12-31 08:37:27 +02:00
|
|
|
// Copyright (c) 2014-2016, The Monero Project
|
2014-10-21 23:33:43 +03:00
|
|
|
// All rights reserved.
|
2015-12-14 06:54:39 +02:00
|
|
|
//
|
2014-10-21 23:33:43 +03:00
|
|
|
// Redistribution and use in source and binary forms, with or without modification, are
|
|
|
|
|
// permitted provided that the following conditions are met:
|
2015-12-14 06:54:39 +02:00
|
|
|
//
|
2014-10-21 23:33:43 +03:00
|
|
|
// 1. Redistributions of source code must retain the above copyright notice, this list of
|
|
|
|
|
// conditions and the following disclaimer.
|
2015-12-14 06:54:39 +02:00
|
|
|
//
|
2014-10-21 23:33:43 +03:00
|
|
|
// 2. Redistributions in binary form must reproduce the above copyright notice, this list
|
|
|
|
|
// of conditions and the following disclaimer in the documentation and/or other
|
|
|
|
|
// materials provided with the distribution.
|
2015-12-14 06:54:39 +02:00
|
|
|
//
|
2014-10-21 23:33:43 +03:00
|
|
|
// 3. Neither the name of the copyright holder nor the names of its contributors may be
|
|
|
|
|
// used to endorse or promote products derived from this software without specific
|
|
|
|
|
// prior written permission.
|
2015-12-14 06:54:39 +02:00
|
|
|
//
|
2014-10-21 23:33:43 +03:00
|
|
|
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY
|
|
|
|
|
// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
|
|
|
|
|
// MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
|
|
|
|
|
// THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
|
|
|
|
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
|
|
|
|
|
// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
|
|
|
|
|
// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
|
|
|
|
|
// STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
|
|
|
|
|
// THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
2015-03-25 17:41:30 +02:00
|
|
|
#pragma once
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2015-05-17 05:05:54 +03:00
|
|
|
#include <atomic>
|
|
|
|
|
|
2015-03-06 22:20:45 +02:00
|
|
|
#include "blockchain_db/blockchain_db.h"
|
2014-10-28 02:45:33 +02:00
|
|
|
#include "cryptonote_protocol/blobdatatype.h" // for type blobdata
|
2014-10-21 23:33:43 +03:00
|
|
|
|
|
|
|
|
#include <lmdb.h>
|
|
|
|
|
|
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 23:09:32 +03:00
|
|
|
#define ENABLE_AUTO_RESIZE
|
|
|
|
|
|
2014-10-21 23:33:43 +03:00
|
|
|
namespace cryptonote
|
|
|
|
|
{
|
|
|
|
|
|
2015-03-14 23:24:51 +02:00
|
|
|
struct mdb_txn_safe
|
2014-10-23 22:37:10 +03:00
|
|
|
{
|
2015-05-16 03:42:47 +03:00
|
|
|
mdb_txn_safe();
|
|
|
|
|
~mdb_txn_safe();
|
2014-10-23 22:37:10 +03:00
|
|
|
|
2015-05-16 03:42:47 +03:00
|
|
|
void commit(std::string message = "");
|
2014-10-23 22:37:10 +03:00
|
|
|
|
2015-02-12 01:55:53 +02:00
|
|
|
// This should only be needed for batch transaction which must be ensured to
|
|
|
|
|
// be aborted before mdb_env_close, not after. So we can't rely on
|
2015-03-14 23:24:51 +02:00
|
|
|
// BlockchainLMDB destructor to call mdb_txn_safe destructor, as that's too late
|
2015-02-12 01:55:53 +02:00
|
|
|
// to properly abort, since mdb_env_close would have been called earlier.
|
2015-05-16 03:42:47 +03:00
|
|
|
void abort();
|
2015-02-12 01:55:53 +02:00
|
|
|
|
2014-10-23 22:37:10 +03:00
|
|
|
operator MDB_txn*()
|
|
|
|
|
{
|
|
|
|
|
return m_txn;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
operator MDB_txn**()
|
|
|
|
|
{
|
|
|
|
|
return &m_txn;
|
|
|
|
|
}
|
|
|
|
|
|
2015-05-27 21:03:46 +03:00
|
|
|
uint64_t num_active_tx() const;
|
2015-05-17 05:05:54 +03:00
|
|
|
|
|
|
|
|
static void prevent_new_txns();
|
|
|
|
|
static void wait_no_active_txns();
|
|
|
|
|
static void allow_new_txns();
|
|
|
|
|
|
2014-10-23 22:37:10 +03:00
|
|
|
MDB_txn* m_txn;
|
2015-02-12 01:55:53 +02:00
|
|
|
bool m_batch_txn = false;
|
2015-05-17 05:05:54 +03:00
|
|
|
static std::atomic<uint64_t> num_active_txns;
|
|
|
|
|
|
|
|
|
|
// could use a mutex here, but this should be sufficient.
|
|
|
|
|
static std::atomic_flag creation_gate;
|
2014-10-23 22:37:10 +03:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
2015-05-17 05:05:54 +03:00
|
|
|
// If m_batch_active is set, a batch transaction exists beyond this class, such
|
|
|
|
|
// as a batch import with verification enabled, or possibly (later) a batch
|
|
|
|
|
// network sync.
|
|
|
|
|
//
|
|
|
|
|
// For some of the lookup methods, such as get_block_timestamp(), tx_exists(),
|
|
|
|
|
// and get_tx(), when m_batch_active is set, the lookup uses the batch
|
|
|
|
|
// transaction. This isn't only because the transaction is available, but it's
|
|
|
|
|
// necessary so that lookups include the database updates only present in the
|
|
|
|
|
// current batch write.
|
|
|
|
|
//
|
|
|
|
|
// A regular network sync without batch writes is expected to open a new read
|
|
|
|
|
// transaction, as those lookups are part of the validation done prior to the
|
|
|
|
|
// write for block and tx data, so no write transaction is open at the time.
|
2014-10-21 23:33:43 +03:00
|
|
|
class BlockchainLMDB : public BlockchainDB
|
|
|
|
|
{
|
|
|
|
|
public:
|
2015-02-12 01:55:53 +02:00
|
|
|
BlockchainLMDB(bool batch_transactions=false);
|
2014-10-21 23:33:43 +03:00
|
|
|
~BlockchainLMDB();
|
|
|
|
|
|
2015-02-12 02:02:20 +02:00
|
|
|
virtual void open(const std::string& filename, const int mdb_flags=0);
|
2014-10-21 23:33:43 +03:00
|
|
|
|
|
|
|
|
virtual void close();
|
|
|
|
|
|
|
|
|
|
virtual void sync();
|
|
|
|
|
|
|
|
|
|
virtual void reset();
|
|
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual std::vector<std::string> get_filenames() const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2015-03-14 03:39:27 +02:00
|
|
|
virtual std::string get_db_name() const;
|
|
|
|
|
|
2014-10-21 23:33:43 +03:00
|
|
|
virtual bool lock();
|
|
|
|
|
|
|
|
|
|
virtual void unlock();
|
|
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual bool block_exists(const crypto::hash& h) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual block get_block(const crypto::hash& h) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual uint64_t get_block_height(const crypto::hash& h) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual block_header get_block_header(const crypto::hash& h) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual block get_block_from_height(const uint64_t& height) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual uint64_t get_block_timestamp(const uint64_t& height) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual uint64_t get_top_block_timestamp() const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual size_t get_block_size(const uint64_t& height) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual difficulty_type get_block_cumulative_difficulty(const uint64_t& height) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual difficulty_type get_block_difficulty(const uint64_t& height) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual uint64_t get_block_already_generated_coins(const uint64_t& height) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual crypto::hash get_block_hash_from_height(const uint64_t& height) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual std::vector<block> get_blocks_range(const uint64_t& h1, const uint64_t& h2) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual std::vector<crypto::hash> get_hashes_range(const uint64_t& h1, const uint64_t& h2) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual crypto::hash top_block_hash() const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual block get_top_block() const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual uint64_t height() const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual bool tx_exists(const crypto::hash& h) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual uint64_t get_tx_unlock_time(const crypto::hash& h) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual transaction get_tx(const crypto::hash& h) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual uint64_t get_tx_count() const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual std::vector<transaction> get_tx_list(const std::vector<crypto::hash>& hlist) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual uint64_t get_tx_block_height(const crypto::hash& h) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual uint64_t get_num_outputs(const uint64_t& amount) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 23:09:32 +03:00
|
|
|
virtual output_data_t get_output_key(const uint64_t& amount, const uint64_t& index);
|
|
|
|
|
virtual output_data_t get_output_key(const uint64_t& global_index) const;
|
|
|
|
|
virtual void get_output_key(const uint64_t &amount, const std::vector<uint64_t> &offsets, std::vector<output_data_t> &outputs);
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-14 22:20:41 +02:00
|
|
|
virtual tx_out_index get_output_tx_and_index_from_global(const uint64_t& index) const;
|
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 23:09:32 +03:00
|
|
|
virtual void get_output_tx_and_index_from_global(const std::vector<uint64_t> &global_indices,
|
2015-12-14 06:54:39 +02:00
|
|
|
std::vector<tx_out_index> &tx_out_indices) const;
|
2014-12-14 22:20:41 +02:00
|
|
|
|
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 23:09:32 +03:00
|
|
|
virtual tx_out_index get_output_tx_and_index(const uint64_t& amount, const uint64_t& index);
|
|
|
|
|
virtual void get_output_tx_and_index(const uint64_t& amount, const std::vector<uint64_t> &offsets, std::vector<tx_out_index> &indices);
|
|
|
|
|
virtual void get_output_global_indices(const uint64_t& amount, const std::vector<uint64_t> &offsets, std::vector<uint64_t> &indices);
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual std::vector<uint64_t> get_tx_output_indices(const crypto::hash& h) const;
|
2015-01-09 22:57:33 +02:00
|
|
|
virtual std::vector<uint64_t> get_tx_amount_output_indices(const crypto::hash& h) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
virtual bool has_key_image(const crypto::key_image& img) const;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2015-10-25 12:45:25 +02:00
|
|
|
virtual bool for_all_key_images(std::function<bool(const crypto::key_image&)>) const;
|
|
|
|
|
virtual bool for_all_blocks(std::function<bool(uint64_t, const crypto::hash&, const cryptonote::block&)>) const;
|
|
|
|
|
virtual bool for_all_transactions(std::function<bool(const crypto::hash&, const cryptonote::transaction&)>) const;
|
|
|
|
|
virtual bool for_all_outputs(std::function<bool(uint64_t amount, const crypto::hash &tx_hash, size_t tx_idx)> f) const;
|
|
|
|
|
|
2014-10-23 22:37:10 +03:00
|
|
|
virtual uint64_t add_block( const block& blk
|
|
|
|
|
, const size_t& block_size
|
|
|
|
|
, const difficulty_type& cumulative_difficulty
|
|
|
|
|
, const uint64_t& coins_generated
|
|
|
|
|
, const std::vector<transaction>& txs
|
|
|
|
|
);
|
|
|
|
|
|
2015-02-12 01:55:53 +02:00
|
|
|
virtual void set_batch_transactions(bool batch_transactions);
|
2015-07-11 22:28:20 +03:00
|
|
|
virtual void batch_start(uint64_t batch_num_blocks=0);
|
2015-02-12 01:55:53 +02:00
|
|
|
virtual void batch_commit();
|
|
|
|
|
virtual void batch_stop();
|
|
|
|
|
virtual void batch_abort();
|
|
|
|
|
|
2014-10-31 00:33:35 +02:00
|
|
|
virtual void pop_block(block& blk, std::vector<transaction>& txs);
|
|
|
|
|
|
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 23:09:32 +03:00
|
|
|
virtual bool can_thread_bulk_indices() const { return true; }
|
2014-10-21 23:33:43 +03:00
|
|
|
private:
|
2015-07-12 08:46:16 +03:00
|
|
|
void do_resize(uint64_t size_increase=0);
|
2015-05-17 05:05:54 +03:00
|
|
|
|
2015-07-12 08:46:16 +03:00
|
|
|
bool need_resize(uint64_t threshold_size=0) const;
|
|
|
|
|
void check_and_resize_for_batch(uint64_t batch_num_blocks);
|
|
|
|
|
uint64_t get_estimated_batch_size(uint64_t batch_num_blocks) const;
|
2015-05-17 05:05:54 +03:00
|
|
|
|
2014-10-21 23:33:43 +03:00
|
|
|
virtual void add_block( const block& blk
|
|
|
|
|
, const size_t& block_size
|
|
|
|
|
, const difficulty_type& cumulative_difficulty
|
|
|
|
|
, const uint64_t& coins_generated
|
2015-02-12 01:55:53 +02:00
|
|
|
, const crypto::hash& block_hash
|
2014-10-21 23:33:43 +03:00
|
|
|
);
|
|
|
|
|
|
2014-10-23 22:37:10 +03:00
|
|
|
virtual void remove_block();
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2015-02-12 01:55:53 +02:00
|
|
|
virtual void add_transaction_data(const crypto::hash& blk_hash, const transaction& tx, const crypto::hash& tx_hash);
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2015-01-12 04:04:04 +02:00
|
|
|
virtual void remove_transaction_data(const crypto::hash& tx_hash, const transaction& tx);
|
2014-10-21 23:33:43 +03:00
|
|
|
|
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 23:09:32 +03:00
|
|
|
virtual void add_output(const crypto::hash& tx_hash, const tx_out& tx_output, const uint64_t& local_index, const uint64_t unlock_time);
|
2014-10-21 23:33:43 +03:00
|
|
|
|
|
|
|
|
virtual void remove_output(const tx_out& tx_output);
|
|
|
|
|
|
2015-01-12 04:04:04 +02:00
|
|
|
void remove_tx_outputs(const crypto::hash& tx_hash, const transaction& tx);
|
2014-10-31 00:33:35 +02:00
|
|
|
|
2015-01-12 04:04:04 +02:00
|
|
|
void remove_output(const uint64_t& out_index, const uint64_t amount);
|
|
|
|
|
void remove_amount_output_index(const uint64_t amount, const uint64_t global_output_index);
|
2014-10-31 00:33:35 +02:00
|
|
|
|
2014-10-21 23:33:43 +03:00
|
|
|
virtual void add_spent_key(const crypto::key_image& k_image);
|
|
|
|
|
|
|
|
|
|
virtual void remove_spent_key(const crypto::key_image& k_image);
|
|
|
|
|
|
2015-09-20 20:41:38 +03:00
|
|
|
// Hard fork
|
|
|
|
|
virtual void set_hard_fork_starting_height(uint8_t version, uint64_t height);
|
|
|
|
|
virtual uint64_t get_hard_fork_starting_height(uint8_t version) const;
|
|
|
|
|
virtual void set_hard_fork_version(uint64_t height, uint8_t version);
|
|
|
|
|
virtual uint8_t get_hard_fork_version(uint64_t height) const;
|
2016-01-15 16:00:58 +02:00
|
|
|
virtual void check_hard_fork_info();
|
2016-02-05 03:15:37 +02:00
|
|
|
virtual void drop_hard_fork_info();
|
2015-09-20 20:41:38 +03:00
|
|
|
|
2014-10-28 02:45:33 +02:00
|
|
|
/**
|
|
|
|
|
* @brief convert a tx output to a blob for storage
|
|
|
|
|
*
|
|
|
|
|
* @param output the output to convert
|
|
|
|
|
*
|
|
|
|
|
* @return the resultant blob
|
|
|
|
|
*/
|
2015-05-27 21:03:46 +03:00
|
|
|
blobdata output_to_blob(const tx_out& output) const;
|
2014-10-28 02:45:33 +02:00
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief convert a tx output blob to a tx output
|
|
|
|
|
*
|
|
|
|
|
* @param blob the blob to convert
|
|
|
|
|
*
|
|
|
|
|
* @return the resultant tx output
|
|
|
|
|
*/
|
2014-12-06 23:37:22 +02:00
|
|
|
tx_out output_from_blob(const blobdata& blob) const;
|
2014-10-28 02:45:33 +02:00
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief get the global index of the index-th output of the given amount
|
|
|
|
|
*
|
|
|
|
|
* @param amount the output amount
|
|
|
|
|
* @param index the index into the set of outputs of that amount
|
|
|
|
|
*
|
|
|
|
|
* @return the global index of the desired output
|
|
|
|
|
*/
|
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 23:09:32 +03:00
|
|
|
uint64_t get_output_global_index(const uint64_t& amount, const uint64_t& index);
|
2014-10-28 02:45:33 +02:00
|
|
|
|
2014-12-06 23:37:22 +02:00
|
|
|
void check_open() const;
|
2014-10-23 22:37:10 +03:00
|
|
|
|
2015-12-27 00:27:35 +02:00
|
|
|
virtual bool is_read_only() const;
|
|
|
|
|
|
2015-12-06 22:48:17 +02:00
|
|
|
// fix up anything that may be wrong due to past bugs
|
|
|
|
|
virtual void fixup();
|
|
|
|
|
|
2014-10-21 23:33:43 +03:00
|
|
|
MDB_env* m_env;
|
|
|
|
|
|
|
|
|
|
MDB_dbi m_blocks;
|
2014-10-28 02:45:33 +02:00
|
|
|
MDB_dbi m_block_heights;
|
2014-10-21 23:33:43 +03:00
|
|
|
MDB_dbi m_block_hashes;
|
2014-10-28 02:45:33 +02:00
|
|
|
MDB_dbi m_block_timestamps;
|
2014-10-21 23:33:43 +03:00
|
|
|
MDB_dbi m_block_sizes;
|
|
|
|
|
MDB_dbi m_block_diffs;
|
|
|
|
|
MDB_dbi m_block_coins;
|
|
|
|
|
|
|
|
|
|
MDB_dbi m_txs;
|
2014-10-28 02:45:33 +02:00
|
|
|
MDB_dbi m_tx_unlocks;
|
2014-10-24 02:47:36 +03:00
|
|
|
MDB_dbi m_tx_heights;
|
|
|
|
|
MDB_dbi m_tx_outputs;
|
2014-10-21 23:33:43 +03:00
|
|
|
|
2014-10-24 02:47:36 +03:00
|
|
|
MDB_dbi m_output_txs;
|
|
|
|
|
MDB_dbi m_output_indices;
|
|
|
|
|
MDB_dbi m_output_amounts;
|
2014-10-31 23:34:15 +02:00
|
|
|
MDB_dbi m_output_keys;
|
2014-10-24 02:47:36 +03:00
|
|
|
|
|
|
|
|
MDB_dbi m_spent_keys;
|
2014-10-23 22:37:10 +03:00
|
|
|
|
2015-09-20 20:41:38 +03:00
|
|
|
MDB_dbi m_hf_starting_heights;
|
|
|
|
|
MDB_dbi m_hf_versions;
|
|
|
|
|
|
2015-10-26 17:52:07 +02:00
|
|
|
MDB_dbi m_properties;
|
|
|
|
|
|
2014-10-23 22:37:10 +03:00
|
|
|
uint64_t m_height;
|
2014-10-24 02:47:36 +03:00
|
|
|
uint64_t m_num_outputs;
|
2014-10-23 22:37:10 +03:00
|
|
|
std::string m_folder;
|
2015-03-14 23:24:51 +02:00
|
|
|
mdb_txn_safe* m_write_txn; // may point to either a short-lived txn or a batch txn
|
2015-05-17 05:05:54 +03:00
|
|
|
mdb_txn_safe* m_write_batch_txn; // persist batch txn outside of BlockchainLMDB
|
2015-02-12 01:55:53 +02:00
|
|
|
|
|
|
|
|
bool m_batch_transactions; // support for batch transactions
|
|
|
|
|
bool m_batch_active; // whether batch transaction is in progress
|
2015-05-17 05:05:54 +03:00
|
|
|
|
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 23:09:32 +03:00
|
|
|
#if defined(__arm__)
|
|
|
|
|
// force a value so it can compile with 32-bit ARM
|
|
|
|
|
constexpr static uint64_t DEFAULT_MAPSIZE = 1LL << 31;
|
|
|
|
|
#else
|
|
|
|
|
#if defined(ENABLE_AUTO_RESIZE)
|
|
|
|
|
constexpr static uint64_t DEFAULT_MAPSIZE = 1LL << 30;
|
|
|
|
|
#else
|
|
|
|
|
constexpr static uint64_t DEFAULT_MAPSIZE = 1LL << 33;
|
|
|
|
|
#endif
|
|
|
|
|
#endif
|
|
|
|
|
|
2015-05-17 05:05:54 +03:00
|
|
|
constexpr static float RESIZE_PERCENT = 0.8f;
|
2014-10-21 23:33:43 +03:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
} // namespace cryptonote
|