oxen-core/src/cryptonote_core/blockchain.h

1212 lines
47 KiB
C
Raw Normal View History

2015-12-31 07:39:56 +01:00
// Copyright (c) 2014-2016, The Monero Project
//
// All rights reserved.
//
// Redistribution and use in source and binary forms, with or without modification, are
// permitted provided that the following conditions are met:
//
// 1. Redistributions of source code must retain the above copyright notice, this list of
// conditions and the following disclaimer.
//
// 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.
//
// 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.
//
// 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.
//
// Parts of this file are originally copyright (c) 2012-2013 The Cryptonote developers
#pragma once
#include <boost/serialization/serialization.hpp>
#include <boost/serialization/version.hpp>
#include <boost/serialization/list.hpp>
#include <boost/multi_index_container.hpp>
#include <boost/multi_index/global_fun.hpp>
#include <boost/multi_index/hashed_index.hpp>
#include <boost/multi_index/member.hpp>
#include <boost/foreach.hpp>
#include <atomic>
#include <unordered_map>
#include <unordered_set>
#include "syncobj.h"
#include "string_tools.h"
#include "cryptonote_basic.h"
#include "common/util.h"
#include "cryptonote_protocol/cryptonote_protocol_defs.h"
#include "rpc/core_rpc_server_commands_defs.h"
#include "difficulty.h"
#include "cryptonote_core/cryptonote_format_utils.h"
#include "verification_context.h"
#include "crypto/hash.h"
#include "checkpoints.h"
2015-09-12 12:14:00 +02:00
#include "hardfork.h"
#include "blockchain_db/blockchain_db.h"
namespace cryptonote
{
class tx_memory_pool;
struct test_options;
/** Declares ways in which the BlockchainDB backend should be told to sync
*
*/
** 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 22:09:32 +02:00
enum blockchain_db_sync_mode
{
db_sync, //!< handle syncing calls instead of the backing db, synchronously
db_async, //!< handle syncing calls instead of the backing db, asynchronously
db_nosync //!< Leave syncing up to the backing db (safest, but slowest because of disk I/O)
** 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 22:09:32 +02:00
};
/************************************************************************/
/* */
/************************************************************************/
class Blockchain
{
public:
/**
* @brief Now-defunct (TODO: remove) struct from in-memory blockchain
*/
struct transaction_chain_entry
{
transaction tx;
uint64_t m_keeper_block_height;
size_t m_blob_size;
std::vector<uint64_t> m_global_output_indexes;
};
/**
* @brief container for passing a block and metadata about it on the blockchain
*/
struct block_extended_info
{
block bl; //!< the block
uint64_t height; //!< the height of the block in the blockchain
size_t block_cumulative_size; //!< the size (in bytes) of the block
difficulty_type cumulative_difficulty; //!< the accumulated difficulty after that block
uint64_t already_generated_coins; //!< the total coins minted after that block
};
/**
* @brief Blockchain constructor
*
* @param tx_pool a reference to the transaction pool to be kept by the Blockchain
*/
Blockchain(tx_memory_pool& tx_pool);
/**
* @brief Initialize the Blockchain state
*
* @param db a pointer to the backing store to use for the blockchain
* @param testnet true if on testnet, else false
* @param test_options test parameters
*
* @return true on success, false if any initialization steps fail
*/
bool init(BlockchainDB* db, const bool testnet = false, const cryptonote::test_options *test_options = NULL);
/**
* @brief Initialize the Blockchain state
*
* @param db a pointer to the backing store to use for the blockchain
* @param hf a structure containing hardfork information
* @param testnet true if on testnet, else false
*
* @return true on success, false if any initialization steps fail
*/
bool init(BlockchainDB* db, HardFork*& hf, const bool testnet = false);
/**
* @brief Uninitializes the blockchain state
*
* Saves to disk any state that needs to be maintained
*
* @return true on success, false if any uninitialization steps fail
*/
bool deinit();
/**
* @brief assign a set of blockchain checkpoint hashes
*
* @param chk_pts the set of checkpoints to assign
*/
void set_checkpoints(checkpoints&& chk_pts) { m_checkpoints = chk_pts; }
/**
* @brief get blocks and transactions from blocks based on start height and count
*
* @param start_offset the height on the blockchain to start at
* @param count the number of blocks to get, if there are as many after start_offset
* @param blocks return-by-reference container to put result blocks in
* @param txs return-by-reference container to put result transactions in
*
* @return false if start_offset > blockchain height, else true
*/
bool get_blocks(uint64_t start_offset, size_t count, std::list<block>& blocks, std::list<transaction>& txs) const;
/**
* @brief get blocks from blocks based on start height and count
*
* @param start_offset the height on the blockchain to start at
* @param count the number of blocks to get, if there are as many after start_offset
* @param blocks return-by-reference container to put result blocks in
*
* @return false if start_offset > blockchain height, else true
*/
bool get_blocks(uint64_t start_offset, size_t count, std::list<block>& blocks) const;
/**
* @brief compiles a list of all blocks stored as alternative chains
*
* @param blocks return-by-reference container to put result blocks in
*
* @return true
*/
bool get_alternative_blocks(std::list<block>& blocks) const;
/**
* @brief returns the number of alternative blocks stored
*
* @return the number of alternative blocks stored
*/
size_t get_alternative_blocks_count() const;
/**
* @brief gets a block's hash given a height
*
* @param height the height of the block
*
* @return the hash of the block at the requested height, or a zeroed hash if there is no such block
*/
crypto::hash get_block_id_by_height(uint64_t height) const;
/**
* @brief gets the block with a given hash
*
* @param h the hash to look for
* @param blk return-by-reference variable to put result block in
*
* @return true if the block was found, else false
*/
bool get_block_by_hash(const crypto::hash &h, block &blk) const;
/**
* @brief get all block hashes (main chain, alt chains, and invalid blocks)
*
* @param main return-by-reference container to put result main chain blocks' hashes in
* @param alt return-by-reference container to put result alt chain blocks' hashes in
* @param invalid return-by-reference container to put result invalid blocks' hashes in
*/
void get_all_known_block_ids(std::list<crypto::hash> &main, std::list<crypto::hash> &alt, std::list<crypto::hash> &invalid) const;
/**
* @brief performs some preprocessing on a group of incoming blocks to speed up verification
*
* @param blocks a list of incoming blocks
*
* @return false on erroneous blocks, else true
*/
** 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 22:09:32 +02:00
bool prepare_handle_incoming_blocks(const std::list<block_complete_entry> &blocks);
/**
* @brief incoming blocks post-processing, cleanup, and disk sync
*
* @param force_sync if true, and Blockchain is handling syncing to disk, always sync
*
* @return true
*/
** 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 22:09:32 +02:00
bool cleanup_handle_incoming_blocks(bool force_sync = false);
/**
* @brief search the blockchain for a transaction by hash
*
* @param id the hash to search for
*
* @return true if the tx exists, else false
*/
bool have_tx(const crypto::hash &id) const;
/**
* @brief check if any key image in a transaction has already been spent
*
* @param tx the transaction to check
*
* @return true if any key image is already spent in the blockchain, else false
*/
bool have_tx_keyimges_as_spent(const transaction &tx) const;
/**
* @brief check if a key image is already spent on the blockchain
*
* Whenever a transaction output is used as an input for another transaction
* (a true input, not just one of a mixing set), a key image is generated
* and stored in the transaction in order to prevent double spending. If
* this key image is seen again, the transaction using it is rejected.
*
* @param key_im the key image to search for
*
* @return true if the key image is already spent in the blockchain, else false
*/
bool have_tx_keyimg_as_spent(const crypto::key_image &key_im) const;
/**
* @brief get the current height of the blockchain
*
* @return the height
*/
uint64_t get_current_blockchain_height() const;
/**
* @brief get the hash of the most recent block on the blockchain
*
* @return the hash
*/
crypto::hash get_tail_id() const;
/**
* @brief get the height and hash of the most recent block on the blockchain
*
* @param height return-by-reference variable to store the height in
*
* @return the hash
*/
crypto::hash get_tail_id(uint64_t& height) const;
/**
* @brief returns the difficulty target the next block to be added must meet
*
* @return the target
*/
** 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 22:09:32 +02:00
difficulty_type get_difficulty_for_next_block();
/**
* @brief adds a block to the blockchain
*
* Adds a new block to the blockchain. If the block's parent is not the
* current top of the blockchain, the block may be added to an alternate
* chain. If the block does not belong, is already in the blockchain
* or an alternate chain, or is invalid, return false.
*
* @param bl_ the block to be added
* @param bvc metadata about the block addition's success/failure
*
* @return true on successful addition to the blockchain, else false
*/
bool add_new_block(const block& bl_, block_verification_context& bvc);
/**
* @brief clears the blockchain and starts a new one
*
* @param b the first block in the new chain (the genesis block)
*
* @return true on success, else false
*/
bool reset_and_set_genesis_block(const block& b);
/**
* @brief creates a new block to mine against
*
* @param b return-by-reference block to be filled in
* @param miner_address address new coins for the block will go to
* @param di return-by-reference tells the miner what the difficulty target is
* @param height return-by-reference tells the miner what height it's mining against
* @param ex_nonce extra data to be added to the miner transaction's extra
*
* @return true if block template filled in successfully, else false
*/
** 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 22:09:32 +02:00
bool create_block_template(block& b, const account_public_address& miner_address, difficulty_type& di, uint64_t& height, const blobdata& ex_nonce);
/**
* @brief checks if a block is known about with a given hash
*
* This function checks the main chain, alternate chains, and invalid blocks
* for a block with the given hash
*
* @param id the hash to search for
*
* @return true if the block is known, else false
*/
bool have_block(const crypto::hash& id) const;
/**
* @brief gets the total number of transactions on the main chain
*
* @return the number of transactions on the main chain
*/
size_t get_total_transactions() const;
/**
* @brief gets the hashes for a subset of the blockchain
*
* puts into list <ids> a list of hashes representing certain blocks
* from the blockchain in reverse chronological order
*
* the blocks chosen, at the time of this writing, are:
* the most recent 11
* powers of 2 less recent from there, so 13, 17, 25, etc...
*
* @param ids return-by-reference list to put the resulting hashes in
*
* @return true
*/
bool get_short_chain_history(std::list<crypto::hash>& ids) const;
/**
* @brief get recent block hashes for a foreign chain
*
* Find the split point between us and foreign blockchain and return
* (by reference) the most recent common block hash along with up to
* BLOCKS_IDS_SYNCHRONIZING_DEFAULT_COUNT additional (more recent) hashes.
*
* @param qblock_ids the foreign chain's "short history" (see get_short_chain_history)
* @param resp return-by-reference the split height and subsequent blocks' hashes
*
* @return true if a block found in common, else false
*/
bool find_blockchain_supplement(const std::list<crypto::hash>& qblock_ids, NOTIFY_RESPONSE_CHAIN_ENTRY::request& resp) const;
/**
* @brief find the most recent common point between ours and a foreign chain
*
* This function takes a list of block hashes from another node
* on the network to find where the split point is between us and them.
* This is used to see what to send another node that needs to sync.
*
* @param qblock_ids the foreign chain's "short history" (see get_short_chain_history)
* @param starter_offset return-by-reference the most recent common block's height
*
* @return true if a block found in common, else false
*/
bool find_blockchain_supplement(const std::list<crypto::hash>& qblock_ids, uint64_t& starter_offset) const;
/**
* @brief get recent blocks for a foreign chain
*
* This function gets recent blocks relative to a foreign chain, starting either at
* a requested height or whatever height is the most recent ours and the foreign
* chain have in common.
*
* @param req_start_block if non-zero, specifies a start point (otherwise find most recent commonality)
* @param qblock_ids the foreign chain's "short history" (see get_short_chain_history)
* @param blocks return-by-reference the blocks and their transactions
* @param total_height return-by-reference our current blockchain height
* @param start_height return-by-reference the height of the first block returned
* @param max_count the max number of blocks to get
*
* @return true if a block found in common or req_start_block specified, else false
*/
bool find_blockchain_supplement(const uint64_t req_start_block, const std::list<crypto::hash>& qblock_ids, std::list<std::pair<block, std::list<transaction> > >& blocks, uint64_t& total_height, uint64_t& start_height, size_t max_count) const;
/**
* @brief retrieves a set of blocks and their transactions, and possibly other transactions
*
* the request object encapsulates a list of block hashes and a (possibly empty) list of
* transaction hashes. for each block hash, the block is fetched along with all of that
* block's transactions. Any transactions requested separately are fetched afterwards.
*
* @param arg the request
* @param rsp return-by-reference the response to fill in
*
* @return true unless any blocks or transactions are missing
*/
bool handle_get_objects(NOTIFY_REQUEST_GET_OBJECTS::request& arg, NOTIFY_RESPONSE_GET_OBJECTS::request& rsp);
/**
* @brief gets random outputs to mix with
*
* This function takes an RPC request for outputs to mix with
* and creates an RPC response with the resultant output indices.
*
* Outputs to mix with are randomly selected from the utxo set
* for each output amount in the request.
*
* @param req the output amounts and number of mixins to select
* @param res return-by-reference the resultant output indices
*
* @return true
*/
bool get_random_outs_for_amounts(const COMMAND_RPC_GET_RANDOM_OUTPUTS_FOR_AMOUNTS::request& req, COMMAND_RPC_GET_RANDOM_OUTPUTS_FOR_AMOUNTS::response& res) const;
/**
* @brief gets specific outputs to mix with
*
* This function takes an RPC request for outputs to mix with
* and creates an RPC response with the resultant output indices.
*
* Outputs to mix with are specified in the request.
*
* @param req the outputs to return
* @param res return-by-reference the resultant output indices and keys
*
* @return true
*/
bool get_outs(const COMMAND_RPC_GET_OUTPUTS::request& req, COMMAND_RPC_GET_OUTPUTS::response& res) const;
/**
* @brief gets random ringct outputs to mix with
*
* This function takes an RPC request for outputs to mix with
* and creates an RPC response with the resultant output indices
* and the matching keys.
*
* Outputs to mix with are randomly selected from the utxo set
* for each output amount in the request.
*
* @param req the output amounts and number of mixins to select
* @param res return-by-reference the resultant output indices
*
* @return true
*/
bool get_random_rct_outs(const COMMAND_RPC_GET_RANDOM_RCT_OUTPUTS::request& req, COMMAND_RPC_GET_RANDOM_RCT_OUTPUTS::response& res) const;
/**
* @brief gets the global indices for outputs from a given transaction
*
* This function gets the global indices for all outputs belonging
* to a specific transaction.
*
* @param tx_id the hash of the transaction to fetch indices for
* @param indexs return-by-reference the global indices for the transaction's outputs
*
* @return false if the transaction does not exist, or if no indices are found, otherwise true
*/
bool get_tx_outputs_gindexs(const crypto::hash& tx_id, std::vector<uint64_t>& indexs) const;
/**
* @brief stores the blockchain
*
* If Blockchain is handling storing of the blockchain (rather than BlockchainDB),
* this initiates a blockchain save.
*
* @return true unless saving the blockchain fails
*/
bool store_blockchain();
** 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 22:09:32 +02:00
/**
* @brief validates a transaction's inputs
*
* validates a transaction's inputs as correctly used and not previously
* spent. also returns the hash and height of the most recent block
* which contains an output that was used as an input to the transaction.
*
* @param tx the transaction to validate
* @param pmax_used_block_height return-by-reference block height of most recent input
* @param max_used_block_id return-by-reference block hash of most recent input
* @param tvc returned information about tx verification
* @param kept_by_block whether or not the transaction is from a previously-verified block
*
* @return false if any input is invalid, otherwise true
*/
bool check_tx_inputs(const transaction& tx, uint64_t& pmax_used_block_height, crypto::hash& max_used_block_id, tx_verification_context &tvc, bool kept_by_block = false);
/**
* @brief check that a transaction's outputs conform to current standards
*
* This function checks, for example at the time of this writing, that
* each output is of the form a * 10^b (phrased differently, that if
* written out would have only one non-zero digit in base 10).
*
* @param tx the transaction to check the outputs of
* @param tvc returned info about tx verification
*
* @return false if any outputs do not conform, otherwise true
*/
bool check_tx_outputs(const transaction& tx, tx_verification_context &tvc);
/**
* @brief gets the blocksize limit based on recent blocks
*
* @return the limit
*/
uint64_t get_current_cumulative_blocksize_limit() const;
/**
* @brief checks if the blockchain is currently being stored
*
* Note: this should be meaningless in cases where Blockchain is not
* directly managing saving the blockchain to disk.
*
* @return true if Blockchain is having the chain stored currently, else false
*/
bool is_storing_blockchain()const{return m_is_blockchain_storing;}
/**
* @brief gets the difficulty of the block with a given height
*
* @param i the height
*
* @return the difficulty
*/
uint64_t block_difficulty(uint64_t i) const;
/**
* @brief gets blocks based on a list of block hashes
*
* @tparam t_ids_container a standard-iterable container
* @tparam t_blocks_container a standard-iterable container
* @tparam t_missed_container a standard-iterable container
* @param block_ids a container of block hashes for which to get the corresponding blocks
* @param blocks return-by-reference a container to store result blocks in
* @param missed_bs return-by-reference a container to store missed blocks in
*
* @return false if an unexpected exception occurs, else true
*/
template<class t_ids_container, class t_blocks_container, class t_missed_container>
bool get_blocks(const t_ids_container& block_ids, t_blocks_container& blocks, t_missed_container& missed_bs) const;
/**
* @brief gets transactions based on a list of transaction hashes
*
* @tparam t_ids_container a standard-iterable container
* @tparam t_tx_container a standard-iterable container
* @tparam t_missed_container a standard-iterable container
* @param txs_ids a container of hashes for which to get the corresponding transactions
* @param txs return-by-reference a container to store result transactions in
* @param missed_txs return-by-reference a container to store missed transactions in
*
* @return false if an unexpected exception occurs, else true
*/
template<class t_ids_container, class t_tx_container, class t_missed_container>
bool get_transactions(const t_ids_container& txs_ids, t_tx_container& txs, t_missed_container& missed_txs) const;
//debug functions
/**
* @brief prints data about a snippet of the blockchain
*
* if start_index is greater than the blockchain height, do nothing
*
* @param start_index height on chain to start at
* @param end_index height on chain to end at
*/
2015-09-19 12:25:57 +02:00
void print_blockchain(uint64_t start_index, uint64_t end_index) const;
/**
* @brief prints every block's hash
*
* WARNING: This function will absolutely crush a terminal in prints, so
* it is recommended to redirect this output to a log file (or null sink
* if a log file is already set up, as should be the default)
*/
2015-09-19 12:25:57 +02:00
void print_blockchain_index() const;
/**
* @brief currently does nothing, candidate for removal
*
* @param file
*/
2015-09-19 12:25:57 +02:00
void print_blockchain_outs(const std::string& file) const;
/**
* @brief check the blockchain against a set of checkpoints
*
* If a block fails a checkpoint and enforce is enabled, the blockchain
* will be rolled back to two blocks prior to that block. If enforce
* is disabled, as is currently the default case with DNS-based checkpoints,
* an error will be printed to the user but no other action will be taken.
*
* @param points the checkpoints to check against
* @param enforce whether or not to take action on failure
*/
void check_against_checkpoints(const checkpoints& points, bool enforce);
/**
* @brief configure whether or not to enforce DNS-based checkpoints
*
* @param enforce the new enforcement setting
*/
void set_enforce_dns_checkpoints(bool enforce);
/**
* @brief loads new checkpoints from a file and optionally from DNS
*
* @param file_path the path of the file to look for and load checkpoints from
* @param check_dns whether or not to check for new DNS-based checkpoints
*
* @return false if any enforced checkpoint type fails to load, otherwise true
*/
bool update_checkpoints(const std::string& file_path, bool check_dns);
** 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 22:09:32 +02:00
// user options, must be called before calling init()
//FIXME: parameter names don't match function definition in .cpp file
/**
* @brief sets various performance options
*
* @param block_threads max number of threads when preparing blocks for addition
* @param blocks_per_sync number of blocks to cache before syncing to database
* @param sync_mode the ::blockchain_db_sync_mode to use
* @param fast_sync sync using built-in block hashes as trusted
*/
** 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 22:09:32 +02:00
void set_user_options(uint64_t block_threads, uint64_t blocks_per_sync,
blockchain_db_sync_mode sync_mode, bool fast_sync);
** 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 22:09:32 +02:00
/**
* @brief set whether or not to show/print time statistics
*
* @param stats the new time stats setting
*/
** 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 22:09:32 +02:00
void set_show_time_stats(bool stats) { m_show_time_stats = stats; }
/**
* @brief gets the hardfork voting state object
*
* @return the HardFork object
*/
HardFork::State get_hard_fork_state() const;
/**
* @brief gets the current hardfork version in use/voted for
*
* @return the version
*/
uint8_t get_current_hard_fork_version() const { return m_hardfork->get_current_version(); }
/**
* @brief returns the newest hardfork version known to the blockchain
*
* @return the version
*/
uint8_t get_ideal_hard_fork_version() const { return m_hardfork->get_ideal_version(); }
2016-08-12 20:19:25 +02:00
/**
* @brief returns the next hardfork version
*
* @return the version
*/
uint8_t get_next_hard_fork_version() const { return m_hardfork->get_next_version(); }
/**
* @brief returns the newest hardfork version voted to be enabled
* as of a certain height
*
* @param height the height for which to check version info
*
* @return the version
*/
uint8_t get_ideal_hard_fork_version(uint64_t height) const { return m_hardfork->get_ideal_version(height); }
/**
* @brief get information about hardfork voting for a version
*
* @param version the version in question
* @param window the size of the voting window
* @param votes the number of votes to enable <version>
* @param threshold the number of votes required to enable <version>
* @param earliest_height the earliest height at which <version> is allowed
* @param voting which version this node is voting for/using
*
* @return whether the version queried is enabled
*/
bool get_hard_fork_voting_info(uint8_t version, uint32_t &window, uint32_t &votes, uint32_t &threshold, uint64_t &earliest_height, uint8_t &voting) const;
/**
* @brief remove transactions from the transaction pool (if present)
*
* @param txids a list of hashes of transactions to be removed
*
* @return false if any removals fail, otherwise true
*/
bool flush_txes_from_pool(const std::list<crypto::hash> &txids);
/**
* @brief return a histogram of outputs on the blockchain
*
* @param amounts optional set of amounts to lookup
* @param unlocked whether to restrict instances to unlocked ones
*
* @return a set of amount/instances
*/
std::map<uint64_t, uint64_t> get_output_histogram(const std::vector<uint64_t> &amounts, bool unlocked) const;
/**
* @brief perform a check on all key images in the blockchain
*
* @param std::function the check to perform, pass/fail
*
* @return false if any key image fails the check, otherwise true
*/
bool for_all_key_images(std::function<bool(const crypto::key_image&)>) const;
/**
* @brief perform a check on all blocks in the blockchain
*
* @param std::function the check to perform, pass/fail
*
* @return false if any block fails the check, otherwise true
*/
bool for_all_blocks(std::function<bool(uint64_t, const crypto::hash&, const block&)>) const;
/**
* @brief perform a check on all transactions in the blockchain
*
* @param std::function the check to perform, pass/fail
*
* @return false if any transaction fails the check, otherwise true
*/
bool for_all_transactions(std::function<bool(const crypto::hash&, const cryptonote::transaction&)>) const;
/**
* @brief perform a check on all outputs in the blockchain
*
* @param std::function the check to perform, pass/fail
*
* @return false if any output fails the check, otherwise true
*/
bool for_all_outputs(std::function<bool(uint64_t amount, const crypto::hash &tx_hash, size_t tx_idx)>) const;
/**
* @brief get a reference to the BlockchainDB in use by Blockchain
*
* @return a reference to the BlockchainDB instance
*/
BlockchainDB& get_db()
{
return *m_db;
}
/**
* @brief get a number of outputs of a specific amount
*
* @param amount the amount
* @param offsets the indices (indexed to the amount) of the outputs
* @param outputs return-by-reference the outputs collected
* @param txs unused, candidate for removal
*/
void output_scan_worker(const uint64_t amount,const std::vector<uint64_t> &offsets,
std::vector<output_data_t> &outputs, std::unordered_map<crypto::hash,
cryptonote::transaction> &txs) 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 22:09:32 +02:00
/**
* @brief computes the "short" and "long" hashes for a set of blocks
*
* @param height the height of the first block
* @param blocks the blocks to be hashed
* @param map return-by-reference the hashes for each block
*/
void block_longhash_worker(const uint64_t height, const std::vector<block> &blocks,
std::unordered_map<crypto::hash, crypto::hash> &map) const;
private:
// TODO: evaluate whether or not each of these typedefs are left over from blockchain_storage
typedef std::unordered_map<crypto::hash, size_t> blocks_by_id_index;
typedef std::unordered_map<crypto::hash, transaction_chain_entry> transactions_container;
typedef std::unordered_set<crypto::key_image> key_images_container;
typedef std::vector<block_extended_info> blocks_container;
typedef std::unordered_map<crypto::hash, block_extended_info> blocks_ext_by_hash;
typedef std::unordered_map<crypto::hash, block> blocks_by_hash;
typedef std::map<uint64_t, std::vector<std::pair<crypto::hash, size_t>>> outputs_container; //crypto::hash - tx hash, size_t - index of out in transaction
BlockchainDB* m_db;
tx_memory_pool& m_tx_pool;
mutable epee::critical_section m_blockchain_lock; // TODO: add here reader/writer lock
// main chain
transactions_container m_transactions;
size_t m_current_block_cumul_sz_limit;
// metadata containers
** 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 22:09:32 +02:00
std::unordered_map<crypto::hash, std::unordered_map<crypto::key_image, std::vector<output_data_t>>> m_scan_table;
std::unordered_map<crypto::hash, crypto::hash> m_blocks_longhash_table;
std::unordered_map<crypto::hash, std::unordered_map<crypto::key_image, bool>> m_check_txin_table;
** 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 22:09:32 +02:00
// SHA-3 hashes for each block and for fast pow checking
std::vector<crypto::hash> m_blocks_hash_check;
std::vector<crypto::hash> m_blocks_txs_check;
blockchain_db_sync_mode m_db_sync_mode;
bool m_fast_sync;
bool m_show_time_stats;
uint64_t m_db_blocks_per_sync;
uint64_t m_max_prepare_blocks_threads;
** 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 22:09:32 +02:00
uint64_t m_fake_pow_calc_time;
uint64_t m_fake_scan_time;
uint64_t m_sync_counter;
std::vector<uint64_t> m_timestamps;
std::vector<difficulty_type> m_difficulties;
uint64_t m_timestamps_and_difficulties_height;
** 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 22:09:32 +02:00
boost::asio::io_service m_async_service;
boost::thread_group m_async_pool;
std::unique_ptr<boost::asio::io_service::work> m_async_work_idle;
// all alternative chains
blocks_ext_by_hash m_alternative_chains; // crypto::hash -> block_extended_info
// some invalid blocks
blocks_ext_by_hash m_invalid_blocks; // crypto::hash -> block_extended_info
checkpoints m_checkpoints;
std::atomic<bool> m_is_in_checkpoint_zone;
std::atomic<bool> m_is_blockchain_storing;
bool m_enforce_dns_checkpoints;
HardFork *m_hardfork;
2015-09-12 12:14:00 +02:00
bool m_testnet;
/**
* @brief collects the keys for all outputs being "spent" as an input
*
* This function makes sure that each "input" in an input (mixins) exists
* and collects the public key for each from the transaction it was included in
* via the visitor passed to it.
*
* If pmax_related_block_height is not NULL, its value is set to the height
* of the most recent block which contains an output used in the input set
*
* @tparam visitor_t a class encapsulating tx is unlocked and collect tx key
* @param tx_in_to_key a transaction input instance
* @param vis an instance of the visitor to use
* @param tx_prefix_hash the hash of the associated transaction_prefix
* @param pmax_related_block_height return-by-pointer the height of the most recent block in the input set
*
* @return false if any keys are not found or any inputs are not unlocked, otherwise true
*/
** 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 22:09:32 +02:00
template<class visitor_t>
inline bool scan_outputkeys_for_indexes(const txin_to_key& tx_in_to_key, visitor_t &vis, const crypto::hash &tx_prefix_hash, uint64_t* pmax_related_block_height = NULL) const;
/**
* @brief collect output public keys of a transaction input set
*
* This function locates all outputs associated with a given input set (mixins)
* and validates that they exist and are usable
* (unlocked, unspent is checked elsewhere).
*
* If pmax_related_block_height is not NULL, its value is set to the height
* of the most recent block which contains an output used in the input set
*
* @param txin the transaction input
* @param tx_prefix_hash the transaction prefix hash, for caching organization
* @param sig the input signature
* @param output_keys return-by-reference the public keys of the outputs in the input set
* @param pmax_related_block_height return-by-pointer the height of the most recent block in the input set
*
* @return false if any output is not yet unlocked, or is missing, otherwise true
*/
** 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 22:09:32 +02:00
bool check_tx_input(const txin_to_key& txin, const crypto::hash& tx_prefix_hash, const std::vector<crypto::signature>& sig, std::vector<crypto::public_key> &output_keys, uint64_t* pmax_related_block_height);
/**
* @brief validate a transaction's inputs and their keys
*
* This function validates transaction inputs and their keys. Previously
* it also performed double spend checking, but that has been moved to its
* own function.
*
* If pmax_related_block_height is not NULL, its value is set to the height
* of the most recent block which contains an output used in any input set
*
* Currently this function calls ring signature validation for each
* transaction.
*
* @param tx the transaction to validate
* @param tvc returned information about tx verification
* @param pmax_related_block_height return-by-pointer the height of the most recent block in the input set
*
* @return false if any validation step fails, otherwise true
*/
bool check_tx_inputs(const transaction& tx, tx_verification_context &tvc, uint64_t* pmax_used_block_height = NULL);
** 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 22:09:32 +02:00
/**
* @brief performs a blockchain reorganization according to the longest chain rule
*
* This function aggregates all the actions necessary to switch to a
* newly-longer chain. If any step in the reorganization process fails,
* the blockchain is reverted to its previous state.
*
* @param alt_chain the chain to switch to
* @param discard_disconnected_chain whether or not to keep the old chain as an alternate
*
* @return false if the reorganization fails, otherwise true
*/
bool switch_to_alternative_blockchain(std::list<blocks_ext_by_hash::iterator>& alt_chain, bool discard_disconnected_chain);
/**
* @brief removes the most recent block from the blockchain
*
* @return the block removed
*/
block pop_block_from_blockchain();
/**
* @brief validate and add a new block to the end of the blockchain
*
* This function is merely a convenience wrapper around the other
* of the same name. This one passes the block's hash to the other
* as well as the block and verification context.
*
* @param bl the block to be added
* @param bvc metadata concerning the block's validity
*
* @return true if the block was added successfully, otherwise false
*/
bool handle_block_to_main_chain(const block& bl, block_verification_context& bvc);
/**
* @brief validate and add a new block to the end of the blockchain
*
* When a block is given to Blockchain to be added to the blockchain, it
* is passed here if it is determined to belong at the end of the current
* chain.
*
* @param bl the block to be added
* @param id the hash of the block
* @param bvc metadata concerning the block's validity
*
* @return true if the block was added successfully, otherwise false
*/
bool handle_block_to_main_chain(const block& bl, const crypto::hash& id, block_verification_context& bvc);
/**
* @brief validate and add a new block to an alternate blockchain
*
* If a block to be added does not belong to the main chain, but there
* is an alternate chain to which it should be added, that is handled
* here.
*
* @param b the block to be added
* @param id the hash of the block
* @param bvc metadata concerning the block's validity
*
* @return true if the block was added successfully, otherwise false
*/
bool handle_alternative_block(const block& b, const crypto::hash& id, block_verification_context& bvc);
/**
* @brief gets the difficulty requirement for a new block on an alternate chain
*
* @param alt_chain the chain to be added to
* @param bei the block being added (and metadata, see ::block_extended_info)
*
* @return the difficulty requirement
*/
difficulty_type get_next_difficulty_for_alternative_chain(const std::list<blocks_ext_by_hash::iterator>& alt_chain, block_extended_info& bei) const;
/**
* @brief sanity checks a miner transaction before validating an entire block
*
* This function merely checks basic things like the structure of the miner
* transaction, the unlock time, and that the amount doesn't overflow.
*
* @param b the block containing the miner transaction
* @param height the height at which the block will be added
*
* @return false if anything is found wrong with the miner transaction, otherwise true
*/
bool prevalidate_miner_transaction(const block& b, uint64_t height);
/**
* @brief validates a miner (coinbase) transaction
*
* This function makes sure that the miner calculated his reward correctly
* and that his miner transaction totals reward + fee.
*
* @param b the block containing the miner transaction to be validated
* @param cumulative_block_size the block's size
* @param fee the total fees collected in the block
* @param base_reward return-by-reference the new block's generated coins
* @param already_generated_coins the amount of currency generated prior to this block
* @param partial_block_reward return-by-reference true if miner accepted only partial reward
* @param version hard fork version for that transaction
*
* @return false if anything is found wrong with the miner transaction, otherwise true
*/
bool validate_miner_transaction(const block& b, size_t cumulative_block_size, uint64_t fee, uint64_t& base_reward, uint64_t already_generated_coins, bool &partial_block_reward, uint8_t version);
/**
* @brief reverts the blockchain to its previous state following a failed switch
*
* If Blockchain fails to switch to an alternate chain when it means
* to do so, this function reverts the blockchain to how it was before
* the attempted switch.
*
* @param original_chain the chain to switch back to
* @param rollback_height the height to revert to before appending the original chain
*
* @return false if something goes wrong with reverting (very bad), otherwise true
*/
bool rollback_blockchain_switching(std::list<block>& original_chain, uint64_t rollback_height);
/**
* @brief gets recent block sizes for median calculation
*
* get the block sizes of the last <count> blocks, and return by reference <sz>.
*
* @param sz return-by-reference the list of sizes
* @param count the number of blocks to get sizes for
*/
void get_last_n_blocks_sizes(std::vector<size_t>& sz, size_t count) const;
/**
* @brief adds the given output to the requested set of random outputs
*
* @param result_outs return-by-reference the set the output is to be added to
* @param amount the output amount
* @param i the output index (indexed to amount)
*/
void add_out_to_get_random_outs(COMMAND_RPC_GET_RANDOM_OUTPUTS_FOR_AMOUNTS::outs_for_amount& result_outs, uint64_t amount, size_t i) const;
/**
* @brief adds the given output to the requested set of random ringct outputs
*
* @param outs return-by-reference the set the output is to be added to
* @param i the rct output index
*/
void add_out_to_get_rct_random_outs(std::list<COMMAND_RPC_GET_RANDOM_RCT_OUTPUTS::out_entry>& outs, size_t i) const;
/**
* @brief checks if a transaction is unlocked (its outputs spendable)
*
* This function checks to see if a transaction is unlocked.
* unlock_time is either a block index or a unix time.
*
* @param unlock_time the unlock parameter (height or time)
*
* @return true if spendable, otherwise false
*/
bool is_tx_spendtime_unlocked(uint64_t unlock_time) const;
/**
* @brief stores an invalid block in a separate container
*
* Storing invalid blocks allows quick dismissal of the same block
* if it is seen again.
*
* @param bl the invalid block
* @param h the block's hash
*
* @return false if the block cannot be stored for some reason, otherwise true
*/
bool add_block_as_invalid(const block& bl, const crypto::hash& h);
/**
* @brief stores an invalid block in a separate container
*
* Storing invalid blocks allows quick dismissal of the same block
* if it is seen again.
*
* @param bei the invalid block (see ::block_extended_info)
* @param h the block's hash
*
* @return false if the block cannot be stored for some reason, otherwise true
*/
bool add_block_as_invalid(const block_extended_info& bei, const crypto::hash& h);
/**
* @brief checks a block's timestamp
*
* This function grabs the timestamps from the most recent <n> blocks,
* where n = BLOCKCHAIN_TIMESTAMP_CHECK_WINDOW. If there are not those many
* blocks in the blockchain, the timestap is assumed to be valid. If there
* are, this function returns:
* true if the block's timestamp is not less than the timestamp of the
* median of the selected blocks
* false otherwise
*
* @param b the block to be checked
*
* @return true if the block's timestamp is valid, otherwise false
*/
bool check_block_timestamp(const block& b) const;
/**
* @brief checks a block's timestamp
*
* If the block is not more recent than the median of the recent
* timestamps passed here, it is considered invalid.
*
* @param timestamps a list of the most recent timestamps to check against
* @param b the block to be checked
*
* @return true if the block's timestamp is valid, otherwise false
*/
bool check_block_timestamp(std::vector<uint64_t>& timestamps, const block& b) const;
/**
* @brief get the "adjusted time"
*
* Currently this simply returns the current time according to the
* user's machine.
*
* @return the current time
*/
uint64_t get_adjusted_time() const;
/**
* @brief finish an alternate chain's timestamp window from the main chain
*
* for an alternate chain, get the timestamps from the main chain to complete
* the needed number of timestamps for the BLOCKCHAIN_TIMESTAMP_CHECK_WINDOW.
*
* @param start_height the alternate chain's attachment height to the main chain
* @param timestamps return-by-value the timestamps set to be populated
*
* @return true unless start_height is greater than the current blockchain height
*/
bool complete_timestamps_vector(uint64_t start_height, std::vector<uint64_t>& timestamps);
/**
* @brief calculate the block size limit for the next block to be added
*
* @return true
*/
bool update_next_cumulative_size_limit();
void return_tx_to_pool(const std::vector<transaction> &txs);
/**
* @brief make sure a transaction isn't attempting a double-spend
*
* @param tx the transaction to check
* @param keys_this_block a cumulative list of spent keys for the current block
*
* @return false if a double spend was detected, otherwise true
*/
bool check_for_double_spend(const transaction& tx, key_images_container& keys_this_block) const;
/**
* @brief validates a transaction input's ring signature
*
* @param tx_prefix_hash the transaction prefix' hash
* @param key_image the key image generated from the true input
* @param pubkeys the public keys for each input in the ring signature
* @param sig the signature generated for each input in the ring signature
* @param result false if the ring signature is invalid, otherwise true
*/
** 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 22:09:32 +02:00
void check_ring_signature(const crypto::hash &tx_prefix_hash, const crypto::key_image &key_image,
const std::vector<crypto::public_key> &pubkeys, const std::vector<crypto::signature> &sig, uint64_t &result);
/**
* @brief loads block hashes from compiled-in data set
*
* A (possibly empty) set of block hashes can be compiled into the
* monero daemon binary. This function loads those hashes into
* a useful state.
*/
void load_compiled_in_block_hashes();
};
} // namespace cryptonote