05300c1439 Use SelectionResult in SelectCoins (Andrew Chow)
9d9b101d20 Use SelectionResult in AttemptSelection (Andrew Chow)
bb50850a44 Use SelectionResult for waste calculation (Andrew Chow)
e8f7ae5eb3 Make an OutputGroup for preset inputs (Andrew Chow)
51a9c00b4d Return SelectionResult from SelectCoinsSRD (Andrew Chow)
0ef6184575 Return SelectionResult from KnapsackSolver (Andrew Chow)
60d2ca72e3 Return SelectionResult from SelectCoinsBnB (Andrew Chow)
a339add471 Make member variables of SelectionResult private (Andrew Chow)
cbf0b9f4ff scripted-diff: Use SelectionResult in coin selector tests (Andrew Chow)
9d1d86da04 Introduce SelectionResult struct (Andrew Chow)
94d851d28c Fix bnb_search_test to use set equivalence for (Andrew Chow)
Pull request description:
Instead of returning a set of selected coins and their total value as separate items, encapsulate both of these, and other variables, into a new `SelectionResult` struct. This allows us to have all of the things relevant to a coin selection solution be in a single object. `SelectionResult` enables us to implement the waste calculation in a cleaner way.
All of the coin selection functions (`SelectCoinsBnB`, `KnapsackSolver`, `AttemptSelection`, and `SelectCoins`) are changed to use a `SelectionResult` as the output parameter.
Based on #22009
ACKs for top commit:
laanwj:
Code review ACK 05300c1439
Tree-SHA512: e4dbb4d78a6cda9c237d230b19e7265591efac5a101a64e6970f0654e2c4f93d13bb5d07b98e8c7b8d37321753dbfc94c28c3a7810cb1c59b5bc29b08a8493ef
Current CWalletTx state representation makes it possible to set
inconsistent states that won't be handled correctly by wallet sync code
or serialized & deserialized back into the same form.
For example, it is possible to call setConflicted without setting a
conflicting block hash, or setConfirmed with no transaction index. And
it's possible update individual m_confirm and fInMempool data fields
without setting an overall consistent state that can be serialized and
handled correctly.
Fix this without changing behavior by using std::variant, instead of an
enum and collection of fields, to represent sync state, so state
tracking code is safer and more legible.
This is a first step to fixing state tracking bugs
https://github.com/bitcoin-core/bitcoin-devwiki/wiki/Wallet-Transaction-Conflict-Tracking,
by adding an extra margin of safety that can prevent new bugs from being
introduced as existing bugs are fixed.
Followup to commit "MOVEONLY: CWallet transaction code out of
wallet.cpp/.h" that detaches and renames some CWalletTx methods, making
into them into standalone functions or CWallet methods instead.
There are no changes in behavior and no code changes that aren't purely
mechanical. It just gives spend and receive functions more consistent
names and removes the circular dependencies added by the earlier
MOVEONLY commit.
There are also no comment or documentation changes. Removed comments
from transaction.h are just migrated to spend.h, receive.h, and
wallet.h.
SelectCoinsMinConf is a bit of a misnomer now since it really just does
all of the coin selection given some parameters. So rename this to
something less annoying to say and makes a bit more sense.
-BEGIN VERIFY SCRIPT-
sed -i 's/SelectCoinsMinConf/AttemptSelection/g' $(git grep -l SelectCoinsMinConf ./src)
-END VERIFY SCRIPT-
The fees for transaction overhead and recipient outputs are now included
in nTargetValue instead of being a separate parameter. For the coin
selection algorithms, it doesn't matter that these are separate as in
either case, the algorithm needs to select enough to cover these fees.
Note that setting nValueToSelect is changed as it now includes
not_input_fees. Without the change to how nValueToSelect is increased
for KnapsackSolver, this would result in overpaying fees. The change to
increase by the difference between nFeeRet and not_input_fees allows
this to have the same behavior as previously.
Additionally, because we assume that KnapsackSolver will always find a
solution that requires change (we assume that BnB always finds a
non-change solution), we also include the fee for the change output in
KnapsackSolver's target. As part of this, we also use the changeless
nFeeRet when iterating for KnapsackSolver. This is because we include
the change fee when doing KnapsackSolver, so nFeeRet on further
iterations won't include the change fee.
f9cd2bfbcc Rename CoinSelectionParams::effective_fee to m_effective_feerate (Andrew Chow)
bdd0c2934b wallet: Move discard feerate fetching to CreateTransaction (Andrew Chow)
448d04b931 wallet: Move long term feerate setting to CreateTransaction (Andrew Chow)
e2f429e6bb wallet: Replace nFeeRateNeeded with effective_fee (Andrew Chow)
1a6a0b0dfb wallet: Use existing feerate instead of getting a new one (Andrew Chow)
Pull request description:
During coin selection, there are various places where we need to have a feerate. We need the feerate for the transaction itself, the discard fee rate, and long term feerate. Fetching these each time we need them can lead to a race condition where two feerates that should be the same are actually different. One particular instance where this can happen is during the loop in `CreateTransactionInternal`. After inputs are chosen, the expected transaction fee is calculated using a newly fetched feerate. If `pick_new_inputs == false`, the loop will go again with the assumption that the fee for the transaction remains the same. However because the feerate is fetched again, it is possible that it actually isn't and this causes coin selection to fail.
Instead of fetching the feerate each time it is needed, we fetch them all at once at the top of `CreateTransactionInternal`, store them in `CoinSelectionParams`, and use them where needed.
While some of these fee rates probably don't need this caching, I've done it for consistency and the guarantee that they remain the same.
Fixes #19229
ACKs for top commit:
glozow:
reACK f9cd2bfbcc
fjahr:
Code review re-ACK f9cd2bfbcc
Xekyo:
tACK f9cd2bfbcc
meshcollider:
Code review + test run ACK f9cd2bfbcc
Tree-SHA512: be83ff64ba473c3cdd3469c812e214659b6e2a9584c22ed2b1595618fce0d4b35d0901e61068cd1069fc1a8fb911db01dd7312d05c3b8cbafbe2504ab7a3e863
Instead of fetching the discard feerate for each SelectCoinsMinConf
iteration, fetch and cache it once during CreateTransaction so that it
is shared for each SelectCoinsMinConf through
coin_selection_params.m_discard_feerate.
Does not change behavior.
Instead of setting the long term feerate for each SelectCoinsMinConf
iteration, set it once during CreateTransaction and let it be shared
with each SelectCoinsMinConf through
coin_selection_params.m_long_term_feerate.
Does not change behavior.
5d4597666d Rewrite OutputGroups to be clearer and to use scriptPubKeys (Andrew Chow)
f6b3052739 Explicitly filter out partial groups when we don't want them (Andrew Chow)
416d74fb16 Move OutputGroup positive only filtering into Insert (Andrew Chow)
d895e98b59 Move EligibleForSpending into GroupOutputs (Andrew Chow)
99b399aba5 Move fee setting of OutputGroup to Insert (Andrew Chow)
6148a8acda Move GroupOutputs into SelectCoinsMinConf (Andrew Chow)
2acad03657 Remove OutputGroup non-default constructors (Andrew Chow)
Pull request description:
Even after #17458, we still deal with setting fees of an `OutputGroup` and filtering the `OutputGroup` outside of the struct. We currently make all of the `OutputGroup`s in `SelectCoins` and then copy and modify them within each `SelectCoinsMinConf` scenario. This PR changes this to constructing the `OutputGroup`s within the `SelectCoinsMinConf` so that the scenario can be taken into account during the group construction. Furthermore, setting of fees and filtering for effective value is moved into `OutputGroup::Insert` itself so that we don't add undesirable outputs to an `OutputGroup` rather than deleting them afterwards.
To facilitate fee calculation and effective value filtering during `OutputGroup::Insert`, `OutputGroup` now takes the feerates in its constructor and computes the fees and effective value for each output during `Insert`.
While removing `OutputGroup`s in accordance with the `CoinEligibilityFilter` still requires creating the `OutputGroup`s first, we can do that within the function that makes them - `GroupOutput`s.
ACKs for top commit:
Xekyo:
Code review ACK: 5d4597666d
fjahr:
Code review ACK 5d4597666d
meshcollider:
Light utACK 5d4597666d
Tree-SHA512: 35965b6d49a87f4ebb366ec4f00aafaaf78e9282481ae2c9682b515a3a9f2cbcd3cd6e202fee29489d48fe7f3a7cede4270796f5e72bbaff76da647138fb3059
This removes a source of complexity and indirection that makes it harder to
understand path checking code. Path checks will be simplified in upcoming
commits.
There is no change in behavior in this commit other than a slightly more
descriptive error message in `loadwallet` if the default "" wallet can't be
found. (The error message is improved more in upcoming commit "wallet: Remove
path checking code from loadwallet RPC".)
78c312c983 Replace current benchmarking framework with nanobench (Martin Ankerl)
Pull request description:
Replace current benchmarking framework with nanobench
This replaces the current benchmarking framework with nanobench [1], an
MIT licensed single-header benchmarking library, of which I am the
autor. This has in my opinion several advantages, especially on Linux:
* fast: Running all benchmarks takes ~6 seconds instead of 4m13s on
an Intel i7-8700 CPU @ 3.20GHz.
* accurate: I ran e.g. the benchmark for SipHash_32b 10 times and
calculate standard deviation / mean = coefficient of variation:
* 0.57% CV for old benchmarking framework
* 0.20% CV for nanobench
So the benchmark results with nanobench seem to vary less than with
the old framework.
* It automatically determines runtime based on clock precision, no need
to specify number of evaluations.
* measure instructions, cycles, branches, instructions per cycle,
branch misses (only Linux, when performance counters are available)
* output in markdown table format.
* Warn about unstable environment (frequency scaling, turbo, ...)
* For better profiling, it is possible to set the environment variable
NANOBENCH_ENDLESS to force endless running of a particular benchmark
without the need to recompile. This makes it to e.g. run "perf top"
and look at hotspots.
Here is an example copy & pasted from the terminal output:
| ns/byte | byte/s | err% | ins/byte | cyc/byte | IPC | bra/byte | miss% | total | benchmark
|--------------------:|--------------------:|--------:|----------------:|----------------:|-------:|---------------:|--------:|----------:|:----------
| 2.52 | 396,529,415.94 | 0.6% | 25.42 | 8.02 | 3.169 | 0.06 | 0.0% | 0.03 | `bench/crypto_hash.cpp RIPEMD160`
| 1.87 | 535,161,444.83 | 0.3% | 21.36 | 5.95 | 3.589 | 0.06 | 0.0% | 0.02 | `bench/crypto_hash.cpp SHA1`
| 3.22 | 310,344,174.79 | 1.1% | 36.80 | 10.22 | 3.601 | 0.09 | 0.0% | 0.04 | `bench/crypto_hash.cpp SHA256`
| 2.01 | 496,375,796.23 | 0.0% | 18.72 | 6.43 | 2.911 | 0.01 | 1.0% | 0.00 | `bench/crypto_hash.cpp SHA256D64_1024`
| 7.23 | 138,263,519.35 | 0.1% | 82.66 | 23.11 | 3.577 | 1.63 | 0.1% | 0.00 | `bench/crypto_hash.cpp SHA256_32b`
| 3.04 | 328,780,166.40 | 0.3% | 35.82 | 9.69 | 3.696 | 0.03 | 0.0% | 0.03 | `bench/crypto_hash.cpp SHA512`
[1] https://github.com/martinus/nanobench
ACKs for top commit:
laanwj:
ACK 78c312c983
Tree-SHA512: 9e18770b18b6f95a7d0105a4a5497d31cf4eb5efe6574f4482f6f1b4c88d7e0946b9a4a1e9e8e6ecbf41a3f2d7571240677dcb45af29a6f0584e89b25f32e49e
This replaces the current benchmarking framework with nanobench [1], an
MIT licensed single-header benchmarking library, of which I am the
autor. This has in my opinion several advantages, especially on Linux:
* fast: Running all benchmarks takes ~6 seconds instead of 4m13s on
an Intel i7-8700 CPU @ 3.20GHz.
* accurate: I ran e.g. the benchmark for SipHash_32b 10 times and
calculate standard deviation / mean = coefficient of variation:
* 0.57% CV for old benchmarking framework
* 0.20% CV for nanobench
So the benchmark results with nanobench seem to vary less than with
the old framework.
* It automatically determines runtime based on clock precision, no need
to specify number of evaluations.
* measure instructions, cycles, branches, instructions per cycle,
branch misses (only Linux, when performance counters are available)
* output in markdown table format.
* Warn about unstable environment (frequency scaling, turbo, ...)
* For better profiling, it is possible to set the environment variable
NANOBENCH_ENDLESS to force endless running of a particular benchmark
without the need to recompile. This makes it to e.g. run "perf top"
and look at hotspots.
Here is an example copy & pasted from the terminal output:
| ns/byte | byte/s | err% | ins/byte | cyc/byte | IPC | bra/byte | miss% | total | benchmark
|--------------------:|--------------------:|--------:|----------------:|----------------:|-------:|---------------:|--------:|----------:|:----------
| 2.52 | 396,529,415.94 | 0.6% | 25.42 | 8.02 | 3.169 | 0.06 | 0.0% | 0.03 | `bench/crypto_hash.cpp RIPEMD160`
| 1.87 | 535,161,444.83 | 0.3% | 21.36 | 5.95 | 3.589 | 0.06 | 0.0% | 0.02 | `bench/crypto_hash.cpp SHA1`
| 3.22 | 310,344,174.79 | 1.1% | 36.80 | 10.22 | 3.601 | 0.09 | 0.0% | 0.04 | `bench/crypto_hash.cpp SHA256`
| 2.01 | 496,375,796.23 | 0.0% | 18.72 | 6.43 | 2.911 | 0.01 | 1.0% | 0.00 | `bench/crypto_hash.cpp SHA256D64_1024`
| 7.23 | 138,263,519.35 | 0.1% | 82.66 | 23.11 | 3.577 | 1.63 | 0.1% | 0.00 | `bench/crypto_hash.cpp SHA256_32b`
| 3.04 | 328,780,166.40 | 0.3% | 35.82 | 9.69 | 3.696 | 0.03 | 0.0% | 0.03 | `bench/crypto_hash.cpp SHA512`
[1] https://github.com/martinus/nanobench
* Adds support for asymptotes
This adds support to calculate asymptotic complexity of a benchmark.
This is similar to #17375, but currently only one asymptote is
supported, and I have added support in the benchmark `ComplexMemPool`
as an example.
Usage is e.g. like this:
```
./bench_bitcoin -filter=ComplexMemPool -asymptote=25,50,100,200,400,600,800
```
This runs the benchmark `ComplexMemPool` several times but with
different complexityN settings. The benchmark can extract that number
and use it accordingly. Here, it's used for `childTxs`. The output is
this:
| complexityN | ns/op | op/s | err% | ins/op | cyc/op | IPC | total | benchmark
|------------:|--------------------:|--------------------:|--------:|----------------:|----------------:|-------:|----------:|:----------
| 25 | 1,064,241.00 | 939.64 | 1.4% | 3,960,279.00 | 2,829,708.00 | 1.400 | 0.01 | `ComplexMemPool`
| 50 | 1,579,530.00 | 633.10 | 1.0% | 6,231,810.00 | 4,412,674.00 | 1.412 | 0.02 | `ComplexMemPool`
| 100 | 4,022,774.00 | 248.58 | 0.6% | 16,544,406.00 | 11,889,535.00 | 1.392 | 0.04 | `ComplexMemPool`
| 200 | 15,390,986.00 | 64.97 | 0.2% | 63,904,254.00 | 47,731,705.00 | 1.339 | 0.17 | `ComplexMemPool`
| 400 | 69,394,711.00 | 14.41 | 0.1% | 272,602,461.00 | 219,014,691.00 | 1.245 | 0.76 | `ComplexMemPool`
| 600 | 168,977,165.00 | 5.92 | 0.1% | 639,108,082.00 | 535,316,887.00 | 1.194 | 1.86 | `ComplexMemPool`
| 800 | 310,109,077.00 | 3.22 | 0.1% |1,149,134,246.00 | 984,620,812.00 | 1.167 | 3.41 | `ComplexMemPool`
| coefficient | err% | complexity
|--------------:|-------:|------------
| 4.78486e-07 | 4.5% | O(n^2)
| 6.38557e-10 | 21.7% | O(n^3)
| 3.42338e-05 | 38.0% | O(n log n)
| 0.000313914 | 46.9% | O(n)
| 0.0129823 | 114.4% | O(log n)
| 0.0815055 | 133.8% | O(1)
The best fitting curve is O(n^2), so the algorithm seems to scale
quadratic with `childTxs` in the range 25 to 800.
This commit does not change behavior. All it does is pass new function
parameters.
It is easiest to review this change with:
git log -p -n1 -U0 --word-diff-regex=.
3fc20632a3 qt: Set BLOCK_CHAIN_SIZE = 220 (DrahtBot)
2b6a2f4a28 Regenerate manpages (DrahtBot)
eb7daf4d60 Update copyright headers to 2018 (DrahtBot)
Pull request description:
Some trivial maintenance to avoid having to do it again after the 0.17 branch off.
(The scripts to do this are in `./contrib/`)
Tree-SHA512: 16b2af45e0351b1c691c5311d48025dc6828079e98c2aa2e600dc5910ee8aa01858ca6c356538150dc46fe14c8819ed8ec8e4ec9a0f682b9950dd41bc50518fa
Allows SelectCoinsMinConf and SelectCoins be able to switch between
using BnB or Knapsack for choosing coins.
Has SelectCoinsMinConf do the preprocessing necessary to support either
BnB or Knapsack. This includes calculating the filtering the effective
values for each input.
Uses BnB in CreateTransaction to find an exact match for the output.
If BnB fails, it will fallback to the Knapsack solver.
Remove requirement that two wallet files can only be opened at the same time if
they are contained in the same directory.
This change mostly consists of updates to function signatures (updating
functions to take fs::path arguments, instead of combinations of strings,
fs::path, and CDBEnv / CWalletDBWrapper arguments).
* inline performance critical code
* Average runtime is specified and used to calculate iterations.
* Console: show median of multiple runs
* plot: show box plot
* filter benchmarks
* specify scaling factor
* ignore src/test and src/bench in command line check script
* number of iterations instead of time
* Replaced runtime in BENCHMARK makro number of iterations.
* Added -? to bench_bitcoin
* Benchmark plotly.js URL, width, height can be customized
* Fixed incorrect precision warning
