2016-02-10 00:12:00 +01:00
|
|
|
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
2017-07-16 01:03:42 +02:00
|
|
|
// This source code is licensed under both the GPLv2 (found in the
|
|
|
|
// COPYING file in the root directory) and Apache 2.0 License
|
|
|
|
// (found in the LICENSE.Apache file in the root directory).
|
2013-10-31 21:38:54 +01:00
|
|
|
|
2014-05-09 17:34:18 +02:00
|
|
|
#ifndef GFLAGS
|
|
|
|
#include <cstdio>
|
|
|
|
int main() {
|
|
|
|
fprintf(stderr, "Please install gflags to run rocksdb tools\n");
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
|
2017-04-06 04:02:00 +02:00
|
|
|
#include "db/db_impl.h"
|
|
|
|
#include "db/dbformat.h"
|
|
|
|
#include "monitoring/histogram.h"
|
2013-10-31 21:38:54 +01:00
|
|
|
#include "rocksdb/db.h"
|
2013-10-29 04:34:02 +01:00
|
|
|
#include "rocksdb/slice_transform.h"
|
2013-10-31 21:38:54 +01:00
|
|
|
#include "rocksdb/table.h"
|
|
|
|
#include "table/block_based_table_factory.h"
|
2017-04-06 04:02:00 +02:00
|
|
|
#include "table/get_context.h"
|
2015-10-13 00:06:38 +02:00
|
|
|
#include "table/internal_iterator.h"
|
2014-01-28 06:58:46 +01:00
|
|
|
#include "table/plain_table_factory.h"
|
2014-02-13 03:09:24 +01:00
|
|
|
#include "table/table_builder.h"
|
Move rate_limiter, write buffering, most perf context instrumentation and most random kill out of Env
Summary: We want to keep Env a think layer for better portability. Less platform dependent codes should be moved out of Env. In this patch, I create a wrapper of file readers and writers, and put rate limiting, write buffering, as well as most perf context instrumentation and random kill out of Env. It will make it easier to maintain multiple Env in the future.
Test Plan: Run all existing unit tests.
Reviewers: anthony, kradhakrishnan, IslamAbdelRahman, yhchiang, igor
Reviewed By: igor
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D42321
2015-07-18 01:16:11 +02:00
|
|
|
#include "util/file_reader_writer.h"
|
2017-12-01 19:40:45 +01:00
|
|
|
#include "util/gflags_compat.h"
|
2013-10-31 21:38:54 +01:00
|
|
|
#include "util/testharness.h"
|
|
|
|
#include "util/testutil.h"
|
|
|
|
|
2017-12-01 19:40:45 +01:00
|
|
|
using GFLAGS_NAMESPACE::ParseCommandLineFlags;
|
|
|
|
using GFLAGS_NAMESPACE::SetUsageMessage;
|
2014-05-09 17:34:18 +02:00
|
|
|
|
2013-10-31 21:38:54 +01:00
|
|
|
namespace rocksdb {
|
2014-04-10 06:17:14 +02:00
|
|
|
|
|
|
|
namespace {
|
2013-10-31 21:38:54 +01:00
|
|
|
// Make a key that i determines the first 4 characters and j determines the
|
|
|
|
// last 4 characters.
|
2013-11-16 07:23:12 +01:00
|
|
|
static std::string MakeKey(int i, int j, bool through_db) {
|
2013-10-31 21:38:54 +01:00
|
|
|
char buf[100];
|
2013-11-16 07:23:12 +01:00
|
|
|
snprintf(buf, sizeof(buf), "%04d__key___%04d", i, j);
|
|
|
|
if (through_db) {
|
|
|
|
return std::string(buf);
|
|
|
|
}
|
|
|
|
// If we directly query table, which operates on internal keys
|
|
|
|
// instead of user keys, we need to add 8 bytes of internal
|
|
|
|
// information (row type etc) to user key to make an internal
|
|
|
|
// key.
|
|
|
|
InternalKey key(std::string(buf), 0, ValueType::kTypeValue);
|
|
|
|
return key.Encode().ToString();
|
2013-10-31 21:38:54 +01:00
|
|
|
}
|
|
|
|
|
Benchmark table reader wiht nanoseconds
Summary: nanosecnods gave us better view of the performance, especially when some operations are fast so that micro seconds may only reveal less informative results.
Test Plan:
sample output:
./table_reader_bench --plain_table --time_unit=nanosecond
=======================================================================================================
InMemoryTableSimpleBenchmark: PlainTable num_key1: 4096 num_key2: 512 non_empty
=======================================================================================================
Histogram (unit: nanosecond):
Count: 6291456 Average: 475.3867 StdDev: 556.05
Min: 135.0000 Median: 400.1817 Max: 33370.0000
Percentiles: P50: 400.18 P75: 530.02 P99: 887.73 P99.9: 8843.26 P99.99: 9941.21
------------------------------------------------------
[ 120, 140 ) 2 0.000% 0.000%
[ 140, 160 ) 452 0.007% 0.007%
[ 160, 180 ) 13683 0.217% 0.225%
[ 180, 200 ) 54353 0.864% 1.089%
[ 200, 250 ) 101004 1.605% 2.694%
[ 250, 300 ) 729791 11.600% 14.294% ##
[ 300, 350 ) 616070 9.792% 24.086% ##
[ 350, 400 ) 1628021 25.877% 49.963% #####
[ 400, 450 ) 647220 10.287% 60.250% ##
[ 450, 500 ) 577206 9.174% 69.424% ##
[ 500, 600 ) 1168585 18.574% 87.999% ####
[ 600, 700 ) 506875 8.057% 96.055% ##
[ 700, 800 ) 147878 2.350% 98.406%
[ 800, 900 ) 42633 0.678% 99.083%
[ 900, 1000 ) 16304 0.259% 99.342%
[ 1000, 1200 ) 7811 0.124% 99.466%
[ 1200, 1400 ) 1453 0.023% 99.490%
[ 1400, 1600 ) 307 0.005% 99.494%
[ 1600, 1800 ) 81 0.001% 99.496%
[ 1800, 2000 ) 18 0.000% 99.496%
[ 2000, 2500 ) 8 0.000% 99.496%
[ 2500, 3000 ) 6 0.000% 99.496%
[ 3500, 4000 ) 3 0.000% 99.496%
[ 4000, 4500 ) 116 0.002% 99.498%
[ 4500, 5000 ) 1144 0.018% 99.516%
[ 5000, 6000 ) 1087 0.017% 99.534%
[ 6000, 7000 ) 2403 0.038% 99.572%
[ 7000, 8000 ) 9840 0.156% 99.728%
[ 8000, 9000 ) 12820 0.204% 99.932%
[ 9000, 10000 ) 3881 0.062% 99.994%
[ 10000, 12000 ) 135 0.002% 99.996%
[ 12000, 14000 ) 159 0.003% 99.998%
[ 14000, 16000 ) 58 0.001% 99.999%
[ 16000, 18000 ) 30 0.000% 100.000%
[ 18000, 20000 ) 14 0.000% 100.000%
[ 20000, 25000 ) 2 0.000% 100.000%
[ 25000, 30000 ) 2 0.000% 100.000%
[ 30000, 35000 ) 1 0.000% 100.000%
Reviewers: haobo, dhruba, sdong
CC: leveldb
Differential Revision: https://reviews.facebook.net/D16113
2014-02-13 22:55:04 +01:00
|
|
|
uint64_t Now(Env* env, bool measured_by_nanosecond) {
|
|
|
|
return measured_by_nanosecond ? env->NowNanos() : env->NowMicros();
|
|
|
|
}
|
2014-04-10 06:17:14 +02:00
|
|
|
} // namespace
|
Benchmark table reader wiht nanoseconds
Summary: nanosecnods gave us better view of the performance, especially when some operations are fast so that micro seconds may only reveal less informative results.
Test Plan:
sample output:
./table_reader_bench --plain_table --time_unit=nanosecond
=======================================================================================================
InMemoryTableSimpleBenchmark: PlainTable num_key1: 4096 num_key2: 512 non_empty
=======================================================================================================
Histogram (unit: nanosecond):
Count: 6291456 Average: 475.3867 StdDev: 556.05
Min: 135.0000 Median: 400.1817 Max: 33370.0000
Percentiles: P50: 400.18 P75: 530.02 P99: 887.73 P99.9: 8843.26 P99.99: 9941.21
------------------------------------------------------
[ 120, 140 ) 2 0.000% 0.000%
[ 140, 160 ) 452 0.007% 0.007%
[ 160, 180 ) 13683 0.217% 0.225%
[ 180, 200 ) 54353 0.864% 1.089%
[ 200, 250 ) 101004 1.605% 2.694%
[ 250, 300 ) 729791 11.600% 14.294% ##
[ 300, 350 ) 616070 9.792% 24.086% ##
[ 350, 400 ) 1628021 25.877% 49.963% #####
[ 400, 450 ) 647220 10.287% 60.250% ##
[ 450, 500 ) 577206 9.174% 69.424% ##
[ 500, 600 ) 1168585 18.574% 87.999% ####
[ 600, 700 ) 506875 8.057% 96.055% ##
[ 700, 800 ) 147878 2.350% 98.406%
[ 800, 900 ) 42633 0.678% 99.083%
[ 900, 1000 ) 16304 0.259% 99.342%
[ 1000, 1200 ) 7811 0.124% 99.466%
[ 1200, 1400 ) 1453 0.023% 99.490%
[ 1400, 1600 ) 307 0.005% 99.494%
[ 1600, 1800 ) 81 0.001% 99.496%
[ 1800, 2000 ) 18 0.000% 99.496%
[ 2000, 2500 ) 8 0.000% 99.496%
[ 2500, 3000 ) 6 0.000% 99.496%
[ 3500, 4000 ) 3 0.000% 99.496%
[ 4000, 4500 ) 116 0.002% 99.498%
[ 4500, 5000 ) 1144 0.018% 99.516%
[ 5000, 6000 ) 1087 0.017% 99.534%
[ 6000, 7000 ) 2403 0.038% 99.572%
[ 7000, 8000 ) 9840 0.156% 99.728%
[ 8000, 9000 ) 12820 0.204% 99.932%
[ 9000, 10000 ) 3881 0.062% 99.994%
[ 10000, 12000 ) 135 0.002% 99.996%
[ 12000, 14000 ) 159 0.003% 99.998%
[ 14000, 16000 ) 58 0.001% 99.999%
[ 16000, 18000 ) 30 0.000% 100.000%
[ 18000, 20000 ) 14 0.000% 100.000%
[ 20000, 25000 ) 2 0.000% 100.000%
[ 25000, 30000 ) 2 0.000% 100.000%
[ 30000, 35000 ) 1 0.000% 100.000%
Reviewers: haobo, dhruba, sdong
CC: leveldb
Differential Revision: https://reviews.facebook.net/D16113
2014-02-13 22:55:04 +01:00
|
|
|
|
2013-10-31 21:38:54 +01:00
|
|
|
// A very simple benchmark that.
|
|
|
|
// Create a table with roughly numKey1 * numKey2 keys,
|
|
|
|
// where there are numKey1 prefixes of the key, each has numKey2 number of
|
|
|
|
// distinguished key, differing in the suffix part.
|
|
|
|
// If if_query_empty_keys = false, query the existing keys numKey1 * numKey2
|
|
|
|
// times randomly.
|
|
|
|
// If if_query_empty_keys = true, query numKey1 * numKey2 random empty keys.
|
|
|
|
// Print out the total time.
|
2013-11-16 07:23:12 +01:00
|
|
|
// If through_db=true, a full DB will be created and queries will be against
|
|
|
|
// it. Otherwise, operations will be directly through table level.
|
2013-10-31 21:38:54 +01:00
|
|
|
//
|
|
|
|
// If for_terator=true, instead of just query one key each time, it queries
|
|
|
|
// a range sharing the same prefix.
|
2014-04-10 06:17:14 +02:00
|
|
|
namespace {
|
2013-10-31 21:38:54 +01:00
|
|
|
void TableReaderBenchmark(Options& opts, EnvOptions& env_options,
|
2013-11-16 07:23:12 +01:00
|
|
|
ReadOptions& read_options, int num_keys1,
|
2018-04-13 02:55:14 +02:00
|
|
|
int num_keys2, int num_iter, int /*prefix_len*/,
|
2013-11-16 07:23:12 +01:00
|
|
|
bool if_query_empty_keys, bool for_iterator,
|
Benchmark table reader wiht nanoseconds
Summary: nanosecnods gave us better view of the performance, especially when some operations are fast so that micro seconds may only reveal less informative results.
Test Plan:
sample output:
./table_reader_bench --plain_table --time_unit=nanosecond
=======================================================================================================
InMemoryTableSimpleBenchmark: PlainTable num_key1: 4096 num_key2: 512 non_empty
=======================================================================================================
Histogram (unit: nanosecond):
Count: 6291456 Average: 475.3867 StdDev: 556.05
Min: 135.0000 Median: 400.1817 Max: 33370.0000
Percentiles: P50: 400.18 P75: 530.02 P99: 887.73 P99.9: 8843.26 P99.99: 9941.21
------------------------------------------------------
[ 120, 140 ) 2 0.000% 0.000%
[ 140, 160 ) 452 0.007% 0.007%
[ 160, 180 ) 13683 0.217% 0.225%
[ 180, 200 ) 54353 0.864% 1.089%
[ 200, 250 ) 101004 1.605% 2.694%
[ 250, 300 ) 729791 11.600% 14.294% ##
[ 300, 350 ) 616070 9.792% 24.086% ##
[ 350, 400 ) 1628021 25.877% 49.963% #####
[ 400, 450 ) 647220 10.287% 60.250% ##
[ 450, 500 ) 577206 9.174% 69.424% ##
[ 500, 600 ) 1168585 18.574% 87.999% ####
[ 600, 700 ) 506875 8.057% 96.055% ##
[ 700, 800 ) 147878 2.350% 98.406%
[ 800, 900 ) 42633 0.678% 99.083%
[ 900, 1000 ) 16304 0.259% 99.342%
[ 1000, 1200 ) 7811 0.124% 99.466%
[ 1200, 1400 ) 1453 0.023% 99.490%
[ 1400, 1600 ) 307 0.005% 99.494%
[ 1600, 1800 ) 81 0.001% 99.496%
[ 1800, 2000 ) 18 0.000% 99.496%
[ 2000, 2500 ) 8 0.000% 99.496%
[ 2500, 3000 ) 6 0.000% 99.496%
[ 3500, 4000 ) 3 0.000% 99.496%
[ 4000, 4500 ) 116 0.002% 99.498%
[ 4500, 5000 ) 1144 0.018% 99.516%
[ 5000, 6000 ) 1087 0.017% 99.534%
[ 6000, 7000 ) 2403 0.038% 99.572%
[ 7000, 8000 ) 9840 0.156% 99.728%
[ 8000, 9000 ) 12820 0.204% 99.932%
[ 9000, 10000 ) 3881 0.062% 99.994%
[ 10000, 12000 ) 135 0.002% 99.996%
[ 12000, 14000 ) 159 0.003% 99.998%
[ 14000, 16000 ) 58 0.001% 99.999%
[ 16000, 18000 ) 30 0.000% 100.000%
[ 18000, 20000 ) 14 0.000% 100.000%
[ 20000, 25000 ) 2 0.000% 100.000%
[ 25000, 30000 ) 2 0.000% 100.000%
[ 30000, 35000 ) 1 0.000% 100.000%
Reviewers: haobo, dhruba, sdong
CC: leveldb
Differential Revision: https://reviews.facebook.net/D16113
2014-02-13 22:55:04 +01:00
|
|
|
bool through_db, bool measured_by_nanosecond) {
|
2014-02-13 03:09:24 +01:00
|
|
|
rocksdb::InternalKeyComparator ikc(opts.comparator);
|
|
|
|
|
2018-07-14 02:18:39 +02:00
|
|
|
std::string file_name =
|
|
|
|
test::PerThreadDBPath("rocksdb_table_reader_benchmark");
|
|
|
|
std::string dbname = test::PerThreadDBPath("rocksdb_table_reader_bench_db");
|
2013-11-16 07:23:12 +01:00
|
|
|
WriteOptions wo;
|
2013-10-31 21:38:54 +01:00
|
|
|
Env* env = Env::Default();
|
2013-11-16 07:23:12 +01:00
|
|
|
TableBuilder* tb = nullptr;
|
|
|
|
DB* db = nullptr;
|
|
|
|
Status s;
|
2014-09-05 01:18:36 +02:00
|
|
|
const ImmutableCFOptions ioptions(opts);
|
2018-05-21 23:33:55 +02:00
|
|
|
const ColumnFamilyOptions cfo(opts);
|
|
|
|
const MutableCFOptions moptions(cfo);
|
2018-11-09 20:17:34 +01:00
|
|
|
std::unique_ptr<WritableFileWriter> file_writer;
|
2013-11-16 07:23:12 +01:00
|
|
|
if (!through_db) {
|
2018-11-09 20:17:34 +01:00
|
|
|
std::unique_ptr<WritableFile> file;
|
2013-11-16 07:23:12 +01:00
|
|
|
env->NewWritableFile(file_name, &file, env_options);
|
A new call back to TablePropertiesCollector to allow users know the entry is add, delete or merge
Summary:
Currently users have no idea a key is add, delete or merge from TablePropertiesCollector call back. Add a new function to add it.
Also refactor the codes so that
(1) make table property collector and internal table property collector two separate data structures with the later one now exposed
(2) table builders only receive internal table properties
Test Plan: Add cases in table_properties_collector_test to cover both of old and new ways of using TablePropertiesCollector.
Reviewers: yhchiang, igor.sugak, rven, igor
Reviewed By: rven, igor
Subscribers: meyering, yoshinorim, maykov, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D35373
2015-04-06 19:04:30 +02:00
|
|
|
|
|
|
|
std::vector<std::unique_ptr<IntTblPropCollectorFactory> >
|
|
|
|
int_tbl_prop_collector_factories;
|
|
|
|
|
2018-08-23 19:04:10 +02:00
|
|
|
file_writer.reset(
|
|
|
|
new WritableFileWriter(std::move(file), file_name, env_options));
|
2016-09-18 07:30:43 +02:00
|
|
|
int unknown_level = -1;
|
A new call back to TablePropertiesCollector to allow users know the entry is add, delete or merge
Summary:
Currently users have no idea a key is add, delete or merge from TablePropertiesCollector call back. Add a new function to add it.
Also refactor the codes so that
(1) make table property collector and internal table property collector two separate data structures with the later one now exposed
(2) table builders only receive internal table properties
Test Plan: Add cases in table_properties_collector_test to cover both of old and new ways of using TablePropertiesCollector.
Reviewers: yhchiang, igor.sugak, rven, igor
Reviewed By: rven, igor
Subscribers: meyering, yoshinorim, maykov, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D35373
2015-04-06 19:04:30 +02:00
|
|
|
tb = opts.table_factory->NewTableBuilder(
|
2018-05-21 23:33:55 +02:00
|
|
|
TableBuilderOptions(
|
|
|
|
ioptions, moptions, ikc, &int_tbl_prop_collector_factories,
|
|
|
|
CompressionType::kNoCompression, CompressionOptions(),
|
Reduce scope of compression dictionary to single SST (#4952)
Summary:
Our previous approach was to train one compression dictionary per compaction, using the first output SST to train a dictionary, and then applying it on subsequent SSTs in the same compaction. While this was great for minimizing CPU/memory/I/O overhead, it did not achieve good compression ratios in practice. In our most promising potential use case, moderate reductions in a dictionary's scope make a major difference on compression ratio.
So, this PR changes compression dictionary to be scoped per-SST. It accepts the tradeoff during table building to use more memory and CPU. Important changes include:
- The `BlockBasedTableBuilder` has a new state when dictionary compression is in-use: `kBuffered`. In that state it accumulates uncompressed data in-memory whenever `Add` is called.
- After accumulating target file size bytes or calling `BlockBasedTableBuilder::Finish`, a `BlockBasedTableBuilder` moves to the `kUnbuffered` state. The transition (`EnterUnbuffered()`) involves sampling the buffered data, training a dictionary, and compressing/writing out all buffered data. In the `kUnbuffered` state, a `BlockBasedTableBuilder` behaves the same as before -- blocks are compressed/written out as soon as they fill up.
- Samples are now whole uncompressed data blocks, except the final sample may be a partial data block so we don't breach the user's configured `max_dict_bytes` or `zstd_max_train_bytes`. The dictionary trainer is supposed to work better when we pass it real units of compression. Previously we were passing 64-byte KV samples which was not realistic.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/4952
Differential Revision: D13967980
Pulled By: ajkr
fbshipit-source-id: 82bea6f7537e1529c7a1a4cdee84585f5949300f
2019-02-12 04:42:25 +01:00
|
|
|
false /* skip_filters */, kDefaultColumnFamilyName, unknown_level),
|
2016-04-07 08:10:32 +02:00
|
|
|
0 /* column_family_id */, file_writer.get());
|
2013-11-16 07:23:12 +01:00
|
|
|
} else {
|
|
|
|
s = DB::Open(opts, dbname, &db);
|
|
|
|
ASSERT_OK(s);
|
|
|
|
ASSERT_TRUE(db != nullptr);
|
|
|
|
}
|
2013-10-31 21:38:54 +01:00
|
|
|
// Populate slightly more than 1M keys
|
|
|
|
for (int i = 0; i < num_keys1; i++) {
|
|
|
|
for (int j = 0; j < num_keys2; j++) {
|
2013-11-16 07:23:12 +01:00
|
|
|
std::string key = MakeKey(i * 2, j, through_db);
|
|
|
|
if (!through_db) {
|
|
|
|
tb->Add(key, key);
|
|
|
|
} else {
|
|
|
|
db->Put(wo, key, key);
|
|
|
|
}
|
2013-10-31 21:38:54 +01:00
|
|
|
}
|
|
|
|
}
|
2013-11-16 07:23:12 +01:00
|
|
|
if (!through_db) {
|
|
|
|
tb->Finish();
|
2015-09-17 01:57:43 +02:00
|
|
|
file_writer->Close();
|
2013-11-16 07:23:12 +01:00
|
|
|
} else {
|
|
|
|
db->Flush(FlushOptions());
|
|
|
|
}
|
2013-10-31 21:38:54 +01:00
|
|
|
|
2018-11-09 20:17:34 +01:00
|
|
|
std::unique_ptr<TableReader> table_reader;
|
2013-11-16 07:23:12 +01:00
|
|
|
if (!through_db) {
|
2018-11-09 20:17:34 +01:00
|
|
|
std::unique_ptr<RandomAccessFile> raf;
|
2014-11-08 00:04:30 +01:00
|
|
|
s = env->NewRandomAccessFile(file_name, &raf, env_options);
|
2015-09-17 01:57:43 +02:00
|
|
|
if (!s.ok()) {
|
|
|
|
fprintf(stderr, "Create File Error: %s\n", s.ToString().c_str());
|
|
|
|
exit(1);
|
|
|
|
}
|
2013-11-16 07:23:12 +01:00
|
|
|
uint64_t file_size;
|
|
|
|
env->GetFileSize(file_name, &file_size);
|
2018-11-09 20:17:34 +01:00
|
|
|
std::unique_ptr<RandomAccessFileReader> file_reader(
|
2017-06-29 06:26:03 +02:00
|
|
|
new RandomAccessFileReader(std::move(raf), file_name));
|
2015-09-11 20:36:33 +02:00
|
|
|
s = opts.table_factory->NewTableReader(
|
2018-05-21 23:33:55 +02:00
|
|
|
TableReaderOptions(ioptions, moptions.prefix_extractor.get(),
|
|
|
|
env_options, ikc),
|
|
|
|
std::move(file_reader), file_size, &table_reader);
|
2015-09-17 01:57:43 +02:00
|
|
|
if (!s.ok()) {
|
|
|
|
fprintf(stderr, "Open Table Error: %s\n", s.ToString().c_str());
|
|
|
|
exit(1);
|
|
|
|
}
|
2013-11-16 07:23:12 +01:00
|
|
|
}
|
2013-10-31 21:38:54 +01:00
|
|
|
|
|
|
|
Random rnd(301);
|
2013-11-16 07:23:12 +01:00
|
|
|
std::string result;
|
2013-10-31 21:38:54 +01:00
|
|
|
HistogramImpl hist;
|
|
|
|
|
|
|
|
for (int it = 0; it < num_iter; it++) {
|
|
|
|
for (int i = 0; i < num_keys1; i++) {
|
|
|
|
for (int j = 0; j < num_keys2; j++) {
|
|
|
|
int r1 = rnd.Uniform(num_keys1) * 2;
|
|
|
|
int r2 = rnd.Uniform(num_keys2);
|
2013-11-16 07:23:12 +01:00
|
|
|
if (if_query_empty_keys) {
|
|
|
|
r1++;
|
|
|
|
r2 = num_keys2 * 2 - r2;
|
|
|
|
}
|
|
|
|
|
2013-10-31 21:38:54 +01:00
|
|
|
if (!for_iterator) {
|
|
|
|
// Query one existing key;
|
2013-11-16 07:23:12 +01:00
|
|
|
std::string key = MakeKey(r1, r2, through_db);
|
2014-02-14 00:27:59 +01:00
|
|
|
uint64_t start_time = Now(env, measured_by_nanosecond);
|
2013-11-16 07:23:12 +01:00
|
|
|
if (!through_db) {
|
2017-03-13 19:44:50 +01:00
|
|
|
PinnableSlice value;
|
2014-09-29 20:09:09 +02:00
|
|
|
MergeContext merge_context;
|
Use only "local" range tombstones during Get (#4449)
Summary:
Previously, range tombstones were accumulated from every level, which
was necessary if a range tombstone in a higher level covered a key in a lower
level. However, RangeDelAggregator::AddTombstones's complexity is based on
the number of tombstones that are currently stored in it, which is wasteful in
the Get case, where we only need to know the highest sequence number of range
tombstones that cover the key from higher levels, and compute the highest covering
sequence number at the current level. This change introduces this optimization, and
removes the use of RangeDelAggregator from the Get path.
In the benchmark results, the following command was used to initialize the database:
```
./db_bench -db=/dev/shm/5k-rts -use_existing_db=false -benchmarks=filluniquerandom -write_buffer_size=1048576 -compression_type=lz4 -target_file_size_base=1048576 -max_bytes_for_level_base=4194304 -value_size=112 -key_size=16 -block_size=4096 -level_compaction_dynamic_level_bytes=true -num=5000000 -max_background_jobs=12 -benchmark_write_rate_limit=20971520 -range_tombstone_width=100 -writes_per_range_tombstone=100 -max_num_range_tombstones=50000 -bloom_bits=8
```
...and the following command was used to measure read throughput:
```
./db_bench -db=/dev/shm/5k-rts/ -use_existing_db=true -benchmarks=readrandom -disable_auto_compactions=true -num=5000000 -reads=100000 -threads=32
```
The filluniquerandom command was only run once, and the resulting database was used
to measure read performance before and after the PR. Both binaries were compiled with
`DEBUG_LEVEL=0`.
Readrandom results before PR:
```
readrandom : 4.544 micros/op 220090 ops/sec; 16.9 MB/s (63103 of 100000 found)
```
Readrandom results after PR:
```
readrandom : 11.147 micros/op 89707 ops/sec; 6.9 MB/s (63103 of 100000 found)
```
So it's actually slower right now, but this PR paves the way for future optimizations (see #4493).
----
Pull Request resolved: https://github.com/facebook/rocksdb/pull/4449
Differential Revision: D10370575
Pulled By: abhimadan
fbshipit-source-id: 9a2e152be1ef36969055c0e9eb4beb0d96c11f4d
2018-10-24 21:29:29 +02:00
|
|
|
SequenceNumber max_covering_tombstone_seq = 0;
|
2016-11-04 02:40:23 +01:00
|
|
|
GetContext get_context(ioptions.user_comparator,
|
|
|
|
ioptions.merge_operator, ioptions.info_log,
|
|
|
|
ioptions.statistics, GetContext::kNotFound,
|
|
|
|
Slice(key), &value, nullptr, &merge_context,
|
Use only "local" range tombstones during Get (#4449)
Summary:
Previously, range tombstones were accumulated from every level, which
was necessary if a range tombstone in a higher level covered a key in a lower
level. However, RangeDelAggregator::AddTombstones's complexity is based on
the number of tombstones that are currently stored in it, which is wasteful in
the Get case, where we only need to know the highest sequence number of range
tombstones that cover the key from higher levels, and compute the highest covering
sequence number at the current level. This change introduces this optimization, and
removes the use of RangeDelAggregator from the Get path.
In the benchmark results, the following command was used to initialize the database:
```
./db_bench -db=/dev/shm/5k-rts -use_existing_db=false -benchmarks=filluniquerandom -write_buffer_size=1048576 -compression_type=lz4 -target_file_size_base=1048576 -max_bytes_for_level_base=4194304 -value_size=112 -key_size=16 -block_size=4096 -level_compaction_dynamic_level_bytes=true -num=5000000 -max_background_jobs=12 -benchmark_write_rate_limit=20971520 -range_tombstone_width=100 -writes_per_range_tombstone=100 -max_num_range_tombstones=50000 -bloom_bits=8
```
...and the following command was used to measure read throughput:
```
./db_bench -db=/dev/shm/5k-rts/ -use_existing_db=true -benchmarks=readrandom -disable_auto_compactions=true -num=5000000 -reads=100000 -threads=32
```
The filluniquerandom command was only run once, and the resulting database was used
to measure read performance before and after the PR. Both binaries were compiled with
`DEBUG_LEVEL=0`.
Readrandom results before PR:
```
readrandom : 4.544 micros/op 220090 ops/sec; 16.9 MB/s (63103 of 100000 found)
```
Readrandom results after PR:
```
readrandom : 11.147 micros/op 89707 ops/sec; 6.9 MB/s (63103 of 100000 found)
```
So it's actually slower right now, but this PR paves the way for future optimizations (see #4493).
----
Pull Request resolved: https://github.com/facebook/rocksdb/pull/4449
Differential Revision: D10370575
Pulled By: abhimadan
fbshipit-source-id: 9a2e152be1ef36969055c0e9eb4beb0d96c11f4d
2018-10-24 21:29:29 +02:00
|
|
|
&max_covering_tombstone_seq, env);
|
2018-05-21 23:33:55 +02:00
|
|
|
s = table_reader->Get(read_options, key, &get_context, nullptr);
|
2013-11-16 07:23:12 +01:00
|
|
|
} else {
|
2014-02-13 03:09:24 +01:00
|
|
|
s = db->Get(read_options, key, &result);
|
2013-11-16 07:23:12 +01:00
|
|
|
}
|
2014-02-14 00:27:59 +01:00
|
|
|
hist.Add(Now(env, measured_by_nanosecond) - start_time);
|
2013-10-31 21:38:54 +01:00
|
|
|
} else {
|
2013-11-16 07:23:12 +01:00
|
|
|
int r2_len;
|
|
|
|
if (if_query_empty_keys) {
|
|
|
|
r2_len = 0;
|
|
|
|
} else {
|
|
|
|
r2_len = rnd.Uniform(num_keys2) + 1;
|
|
|
|
if (r2_len + r2 > num_keys2) {
|
|
|
|
r2_len = num_keys2 - r2;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
std::string start_key = MakeKey(r1, r2, through_db);
|
|
|
|
std::string end_key = MakeKey(r1, r2 + r2_len, through_db);
|
2013-10-31 23:26:06 +01:00
|
|
|
uint64_t total_time = 0;
|
2014-02-14 00:27:59 +01:00
|
|
|
uint64_t start_time = Now(env, measured_by_nanosecond);
|
2015-10-13 00:06:38 +02:00
|
|
|
Iterator* iter = nullptr;
|
|
|
|
InternalIterator* iiter = nullptr;
|
2013-11-16 07:23:12 +01:00
|
|
|
if (!through_db) {
|
2018-05-21 23:33:55 +02:00
|
|
|
iiter = table_reader->NewIterator(read_options, nullptr);
|
2013-11-16 07:23:12 +01:00
|
|
|
} else {
|
|
|
|
iter = db->NewIterator(read_options);
|
|
|
|
}
|
2013-10-31 21:38:54 +01:00
|
|
|
int count = 0;
|
2015-10-13 00:06:38 +02:00
|
|
|
for (through_db ? iter->Seek(start_key) : iiter->Seek(start_key);
|
|
|
|
through_db ? iter->Valid() : iiter->Valid();
|
|
|
|
through_db ? iter->Next() : iiter->Next()) {
|
2013-11-16 07:23:12 +01:00
|
|
|
if (if_query_empty_keys) {
|
|
|
|
break;
|
|
|
|
}
|
2013-10-31 21:38:54 +01:00
|
|
|
// verify key;
|
2014-02-14 00:27:59 +01:00
|
|
|
total_time += Now(env, measured_by_nanosecond) - start_time;
|
plain table reader: non-mmap mode to keep two recent buffers
Summary: In plain table reader's non-mmap mode, we only keep the most recent read buffer. However, for binary search, it is likely we come back to a location to read. To avoid one pread in such a case, we keep two read buffers. It should cover most of the cases.
Test Plan:
1. run tests
2. check the optimization works through strace when running
./table_reader_bench -mmap_read=false --num_keys2=1 -num_keys1=5000 -table_factory=plain_table --iterator --through_db
Reviewers: anthony, rven, kradhakrishnan, igor, yhchiang, IslamAbdelRahman
Reviewed By: IslamAbdelRahman
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D51171
2015-12-24 02:30:10 +01:00
|
|
|
assert(Slice(MakeKey(r1, r2 + count, through_db)) ==
|
|
|
|
(through_db ? iter->key() : iiter->key()));
|
2014-02-14 00:27:59 +01:00
|
|
|
start_time = Now(env, measured_by_nanosecond);
|
2013-10-31 21:38:54 +01:00
|
|
|
if (++count >= r2_len) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (count != r2_len) {
|
|
|
|
fprintf(
|
|
|
|
stderr, "Iterator cannot iterate expected number of entries. "
|
|
|
|
"Expected %d but got %d\n", r2_len, count);
|
|
|
|
assert(false);
|
|
|
|
}
|
|
|
|
delete iter;
|
2014-02-14 00:27:59 +01:00
|
|
|
total_time += Now(env, measured_by_nanosecond) - start_time;
|
2013-10-31 23:26:06 +01:00
|
|
|
hist.Add(total_time);
|
2013-10-31 21:38:54 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(
|
|
|
|
stderr,
|
|
|
|
"==================================================="
|
|
|
|
"====================================================\n"
|
|
|
|
"InMemoryTableSimpleBenchmark: %20s num_key1: %5d "
|
|
|
|
"num_key2: %5d %10s\n"
|
|
|
|
"==================================================="
|
|
|
|
"===================================================="
|
Benchmark table reader wiht nanoseconds
Summary: nanosecnods gave us better view of the performance, especially when some operations are fast so that micro seconds may only reveal less informative results.
Test Plan:
sample output:
./table_reader_bench --plain_table --time_unit=nanosecond
=======================================================================================================
InMemoryTableSimpleBenchmark: PlainTable num_key1: 4096 num_key2: 512 non_empty
=======================================================================================================
Histogram (unit: nanosecond):
Count: 6291456 Average: 475.3867 StdDev: 556.05
Min: 135.0000 Median: 400.1817 Max: 33370.0000
Percentiles: P50: 400.18 P75: 530.02 P99: 887.73 P99.9: 8843.26 P99.99: 9941.21
------------------------------------------------------
[ 120, 140 ) 2 0.000% 0.000%
[ 140, 160 ) 452 0.007% 0.007%
[ 160, 180 ) 13683 0.217% 0.225%
[ 180, 200 ) 54353 0.864% 1.089%
[ 200, 250 ) 101004 1.605% 2.694%
[ 250, 300 ) 729791 11.600% 14.294% ##
[ 300, 350 ) 616070 9.792% 24.086% ##
[ 350, 400 ) 1628021 25.877% 49.963% #####
[ 400, 450 ) 647220 10.287% 60.250% ##
[ 450, 500 ) 577206 9.174% 69.424% ##
[ 500, 600 ) 1168585 18.574% 87.999% ####
[ 600, 700 ) 506875 8.057% 96.055% ##
[ 700, 800 ) 147878 2.350% 98.406%
[ 800, 900 ) 42633 0.678% 99.083%
[ 900, 1000 ) 16304 0.259% 99.342%
[ 1000, 1200 ) 7811 0.124% 99.466%
[ 1200, 1400 ) 1453 0.023% 99.490%
[ 1400, 1600 ) 307 0.005% 99.494%
[ 1600, 1800 ) 81 0.001% 99.496%
[ 1800, 2000 ) 18 0.000% 99.496%
[ 2000, 2500 ) 8 0.000% 99.496%
[ 2500, 3000 ) 6 0.000% 99.496%
[ 3500, 4000 ) 3 0.000% 99.496%
[ 4000, 4500 ) 116 0.002% 99.498%
[ 4500, 5000 ) 1144 0.018% 99.516%
[ 5000, 6000 ) 1087 0.017% 99.534%
[ 6000, 7000 ) 2403 0.038% 99.572%
[ 7000, 8000 ) 9840 0.156% 99.728%
[ 8000, 9000 ) 12820 0.204% 99.932%
[ 9000, 10000 ) 3881 0.062% 99.994%
[ 10000, 12000 ) 135 0.002% 99.996%
[ 12000, 14000 ) 159 0.003% 99.998%
[ 14000, 16000 ) 58 0.001% 99.999%
[ 16000, 18000 ) 30 0.000% 100.000%
[ 18000, 20000 ) 14 0.000% 100.000%
[ 20000, 25000 ) 2 0.000% 100.000%
[ 25000, 30000 ) 2 0.000% 100.000%
[ 30000, 35000 ) 1 0.000% 100.000%
Reviewers: haobo, dhruba, sdong
CC: leveldb
Differential Revision: https://reviews.facebook.net/D16113
2014-02-13 22:55:04 +01:00
|
|
|
"\nHistogram (unit: %s): \n%s",
|
2013-11-16 07:23:12 +01:00
|
|
|
opts.table_factory->Name(), num_keys1, num_keys2,
|
Benchmark table reader wiht nanoseconds
Summary: nanosecnods gave us better view of the performance, especially when some operations are fast so that micro seconds may only reveal less informative results.
Test Plan:
sample output:
./table_reader_bench --plain_table --time_unit=nanosecond
=======================================================================================================
InMemoryTableSimpleBenchmark: PlainTable num_key1: 4096 num_key2: 512 non_empty
=======================================================================================================
Histogram (unit: nanosecond):
Count: 6291456 Average: 475.3867 StdDev: 556.05
Min: 135.0000 Median: 400.1817 Max: 33370.0000
Percentiles: P50: 400.18 P75: 530.02 P99: 887.73 P99.9: 8843.26 P99.99: 9941.21
------------------------------------------------------
[ 120, 140 ) 2 0.000% 0.000%
[ 140, 160 ) 452 0.007% 0.007%
[ 160, 180 ) 13683 0.217% 0.225%
[ 180, 200 ) 54353 0.864% 1.089%
[ 200, 250 ) 101004 1.605% 2.694%
[ 250, 300 ) 729791 11.600% 14.294% ##
[ 300, 350 ) 616070 9.792% 24.086% ##
[ 350, 400 ) 1628021 25.877% 49.963% #####
[ 400, 450 ) 647220 10.287% 60.250% ##
[ 450, 500 ) 577206 9.174% 69.424% ##
[ 500, 600 ) 1168585 18.574% 87.999% ####
[ 600, 700 ) 506875 8.057% 96.055% ##
[ 700, 800 ) 147878 2.350% 98.406%
[ 800, 900 ) 42633 0.678% 99.083%
[ 900, 1000 ) 16304 0.259% 99.342%
[ 1000, 1200 ) 7811 0.124% 99.466%
[ 1200, 1400 ) 1453 0.023% 99.490%
[ 1400, 1600 ) 307 0.005% 99.494%
[ 1600, 1800 ) 81 0.001% 99.496%
[ 1800, 2000 ) 18 0.000% 99.496%
[ 2000, 2500 ) 8 0.000% 99.496%
[ 2500, 3000 ) 6 0.000% 99.496%
[ 3500, 4000 ) 3 0.000% 99.496%
[ 4000, 4500 ) 116 0.002% 99.498%
[ 4500, 5000 ) 1144 0.018% 99.516%
[ 5000, 6000 ) 1087 0.017% 99.534%
[ 6000, 7000 ) 2403 0.038% 99.572%
[ 7000, 8000 ) 9840 0.156% 99.728%
[ 8000, 9000 ) 12820 0.204% 99.932%
[ 9000, 10000 ) 3881 0.062% 99.994%
[ 10000, 12000 ) 135 0.002% 99.996%
[ 12000, 14000 ) 159 0.003% 99.998%
[ 14000, 16000 ) 58 0.001% 99.999%
[ 16000, 18000 ) 30 0.000% 100.000%
[ 18000, 20000 ) 14 0.000% 100.000%
[ 20000, 25000 ) 2 0.000% 100.000%
[ 25000, 30000 ) 2 0.000% 100.000%
[ 30000, 35000 ) 1 0.000% 100.000%
Reviewers: haobo, dhruba, sdong
CC: leveldb
Differential Revision: https://reviews.facebook.net/D16113
2014-02-13 22:55:04 +01:00
|
|
|
for_iterator ? "iterator" : (if_query_empty_keys ? "empty" : "non_empty"),
|
|
|
|
measured_by_nanosecond ? "nanosecond" : "microsecond",
|
2013-10-31 21:38:54 +01:00
|
|
|
hist.ToString().c_str());
|
2013-11-16 07:23:12 +01:00
|
|
|
if (!through_db) {
|
|
|
|
env->DeleteFile(file_name);
|
|
|
|
} else {
|
|
|
|
delete db;
|
|
|
|
db = nullptr;
|
|
|
|
DestroyDB(dbname, opts);
|
|
|
|
}
|
2013-10-31 21:38:54 +01:00
|
|
|
}
|
2014-04-10 06:17:14 +02:00
|
|
|
} // namespace
|
Benchmark table reader wiht nanoseconds
Summary: nanosecnods gave us better view of the performance, especially when some operations are fast so that micro seconds may only reveal less informative results.
Test Plan:
sample output:
./table_reader_bench --plain_table --time_unit=nanosecond
=======================================================================================================
InMemoryTableSimpleBenchmark: PlainTable num_key1: 4096 num_key2: 512 non_empty
=======================================================================================================
Histogram (unit: nanosecond):
Count: 6291456 Average: 475.3867 StdDev: 556.05
Min: 135.0000 Median: 400.1817 Max: 33370.0000
Percentiles: P50: 400.18 P75: 530.02 P99: 887.73 P99.9: 8843.26 P99.99: 9941.21
------------------------------------------------------
[ 120, 140 ) 2 0.000% 0.000%
[ 140, 160 ) 452 0.007% 0.007%
[ 160, 180 ) 13683 0.217% 0.225%
[ 180, 200 ) 54353 0.864% 1.089%
[ 200, 250 ) 101004 1.605% 2.694%
[ 250, 300 ) 729791 11.600% 14.294% ##
[ 300, 350 ) 616070 9.792% 24.086% ##
[ 350, 400 ) 1628021 25.877% 49.963% #####
[ 400, 450 ) 647220 10.287% 60.250% ##
[ 450, 500 ) 577206 9.174% 69.424% ##
[ 500, 600 ) 1168585 18.574% 87.999% ####
[ 600, 700 ) 506875 8.057% 96.055% ##
[ 700, 800 ) 147878 2.350% 98.406%
[ 800, 900 ) 42633 0.678% 99.083%
[ 900, 1000 ) 16304 0.259% 99.342%
[ 1000, 1200 ) 7811 0.124% 99.466%
[ 1200, 1400 ) 1453 0.023% 99.490%
[ 1400, 1600 ) 307 0.005% 99.494%
[ 1600, 1800 ) 81 0.001% 99.496%
[ 1800, 2000 ) 18 0.000% 99.496%
[ 2000, 2500 ) 8 0.000% 99.496%
[ 2500, 3000 ) 6 0.000% 99.496%
[ 3500, 4000 ) 3 0.000% 99.496%
[ 4000, 4500 ) 116 0.002% 99.498%
[ 4500, 5000 ) 1144 0.018% 99.516%
[ 5000, 6000 ) 1087 0.017% 99.534%
[ 6000, 7000 ) 2403 0.038% 99.572%
[ 7000, 8000 ) 9840 0.156% 99.728%
[ 8000, 9000 ) 12820 0.204% 99.932%
[ 9000, 10000 ) 3881 0.062% 99.994%
[ 10000, 12000 ) 135 0.002% 99.996%
[ 12000, 14000 ) 159 0.003% 99.998%
[ 14000, 16000 ) 58 0.001% 99.999%
[ 16000, 18000 ) 30 0.000% 100.000%
[ 18000, 20000 ) 14 0.000% 100.000%
[ 20000, 25000 ) 2 0.000% 100.000%
[ 25000, 30000 ) 2 0.000% 100.000%
[ 30000, 35000 ) 1 0.000% 100.000%
Reviewers: haobo, dhruba, sdong
CC: leveldb
Differential Revision: https://reviews.facebook.net/D16113
2014-02-13 22:55:04 +01:00
|
|
|
} // namespace rocksdb
|
2013-10-31 21:38:54 +01:00
|
|
|
|
|
|
|
DEFINE_bool(query_empty, false, "query non-existing keys instead of existing "
|
|
|
|
"ones.");
|
|
|
|
DEFINE_int32(num_keys1, 4096, "number of distinguish prefix of keys");
|
|
|
|
DEFINE_int32(num_keys2, 512, "number of distinguish keys for each prefix");
|
|
|
|
DEFINE_int32(iter, 3, "query non-existing keys instead of existing ones");
|
2013-11-16 07:23:12 +01:00
|
|
|
DEFINE_int32(prefix_len, 16, "Prefix length used for iterators and indexes");
|
2013-10-31 21:38:54 +01:00
|
|
|
DEFINE_bool(iterator, false, "For test iterator");
|
2013-11-16 07:23:12 +01:00
|
|
|
DEFINE_bool(through_db, false, "If enable, a DB instance will be created and "
|
|
|
|
"the query will be against DB. Otherwise, will be directly against "
|
|
|
|
"a table reader.");
|
2015-11-18 03:29:40 +01:00
|
|
|
DEFINE_bool(mmap_read, true, "Whether use mmap read");
|
2014-08-19 21:50:13 +02:00
|
|
|
DEFINE_string(table_factory, "block_based",
|
|
|
|
"Table factory to use: `block_based` (default), `plain_table` or "
|
|
|
|
"`cuckoo_hash`.");
|
Benchmark table reader wiht nanoseconds
Summary: nanosecnods gave us better view of the performance, especially when some operations are fast so that micro seconds may only reveal less informative results.
Test Plan:
sample output:
./table_reader_bench --plain_table --time_unit=nanosecond
=======================================================================================================
InMemoryTableSimpleBenchmark: PlainTable num_key1: 4096 num_key2: 512 non_empty
=======================================================================================================
Histogram (unit: nanosecond):
Count: 6291456 Average: 475.3867 StdDev: 556.05
Min: 135.0000 Median: 400.1817 Max: 33370.0000
Percentiles: P50: 400.18 P75: 530.02 P99: 887.73 P99.9: 8843.26 P99.99: 9941.21
------------------------------------------------------
[ 120, 140 ) 2 0.000% 0.000%
[ 140, 160 ) 452 0.007% 0.007%
[ 160, 180 ) 13683 0.217% 0.225%
[ 180, 200 ) 54353 0.864% 1.089%
[ 200, 250 ) 101004 1.605% 2.694%
[ 250, 300 ) 729791 11.600% 14.294% ##
[ 300, 350 ) 616070 9.792% 24.086% ##
[ 350, 400 ) 1628021 25.877% 49.963% #####
[ 400, 450 ) 647220 10.287% 60.250% ##
[ 450, 500 ) 577206 9.174% 69.424% ##
[ 500, 600 ) 1168585 18.574% 87.999% ####
[ 600, 700 ) 506875 8.057% 96.055% ##
[ 700, 800 ) 147878 2.350% 98.406%
[ 800, 900 ) 42633 0.678% 99.083%
[ 900, 1000 ) 16304 0.259% 99.342%
[ 1000, 1200 ) 7811 0.124% 99.466%
[ 1200, 1400 ) 1453 0.023% 99.490%
[ 1400, 1600 ) 307 0.005% 99.494%
[ 1600, 1800 ) 81 0.001% 99.496%
[ 1800, 2000 ) 18 0.000% 99.496%
[ 2000, 2500 ) 8 0.000% 99.496%
[ 2500, 3000 ) 6 0.000% 99.496%
[ 3500, 4000 ) 3 0.000% 99.496%
[ 4000, 4500 ) 116 0.002% 99.498%
[ 4500, 5000 ) 1144 0.018% 99.516%
[ 5000, 6000 ) 1087 0.017% 99.534%
[ 6000, 7000 ) 2403 0.038% 99.572%
[ 7000, 8000 ) 9840 0.156% 99.728%
[ 8000, 9000 ) 12820 0.204% 99.932%
[ 9000, 10000 ) 3881 0.062% 99.994%
[ 10000, 12000 ) 135 0.002% 99.996%
[ 12000, 14000 ) 159 0.003% 99.998%
[ 14000, 16000 ) 58 0.001% 99.999%
[ 16000, 18000 ) 30 0.000% 100.000%
[ 18000, 20000 ) 14 0.000% 100.000%
[ 20000, 25000 ) 2 0.000% 100.000%
[ 25000, 30000 ) 2 0.000% 100.000%
[ 30000, 35000 ) 1 0.000% 100.000%
Reviewers: haobo, dhruba, sdong
CC: leveldb
Differential Revision: https://reviews.facebook.net/D16113
2014-02-13 22:55:04 +01:00
|
|
|
DEFINE_string(time_unit, "microsecond",
|
|
|
|
"The time unit used for measuring performance. User can specify "
|
|
|
|
"`microsecond` (default) or `nanosecond`");
|
2013-10-31 21:38:54 +01:00
|
|
|
|
|
|
|
int main(int argc, char** argv) {
|
2014-05-09 17:34:18 +02:00
|
|
|
SetUsageMessage(std::string("\nUSAGE:\n") + std::string(argv[0]) +
|
|
|
|
" [OPTIONS]...");
|
|
|
|
ParseCommandLineFlags(&argc, &argv, true);
|
2013-10-31 21:38:54 +01:00
|
|
|
|
2014-08-19 21:50:13 +02:00
|
|
|
std::shared_ptr<rocksdb::TableFactory> tf;
|
2013-10-31 21:38:54 +01:00
|
|
|
rocksdb::Options options;
|
2013-11-16 07:23:12 +01:00
|
|
|
if (FLAGS_prefix_len < 16) {
|
2014-03-10 20:56:46 +01:00
|
|
|
options.prefix_extractor.reset(rocksdb::NewFixedPrefixTransform(
|
|
|
|
FLAGS_prefix_len));
|
2013-11-16 07:23:12 +01:00
|
|
|
}
|
2013-10-31 21:38:54 +01:00
|
|
|
rocksdb::ReadOptions ro;
|
|
|
|
rocksdb::EnvOptions env_options;
|
2013-11-16 07:23:12 +01:00
|
|
|
options.create_if_missing = true;
|
2013-10-29 04:34:02 +01:00
|
|
|
options.compression = rocksdb::CompressionType::kNoCompression;
|
|
|
|
|
2014-08-19 21:50:13 +02:00
|
|
|
if (FLAGS_table_factory == "cuckoo_hash") {
|
2014-11-12 22:05:12 +01:00
|
|
|
#ifndef ROCKSDB_LITE
|
2015-11-18 03:29:40 +01:00
|
|
|
options.allow_mmap_reads = FLAGS_mmap_read;
|
|
|
|
env_options.use_mmap_reads = FLAGS_mmap_read;
|
CuckooTable: add one option to allow identity function for the first hash function
Summary:
MurmurHash becomes expensive when we do millions Get() a second in one
thread. Add this option to allow the first hash function to use identity
function as hash function. It results in QPS increase from 3.7M/s to
~4.3M/s. I did not observe improvement for end to end RocksDB
performance. This may be caused by other bottlenecks that I will address
in a separate diff.
Test Plan:
```
[ljin@dev1964 rocksdb] ./cuckoo_table_reader_test --enable_perf --file_dir=/dev/shm --write --identity_as_first_hash=0
==== Test CuckooReaderTest.WhenKeyExists
==== Test CuckooReaderTest.WhenKeyExistsWithUint64Comparator
==== Test CuckooReaderTest.CheckIterator
==== Test CuckooReaderTest.CheckIteratorUint64
==== Test CuckooReaderTest.WhenKeyNotFound
==== Test CuckooReaderTest.TestReadPerformance
With 125829120 items, utilization is 93.75%, number of hash functions: 2.
Time taken per op is 0.272us (3.7 Mqps) with batch size of 0, # of found keys 125829120
With 125829120 items, utilization is 93.75%, number of hash functions: 2.
Time taken per op is 0.138us (7.2 Mqps) with batch size of 10, # of found keys 125829120
With 125829120 items, utilization is 93.75%, number of hash functions: 2.
Time taken per op is 0.142us (7.1 Mqps) with batch size of 25, # of found keys 125829120
With 125829120 items, utilization is 93.75%, number of hash functions: 2.
Time taken per op is 0.142us (7.0 Mqps) with batch size of 50, # of found keys 125829120
With 125829120 items, utilization is 93.75%, number of hash functions: 2.
Time taken per op is 0.144us (6.9 Mqps) with batch size of 100, # of found keys 125829120
With 104857600 items, utilization is 78.12%, number of hash functions: 2.
Time taken per op is 0.201us (5.0 Mqps) with batch size of 0, # of found keys 104857600
With 104857600 items, utilization is 78.12%, number of hash functions: 2.
Time taken per op is 0.121us (8.3 Mqps) with batch size of 10, # of found keys 104857600
With 104857600 items, utilization is 78.12%, number of hash functions: 2.
Time taken per op is 0.123us (8.1 Mqps) with batch size of 25, # of found keys 104857600
With 104857600 items, utilization is 78.12%, number of hash functions: 2.
Time taken per op is 0.121us (8.3 Mqps) with batch size of 50, # of found keys 104857600
With 104857600 items, utilization is 78.12%, number of hash functions: 2.
Time taken per op is 0.112us (8.9 Mqps) with batch size of 100, # of found keys 104857600
With 83886080 items, utilization is 62.50%, number of hash functions: 2.
Time taken per op is 0.251us (4.0 Mqps) with batch size of 0, # of found keys 83886080
With 83886080 items, utilization is 62.50%, number of hash functions: 2.
Time taken per op is 0.107us (9.4 Mqps) with batch size of 10, # of found keys 83886080
With 83886080 items, utilization is 62.50%, number of hash functions: 2.
Time taken per op is 0.099us (10.1 Mqps) with batch size of 25, # of found keys 83886080
With 83886080 items, utilization is 62.50%, number of hash functions: 2.
Time taken per op is 0.100us (10.0 Mqps) with batch size of 50, # of found keys 83886080
With 83886080 items, utilization is 62.50%, number of hash functions: 2.
Time taken per op is 0.116us (8.6 Mqps) with batch size of 100, # of found keys 83886080
With 73400320 items, utilization is 54.69%, number of hash functions: 2.
Time taken per op is 0.189us (5.3 Mqps) with batch size of 0, # of found keys 73400320
With 73400320 items, utilization is 54.69%, number of hash functions: 2.
Time taken per op is 0.095us (10.5 Mqps) with batch size of 10, # of found keys 73400320
With 73400320 items, utilization is 54.69%, number of hash functions: 2.
Time taken per op is 0.096us (10.4 Mqps) with batch size of 25, # of found keys 73400320
With 73400320 items, utilization is 54.69%, number of hash functions: 2.
Time taken per op is 0.098us (10.2 Mqps) with batch size of 50, # of found keys 73400320
With 73400320 items, utilization is 54.69%, number of hash functions: 2.
Time taken per op is 0.105us (9.5 Mqps) with batch size of 100, # of found keys 73400320
[ljin@dev1964 rocksdb] ./cuckoo_table_reader_test --enable_perf --file_dir=/dev/shm --write --identity_as_first_hash=1
==== Test CuckooReaderTest.WhenKeyExists
==== Test CuckooReaderTest.WhenKeyExistsWithUint64Comparator
==== Test CuckooReaderTest.CheckIterator
==== Test CuckooReaderTest.CheckIteratorUint64
==== Test CuckooReaderTest.WhenKeyNotFound
==== Test CuckooReaderTest.TestReadPerformance
With 125829120 items, utilization is 93.75%, number of hash functions: 2.
Time taken per op is 0.230us (4.3 Mqps) with batch size of 0, # of found keys 125829120
With 125829120 items, utilization is 93.75%, number of hash functions: 2.
Time taken per op is 0.086us (11.7 Mqps) with batch size of 10, # of found keys 125829120
With 125829120 items, utilization is 93.75%, number of hash functions: 2.
Time taken per op is 0.088us (11.3 Mqps) with batch size of 25, # of found keys 125829120
With 125829120 items, utilization is 93.75%, number of hash functions: 2.
Time taken per op is 0.083us (12.1 Mqps) with batch size of 50, # of found keys 125829120
With 125829120 items, utilization is 93.75%, number of hash functions: 2.
Time taken per op is 0.083us (12.1 Mqps) with batch size of 100, # of found keys 125829120
With 104857600 items, utilization is 78.12%, number of hash functions: 2.
Time taken per op is 0.159us (6.3 Mqps) with batch size of 0, # of found keys 104857600
With 104857600 items, utilization is 78.12%, number of hash functions: 2.
Time taken per op is 0.078us (12.8 Mqps) with batch size of 10, # of found keys 104857600
With 104857600 items, utilization is 78.12%, number of hash functions: 2.
Time taken per op is 0.080us (12.6 Mqps) with batch size of 25, # of found keys 104857600
With 104857600 items, utilization is 78.12%, number of hash functions: 2.
Time taken per op is 0.080us (12.5 Mqps) with batch size of 50, # of found keys 104857600
With 104857600 items, utilization is 78.12%, number of hash functions: 2.
Time taken per op is 0.082us (12.2 Mqps) with batch size of 100, # of found keys 104857600
With 83886080 items, utilization is 62.50%, number of hash functions: 2.
Time taken per op is 0.154us (6.5 Mqps) with batch size of 0, # of found keys 83886080
With 83886080 items, utilization is 62.50%, number of hash functions: 2.
Time taken per op is 0.077us (13.0 Mqps) with batch size of 10, # of found keys 83886080
With 83886080 items, utilization is 62.50%, number of hash functions: 2.
Time taken per op is 0.077us (12.9 Mqps) with batch size of 25, # of found keys 83886080
With 83886080 items, utilization is 62.50%, number of hash functions: 2.
Time taken per op is 0.078us (12.8 Mqps) with batch size of 50, # of found keys 83886080
With 83886080 items, utilization is 62.50%, number of hash functions: 2.
Time taken per op is 0.079us (12.6 Mqps) with batch size of 100, # of found keys 83886080
With 73400320 items, utilization is 54.69%, number of hash functions: 2.
Time taken per op is 0.218us (4.6 Mqps) with batch size of 0, # of found keys 73400320
With 73400320 items, utilization is 54.69%, number of hash functions: 2.
Time taken per op is 0.083us (12.0 Mqps) with batch size of 10, # of found keys 73400320
With 73400320 items, utilization is 54.69%, number of hash functions: 2.
Time taken per op is 0.085us (11.7 Mqps) with batch size of 25, # of found keys 73400320
With 73400320 items, utilization is 54.69%, number of hash functions: 2.
Time taken per op is 0.086us (11.6 Mqps) with batch size of 50, # of found keys 73400320
With 73400320 items, utilization is 54.69%, number of hash functions: 2.
Time taken per op is 0.078us (12.8 Mqps) with batch size of 100, # of found keys 73400320
```
Reviewers: sdong, igor, yhchiang
Reviewed By: igor
Subscribers: leveldb
Differential Revision: https://reviews.facebook.net/D23451
2014-09-18 20:00:48 +02:00
|
|
|
rocksdb::CuckooTableOptions table_options;
|
|
|
|
table_options.hash_table_ratio = 0.75;
|
|
|
|
tf.reset(rocksdb::NewCuckooTableFactory(table_options));
|
2014-11-12 22:05:12 +01:00
|
|
|
#else
|
|
|
|
fprintf(stderr, "Plain table is not supported in lite mode\n");
|
|
|
|
exit(1);
|
|
|
|
#endif // ROCKSDB_LITE
|
2014-08-19 21:50:13 +02:00
|
|
|
} else if (FLAGS_table_factory == "plain_table") {
|
2014-11-12 22:05:12 +01:00
|
|
|
#ifndef ROCKSDB_LITE
|
2015-11-18 03:29:40 +01:00
|
|
|
options.allow_mmap_reads = FLAGS_mmap_read;
|
|
|
|
env_options.use_mmap_reads = FLAGS_mmap_read;
|
2014-07-18 09:08:38 +02:00
|
|
|
|
|
|
|
rocksdb::PlainTableOptions plain_table_options;
|
|
|
|
plain_table_options.user_key_len = 16;
|
|
|
|
plain_table_options.bloom_bits_per_key = (FLAGS_prefix_len == 16) ? 0 : 8;
|
|
|
|
plain_table_options.hash_table_ratio = 0.75;
|
|
|
|
|
2014-08-19 21:50:13 +02:00
|
|
|
tf.reset(new rocksdb::PlainTableFactory(plain_table_options));
|
2014-03-10 20:56:46 +01:00
|
|
|
options.prefix_extractor.reset(rocksdb::NewFixedPrefixTransform(
|
|
|
|
FLAGS_prefix_len));
|
2014-11-12 22:05:12 +01:00
|
|
|
#else
|
|
|
|
fprintf(stderr, "Cuckoo table is not supported in lite mode\n");
|
|
|
|
exit(1);
|
|
|
|
#endif // ROCKSDB_LITE
|
2014-08-19 21:50:13 +02:00
|
|
|
} else if (FLAGS_table_factory == "block_based") {
|
|
|
|
tf.reset(new rocksdb::BlockBasedTableFactory());
|
|
|
|
} else {
|
|
|
|
fprintf(stderr, "Invalid table type %s\n", FLAGS_table_factory.c_str());
|
|
|
|
}
|
|
|
|
|
|
|
|
if (tf) {
|
|
|
|
// if user provides invalid options, just fall back to microsecond.
|
|
|
|
bool measured_by_nanosecond = FLAGS_time_unit == "nanosecond";
|
|
|
|
|
|
|
|
options.table_factory = tf;
|
|
|
|
rocksdb::TableReaderBenchmark(options, env_options, ro, FLAGS_num_keys1,
|
|
|
|
FLAGS_num_keys2, FLAGS_iter, FLAGS_prefix_len,
|
|
|
|
FLAGS_query_empty, FLAGS_iterator,
|
|
|
|
FLAGS_through_db, measured_by_nanosecond);
|
2013-10-29 04:34:02 +01:00
|
|
|
} else {
|
2014-08-19 21:50:13 +02:00
|
|
|
return 1;
|
2013-10-29 04:34:02 +01:00
|
|
|
}
|
Benchmark table reader wiht nanoseconds
Summary: nanosecnods gave us better view of the performance, especially when some operations are fast so that micro seconds may only reveal less informative results.
Test Plan:
sample output:
./table_reader_bench --plain_table --time_unit=nanosecond
=======================================================================================================
InMemoryTableSimpleBenchmark: PlainTable num_key1: 4096 num_key2: 512 non_empty
=======================================================================================================
Histogram (unit: nanosecond):
Count: 6291456 Average: 475.3867 StdDev: 556.05
Min: 135.0000 Median: 400.1817 Max: 33370.0000
Percentiles: P50: 400.18 P75: 530.02 P99: 887.73 P99.9: 8843.26 P99.99: 9941.21
------------------------------------------------------
[ 120, 140 ) 2 0.000% 0.000%
[ 140, 160 ) 452 0.007% 0.007%
[ 160, 180 ) 13683 0.217% 0.225%
[ 180, 200 ) 54353 0.864% 1.089%
[ 200, 250 ) 101004 1.605% 2.694%
[ 250, 300 ) 729791 11.600% 14.294% ##
[ 300, 350 ) 616070 9.792% 24.086% ##
[ 350, 400 ) 1628021 25.877% 49.963% #####
[ 400, 450 ) 647220 10.287% 60.250% ##
[ 450, 500 ) 577206 9.174% 69.424% ##
[ 500, 600 ) 1168585 18.574% 87.999% ####
[ 600, 700 ) 506875 8.057% 96.055% ##
[ 700, 800 ) 147878 2.350% 98.406%
[ 800, 900 ) 42633 0.678% 99.083%
[ 900, 1000 ) 16304 0.259% 99.342%
[ 1000, 1200 ) 7811 0.124% 99.466%
[ 1200, 1400 ) 1453 0.023% 99.490%
[ 1400, 1600 ) 307 0.005% 99.494%
[ 1600, 1800 ) 81 0.001% 99.496%
[ 1800, 2000 ) 18 0.000% 99.496%
[ 2000, 2500 ) 8 0.000% 99.496%
[ 2500, 3000 ) 6 0.000% 99.496%
[ 3500, 4000 ) 3 0.000% 99.496%
[ 4000, 4500 ) 116 0.002% 99.498%
[ 4500, 5000 ) 1144 0.018% 99.516%
[ 5000, 6000 ) 1087 0.017% 99.534%
[ 6000, 7000 ) 2403 0.038% 99.572%
[ 7000, 8000 ) 9840 0.156% 99.728%
[ 8000, 9000 ) 12820 0.204% 99.932%
[ 9000, 10000 ) 3881 0.062% 99.994%
[ 10000, 12000 ) 135 0.002% 99.996%
[ 12000, 14000 ) 159 0.003% 99.998%
[ 14000, 16000 ) 58 0.001% 99.999%
[ 16000, 18000 ) 30 0.000% 100.000%
[ 18000, 20000 ) 14 0.000% 100.000%
[ 20000, 25000 ) 2 0.000% 100.000%
[ 25000, 30000 ) 2 0.000% 100.000%
[ 30000, 35000 ) 1 0.000% 100.000%
Reviewers: haobo, dhruba, sdong
CC: leveldb
Differential Revision: https://reviews.facebook.net/D16113
2014-02-13 22:55:04 +01:00
|
|
|
|
2013-10-31 21:38:54 +01:00
|
|
|
return 0;
|
|
|
|
}
|
2014-05-09 17:34:18 +02:00
|
|
|
|
|
|
|
#endif // GFLAGS
|