2019-10-08 05:09:27 +02:00
|
|
|
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
|
|
|
// 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).
|
|
|
|
|
2019-10-28 22:10:51 +01:00
|
|
|
#if !defined(GFLAGS) || defined(ROCKSDB_LITE)
|
2019-10-08 05:09:27 +02:00
|
|
|
#include <cstdio>
|
|
|
|
int main() {
|
2019-10-28 22:10:51 +01:00
|
|
|
fprintf(stderr, "filter_bench requires gflags and !ROCKSDB_LITE\n");
|
2019-10-08 05:09:27 +02:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
|
|
|
|
#include <cinttypes>
|
|
|
|
#include <iostream>
|
2019-10-24 22:07:09 +02:00
|
|
|
#include <sstream>
|
2019-10-08 05:09:27 +02:00
|
|
|
#include <vector>
|
|
|
|
|
2019-10-25 22:25:28 +02:00
|
|
|
#include "memory/arena.h"
|
2019-10-08 05:09:27 +02:00
|
|
|
#include "port/port.h"
|
|
|
|
#include "port/stack_trace.h"
|
2021-01-26 07:07:26 +01:00
|
|
|
#include "rocksdb/system_clock.h"
|
New Bloom filter implementation for full and partitioned filters (#6007)
Summary:
Adds an improved, replacement Bloom filter implementation (FastLocalBloom) for full and partitioned filters in the block-based table. This replacement is faster and more accurate, especially for high bits per key or millions of keys in a single filter.
Speed
The improved speed, at least on recent x86_64, comes from
* Using fastrange instead of modulo (%)
* Using our new hash function (XXH3 preview, added in a previous commit), which is much faster for large keys and only *slightly* slower on keys around 12 bytes if hashing the same size many thousands of times in a row.
* Optimizing the Bloom filter queries with AVX2 SIMD operations. (Added AVX2 to the USE_SSE=1 build.) Careful design was required to support (a) SIMD-optimized queries, (b) compatible non-SIMD code that's simple and efficient, (c) flexible choice of number of probes, and (d) essentially maximized accuracy for a cache-local Bloom filter. Probes are made eight at a time, so any number of probes up to 8 is the same speed, then up to 16, etc.
* Prefetching cache lines when building the filter. Although this optimization could be applied to the old structure as well, it seems to balance out the small added cost of accumulating 64 bit hashes for adding to the filter rather than 32 bit hashes.
Here's nominal speed data from filter_bench (200MB in filters, about 10k keys each, 10 bits filter data / key, 6 probes, avg key size 24 bytes, includes hashing time) on Skylake DE (relatively low clock speed):
$ ./filter_bench -quick -impl=2 -net_includes_hashing # New Bloom filter
Build avg ns/key: 47.7135
Mixed inside/outside queries...
Single filter net ns/op: 26.2825
Random filter net ns/op: 150.459
Average FP rate %: 0.954651
$ ./filter_bench -quick -impl=0 -net_includes_hashing # Old Bloom filter
Build avg ns/key: 47.2245
Mixed inside/outside queries...
Single filter net ns/op: 63.2978
Random filter net ns/op: 188.038
Average FP rate %: 1.13823
Similar build time but dramatically faster query times on hot data (63 ns to 26 ns), and somewhat faster on stale data (188 ns to 150 ns). Performance differences on batched and skewed query loads are between these extremes as expected.
The only other interesting thing about speed is "inside" (query key was added to filter) vs. "outside" (query key was not added to filter) query times. The non-SIMD implementations are substantially slower when most queries are "outside" vs. "inside". This goes against what one might expect or would have observed years ago, as "outside" queries only need about two probes on average, due to short-circuiting, while "inside" always have num_probes (say 6). The problem is probably the nastily unpredictable branch. The SIMD implementation has few branches (very predictable) and has pretty consistent running time regardless of query outcome.
Accuracy
The generally improved accuracy (re: Issue https://github.com/facebook/rocksdb/issues/5857) comes from a better design for probing indices
within a cache line (re: Issue https://github.com/facebook/rocksdb/issues/4120) and improved accuracy for millions of keys in a single filter from using a 64-bit hash function (XXH3p). Design details in code comments.
Accuracy data (generalizes, except old impl gets worse with millions of keys):
Memory bits per key: FP rate percent old impl -> FP rate percent new impl
6: 5.70953 -> 5.69888
8: 2.45766 -> 2.29709
10: 1.13977 -> 0.959254
12: 0.662498 -> 0.411593
16: 0.353023 -> 0.0873754
24: 0.261552 -> 0.0060971
50: 0.225453 -> ~0.00003 (less than 1 in a million queries are FP)
Fixes https://github.com/facebook/rocksdb/issues/5857
Fixes https://github.com/facebook/rocksdb/issues/4120
Unlike the old implementation, this implementation has a fixed cache line size (64 bytes). At 10 bits per key, the accuracy of this new implementation is very close to the old implementation with 128-byte cache line size. If there's sufficient demand, this implementation could be generalized.
Compatibility
Although old releases would see the new structure as corrupt filter data and read the table as if there's no filter, we've decided only to enable the new Bloom filter with new format_version=5. This provides a smooth path for automatic adoption over time, with an option for early opt-in.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6007
Test Plan: filter_bench has been used thoroughly to validate speed, accuracy, and correctness. Unit tests have been carefully updated to exercise new and old implementations, as well as the logic to select an implementation based on context (format_version).
Differential Revision: D18294749
Pulled By: pdillinger
fbshipit-source-id: d44c9db3696e4d0a17caaec47075b7755c262c5f
2019-11-14 01:31:26 +01:00
|
|
|
#include "table/block_based/filter_policy_internal.h"
|
2019-10-08 05:09:27 +02:00
|
|
|
#include "table/block_based/full_filter_block.h"
|
|
|
|
#include "table/block_based/mock_block_based_table.h"
|
2019-10-25 22:25:28 +02:00
|
|
|
#include "table/plain/plain_table_bloom.h"
|
2020-04-29 22:06:27 +02:00
|
|
|
#include "util/cast_util.h"
|
2019-10-08 05:09:27 +02:00
|
|
|
#include "util/gflags_compat.h"
|
|
|
|
#include "util/hash.h"
|
2019-10-09 04:19:43 +02:00
|
|
|
#include "util/random.h"
|
2020-01-18 04:36:09 +01:00
|
|
|
#include "util/stderr_logger.h"
|
2019-10-08 05:09:27 +02:00
|
|
|
#include "util/stop_watch.h"
|
|
|
|
|
|
|
|
using GFLAGS_NAMESPACE::ParseCommandLineFlags;
|
|
|
|
using GFLAGS_NAMESPACE::RegisterFlagValidator;
|
|
|
|
using GFLAGS_NAMESPACE::SetUsageMessage;
|
|
|
|
|
2019-10-09 04:19:43 +02:00
|
|
|
DEFINE_uint32(seed, 0, "Seed for random number generators");
|
2019-10-08 05:09:27 +02:00
|
|
|
|
|
|
|
DEFINE_double(working_mem_size_mb, 200,
|
2020-02-21 22:29:41 +01:00
|
|
|
"MB of memory to get up to among all filters, unless "
|
|
|
|
"m_keys_total_max is specified.");
|
2019-10-08 05:09:27 +02:00
|
|
|
|
|
|
|
DEFINE_uint32(average_keys_per_filter, 10000,
|
|
|
|
"Average number of keys per filter");
|
|
|
|
|
2020-02-21 22:29:41 +01:00
|
|
|
DEFINE_double(vary_key_count_ratio, 0.4,
|
|
|
|
"Vary number of keys by up to +/- vary_key_count_ratio * "
|
|
|
|
"average_keys_per_filter.");
|
|
|
|
|
2019-10-24 22:07:09 +02:00
|
|
|
DEFINE_uint32(key_size, 24, "Average number of bytes for each key");
|
|
|
|
|
|
|
|
DEFINE_bool(vary_key_alignment, true,
|
|
|
|
"Whether to vary key alignment (default: at least 32-bit "
|
|
|
|
"alignment)");
|
|
|
|
|
|
|
|
DEFINE_uint32(vary_key_size_log2_interval, 5,
|
|
|
|
"Use same key size 2^n times, then change. Key size varies from "
|
|
|
|
"-2 to +2 bytes vs. average, unless n>=30 to fix key size.");
|
2019-10-08 05:09:27 +02:00
|
|
|
|
|
|
|
DEFINE_uint32(batch_size, 8, "Number of keys to group in each batch");
|
|
|
|
|
Allow fractional bits/key in BloomFilterPolicy (#6092)
Summary:
There's no technological impediment to allowing the Bloom
filter bits/key to be non-integer (fractional/decimal) values, and it
provides finer control over the memory vs. accuracy trade-off. This is
especially handy in using the format_version=5 Bloom filter in place
of the old one, because bits_per_key=9.55 provides the same accuracy as
the old bits_per_key=10.
This change not only requires refining the logic for choosing the best
num_probes for a given bits/key setting, it revealed a flaw in that logic.
As bits/key gets higher, the best num_probes for a cache-local Bloom
filter is closer to bpk / 2 than to bpk * 0.69, the best choice for a
standard Bloom filter. For example, at 16 bits per key, the best
num_probes is 9 (FP rate = 0.0843%) not 11 (FP rate = 0.0884%).
This change fixes and refines that logic (for the format_version=5
Bloom filter only, just in case) based on empirical tests to find
accuracy inflection points between each num_probes.
Although bits_per_key is now specified as a double, the new Bloom
filter converts/rounds this to "millibits / key" for predictable/precise
internal computations. Just in case of unforeseen compatibility
issues, we round to the nearest whole number bits / key for the
legacy Bloom filter, so as not to unlock new behaviors for it.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6092
Test Plan: unit tests included
Differential Revision: D18711313
Pulled By: pdillinger
fbshipit-source-id: 1aa73295f152a995328cb846ef9157ae8a05522a
2019-11-27 00:49:16 +01:00
|
|
|
DEFINE_double(bits_per_key, 10.0, "Bits per key setting for filters");
|
2019-10-08 05:09:27 +02:00
|
|
|
|
|
|
|
DEFINE_double(m_queries, 200, "Millions of queries for each test mode");
|
|
|
|
|
2020-02-21 22:29:41 +01:00
|
|
|
DEFINE_double(m_keys_total_max, 0,
|
|
|
|
"Maximum total keys added to filters, in millions. "
|
|
|
|
"0 (default) disables. Non-zero overrides working_mem_size_mb "
|
|
|
|
"option.");
|
|
|
|
|
2019-10-08 05:09:27 +02:00
|
|
|
DEFINE_bool(use_full_block_reader, false,
|
|
|
|
"Use FullFilterBlockReader interface rather than FilterBitsReader");
|
|
|
|
|
2019-10-25 22:25:28 +02:00
|
|
|
DEFINE_bool(use_plain_table_bloom, false,
|
|
|
|
"Use PlainTableBloom structure and interface rather than "
|
|
|
|
"FilterBitsReader/FullFilterBlockReader");
|
|
|
|
|
2019-12-16 06:29:43 +01:00
|
|
|
DEFINE_bool(new_builder, false,
|
|
|
|
"Whether to create a new builder for each new filter");
|
|
|
|
|
2019-10-25 22:25:28 +02:00
|
|
|
DEFINE_uint32(impl, 0,
|
|
|
|
"Select filter implementation. Without -use_plain_table_bloom:"
|
Experimental (production candidate) SST schema for Ribbon filter (#7658)
Summary:
Added experimental public API for Ribbon filter:
NewExperimentalRibbonFilterPolicy(). This experimental API will
take a "Bloom equivalent" bits per key, and configure the Ribbon
filter for the same FP rate as Bloom would have but ~30% space
savings. (Note: optimize_filters_for_memory is not yet implemented
for Ribbon filter. That can be added with no effect on schema.)
Internally, the Ribbon filter is configured using a "one_in_fp_rate"
value, which is 1 over desired FP rate. For example, use 100 for 1%
FP rate. I'm expecting this will be used in the future for configuring
Bloom-like filters, as I expect people to more commonly hold constant
the filter accuracy and change the space vs. time trade-off, rather than
hold constant the space (per key) and change the accuracy vs. time
trade-off, though we might make that available.
### Benchmarking
```
$ ./filter_bench -impl=2 -quick -m_keys_total_max=200 -average_keys_per_filter=100000 -net_includes_hashing
Building...
Build avg ns/key: 34.1341
Number of filters: 1993
Total size (MB): 238.488
Reported total allocated memory (MB): 262.875
Reported internal fragmentation: 10.2255%
Bits/key stored: 10.0029
----------------------------
Mixed inside/outside queries...
Single filter net ns/op: 18.7508
Random filter net ns/op: 258.246
Average FP rate %: 0.968672
----------------------------
Done. (For more info, run with -legend or -help.)
$ ./filter_bench -impl=3 -quick -m_keys_total_max=200 -average_keys_per_filter=100000 -net_includes_hashing
Building...
Build avg ns/key: 130.851
Number of filters: 1993
Total size (MB): 168.166
Reported total allocated memory (MB): 183.211
Reported internal fragmentation: 8.94626%
Bits/key stored: 7.05341
----------------------------
Mixed inside/outside queries...
Single filter net ns/op: 58.4523
Random filter net ns/op: 363.717
Average FP rate %: 0.952978
----------------------------
Done. (For more info, run with -legend or -help.)
```
168.166 / 238.488 = 0.705 -> 29.5% space reduction
130.851 / 34.1341 = 3.83x construction time for this Ribbon filter vs. lastest Bloom filter (could make that as little as about 2.5x for less space reduction)
### Working around a hashing "flaw"
bloom_test discovered a flaw in the simple hashing applied in
StandardHasher when num_starts == 1 (num_slots == 128), showing an
excessively high FP rate. The problem is that when many entries, on the
order of number of hash bits or kCoeffBits, are associated with the same
start location, the correlation between the CoeffRow and ResultRow (for
efficiency) can lead to a solution that is "universal," or nearly so, for
entries mapping to that start location. (Normally, variance in start
location breaks the effective association between CoeffRow and
ResultRow; the same value for CoeffRow is effectively different if start
locations are different.) Without kUseSmash and with num_starts > 1 (thus
num_starts ~= num_slots), this flaw should be completely irrelevant. Even
with 10M slots, the chances of a single slot having just 16 (or more)
entries map to it--not enough to cause an FP problem, which would be local
to that slot if it happened--is 1 in millions. This spreadsheet formula
shows that: =1/(10000000*(1 - POISSON(15, 1, TRUE)))
As kUseSmash==false (the setting for Standard128RibbonBitsBuilder) is
intended for CPU efficiency of filters with many more entries/slots than
kCoeffBits, a very reasonable work-around is to disallow num_starts==1
when !kUseSmash, by making the minimum non-zero number of slots
2*kCoeffBits. This is the work-around I've applied. This also means that
the new Ribbon filter schema (Standard128RibbonBitsBuilder) is not
space-efficient for less than a few hundred entries. Because of this, I
have made it fall back on constructing a Bloom filter, under existing
schema, when that is more space efficient for small filters. (We can
change this in the future if we want.)
TODO: better unit tests for this case in ribbon_test, and probably
update StandardHasher for kUseSmash case so that it can scale nicely to
small filters.
### Other related changes
* Add Ribbon filter to stress/crash test
* Add Ribbon filter to filter_bench as -impl=3
* Add option string support, as in "filter_policy=experimental_ribbon:5.678;"
where 5.678 is the Bloom equivalent bits per key.
* Rename internal mode BloomFilterPolicy::kAuto to kAutoBloom
* Add a general BuiltinFilterBitsBuilder::CalculateNumEntry based on
binary searching CalculateSpace (inefficient), so that subclasses
(especially experimental ones) don't have to provide an efficient
implementation inverting CalculateSpace.
* Minor refactor FastLocalBloomBitsBuilder for new base class
XXH3pFilterBitsBuilder shared with new Standard128RibbonBitsBuilder,
which allows the latter to fall back on Bloom construction in some
extreme cases.
* Mostly updated bloom_test for Ribbon filter, though a test like
FullBloomTest::Schema is a next TODO to ensure schema stability
(in case this becomes production-ready schema as it is).
* Add some APIs to ribbon_impl.h for configuring Ribbon filters.
Although these are reasonably covered by bloom_test, TODO more unit
tests in ribbon_test
* Added a "tool" FindOccupancyForSuccessRate to ribbon_test to get data
for constructing the linear approximations in GetNumSlotsFor95PctSuccess.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/7658
Test Plan:
Some unit tests updated but other testing is left TODO. This
is considered experimental but laying down schema compatibility as early
as possible in case it proves production-quality. Also tested in
stress/crash test.
Reviewed By: jay-zhuang
Differential Revision: D24899349
Pulled By: pdillinger
fbshipit-source-id: 9715f3e6371c959d923aea8077c9423c7a9f82b8
2020-11-13 05:45:02 +01:00
|
|
|
"0 = legacy full Bloom filter, 1 = block-based Bloom filter, "
|
|
|
|
"2 = format_version 5 Bloom filter, 3 = Ribbon128 filter. With "
|
2019-10-25 22:25:28 +02:00
|
|
|
"-use_plain_table_bloom: 0 = no locality, 1 = locality.");
|
|
|
|
|
|
|
|
DEFINE_bool(net_includes_hashing, false,
|
|
|
|
"Whether query net ns/op times should include hashing. "
|
|
|
|
"(if not, dry run will include hashing) "
|
|
|
|
"(build times always include hashing)");
|
|
|
|
|
Minimize memory internal fragmentation for Bloom filters (#6427)
Summary:
New experimental option BBTO::optimize_filters_for_memory builds
filters that maximize their use of "usable size" from malloc_usable_size,
which is also used to compute block cache charges.
Rather than always "rounding up," we track state in the
BloomFilterPolicy object to mix essentially "rounding down" and
"rounding up" so that the average FP rate of all generated filters is
the same as without the option. (YMMV as heavily accessed filters might
be unluckily lower accuracy.)
Thus, the option near-minimizes what the block cache considers as
"memory used" for a given target Bloom filter false positive rate and
Bloom filter implementation. There are no forward or backward
compatibility issues with this change, though it only works on the
format_version=5 Bloom filter.
With Jemalloc, we see about 10% reduction in memory footprint (and block
cache charge) for Bloom filters, but 1-2% increase in storage footprint,
due to encoding efficiency losses (FP rate is non-linear with bits/key).
Why not weighted random round up/down rather than state tracking? By
only requiring malloc_usable_size, we don't actually know what the next
larger and next smaller usable sizes for the allocator are. We pick a
requested size, accept and use whatever usable size it has, and use the
difference to inform our next choice. This allows us to narrow in on the
right balance without tracking/predicting usable sizes.
Why not weight history of generated filter false positive rates by
number of keys? This could lead to excess skew in small filters after
generating a large filter.
Results from filter_bench with jemalloc (irrelevant details omitted):
(normal keys/filter, but high variance)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.6278
Number of filters: 5516
Total size (MB): 200.046
Reported total allocated memory (MB): 220.597
Reported internal fragmentation: 10.2732%
Bits/key stored: 10.0097
Average FP rate %: 0.965228
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.5104
Number of filters: 5464
Total size (MB): 200.015
Reported total allocated memory (MB): 200.322
Reported internal fragmentation: 0.153709%
Bits/key stored: 10.1011
Average FP rate %: 0.966313
(very few keys / filter, optimization not as effective due to ~59 byte
internal fragmentation in blocked Bloom filter representation)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.5649
Number of filters: 162950
Total size (MB): 200.001
Reported total allocated memory (MB): 224.624
Reported internal fragmentation: 12.3117%
Bits/key stored: 10.2951
Average FP rate %: 0.821534
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 31.8057
Number of filters: 159849
Total size (MB): 200
Reported total allocated memory (MB): 208.846
Reported internal fragmentation: 4.42297%
Bits/key stored: 10.4948
Average FP rate %: 0.811006
(high keys/filter)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.7017
Number of filters: 164
Total size (MB): 200.352
Reported total allocated memory (MB): 221.5
Reported internal fragmentation: 10.5552%
Bits/key stored: 10.0003
Average FP rate %: 0.969358
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.7131
Number of filters: 160
Total size (MB): 200.928
Reported total allocated memory (MB): 200.938
Reported internal fragmentation: 0.00448054%
Bits/key stored: 10.1852
Average FP rate %: 0.963387
And from db_bench (block cache) with jemalloc:
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ ./db_bench -db=/dev/shm/dbbench -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -optimize_filters_for_memory -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ (for FILE in /dev/shm/dbbench.no_optimize/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17063835
$ (for FILE in /dev/shm/dbbench/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17430747
$ #^ 2.1% additional filter storage
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8440400
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 21087528
rocksdb.bloom.filter.useful COUNT : 4963889
rocksdb.bloom.filter.full.positive COUNT : 1214081
rocksdb.bloom.filter.full.true.positive COUNT : 1161999
$ #^ 1.04 % observed FP rate
$ ./db_bench -db=/dev/shm/dbbench -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -optimize_filters_for_memory -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8448592
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 18220328
rocksdb.bloom.filter.useful COUNT : 5360933
rocksdb.bloom.filter.full.positive COUNT : 1321315
rocksdb.bloom.filter.full.true.positive COUNT : 1262999
$ #^ 1.08 % observed FP rate, 13.6% less memory usage for filters
(Due to specific key density, this example tends to generate filters that are "worse than average" for internal fragmentation. "Better than average" cases can show little or no improvement.)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6427
Test Plan: unit test added, 'make check' with gcc, clang and valgrind
Reviewed By: siying
Differential Revision: D22124374
Pulled By: pdillinger
fbshipit-source-id: f3e3aa152f9043ddf4fae25799e76341d0d8714e
2020-06-22 22:30:57 +02:00
|
|
|
DEFINE_bool(optimize_filters_for_memory, false,
|
|
|
|
"Setting for BlockBasedTableOptions::optimize_filters_for_memory");
|
|
|
|
|
2019-10-08 05:09:27 +02:00
|
|
|
DEFINE_bool(quick, false, "Run more limited set of tests, fewer queries");
|
|
|
|
|
2019-10-25 22:25:28 +02:00
|
|
|
DEFINE_bool(best_case, false, "Run limited tests only for best-case");
|
|
|
|
|
2019-10-08 05:09:27 +02:00
|
|
|
DEFINE_bool(allow_bad_fp_rate, false, "Continue even if FP rate is bad");
|
|
|
|
|
|
|
|
DEFINE_bool(legend, false,
|
|
|
|
"Print more information about interpreting results instead of "
|
|
|
|
"running tests");
|
|
|
|
|
2020-02-21 22:29:41 +01:00
|
|
|
DEFINE_uint32(runs, 1, "Number of times to rebuild and run benchmark tests");
|
|
|
|
|
2019-10-08 05:09:27 +02:00
|
|
|
void _always_assert_fail(int line, const char *file, const char *expr) {
|
|
|
|
fprintf(stderr, "%s: %d: Assertion %s failed\n", file, line, expr);
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
|
|
|
|
#define ALWAYS_ASSERT(cond) \
|
|
|
|
((cond) ? (void)0 : ::_always_assert_fail(__LINE__, __FILE__, #cond))
|
|
|
|
|
Warn on excessive keys for legacy Bloom filter with 32-bit hash (#6317)
Summary:
With many millions of keys, the old Bloom filter implementation
for the block-based table (format_version <= 4) would have excessive FP
rate due to the limitations of feeding the Bloom filter with a 32-bit hash.
This change computes an estimated inflated FP rate due to this effect
and warns in the log whenever an SST filter is constructed (almost
certainly a "full" not "partitioned" filter) that exceeds 1.5x FP rate
due to this effect. The detailed condition is only checked if 3 million
keys or more have been added to a filter, as this should be a lower
bound for common bits/key settings (< 20).
Recommended remedies include smaller SST file size, using
format_version >= 5 (for new Bloom filter), or using partitioned
filters.
This does not change behavior other than generating warnings for some
constructed filters using the old implementation.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6317
Test Plan:
Example with warning, 15M keys @ 15 bits / key: (working_mem_size_mb is just to stop after building one filter if it's large)
$ ./filter_bench -quick -impl=0 -working_mem_size_mb=1 -bits_per_key=15 -average_keys_per_filter=15000000 2>&1 | grep 'FP rate'
[WARN] [/block_based/filter_policy.cc:292] Using legacy SST/BBT Bloom filter with excessive key count (15.0M @ 15bpk), causing estimated 1.8x higher filter FP rate. Consider using new Bloom with format_version>=5, smaller SST file size, or partitioned filters.
Predicted FP rate %: 0.766702
Average FP rate %: 0.66846
Example without warning (150K keys):
$ ./filter_bench -quick -impl=0 -working_mem_size_mb=1 -bits_per_key=15 -average_keys_per_filter=150000 2>&1 | grep 'FP rate'
Predicted FP rate %: 0.422857
Average FP rate %: 0.379301
$
With more samples at 15 bits/key:
150K keys -> no warning; actual: 0.379% FP rate (baseline)
1M keys -> no warning; actual: 0.396% FP rate, 1.045x
9M keys -> no warning; actual: 0.563% FP rate, 1.485x
10M keys -> warning (1.5x); actual: 0.564% FP rate, 1.488x
15M keys -> warning (1.8x); actual: 0.668% FP rate, 1.76x
25M keys -> warning (2.4x); actual: 0.880% FP rate, 2.32x
At 10 bits/key:
150K keys -> no warning; actual: 1.17% FP rate (baseline)
1M keys -> no warning; actual: 1.16% FP rate
10M keys -> no warning; actual: 1.32% FP rate, 1.13x
25M keys -> no warning; actual: 1.63% FP rate, 1.39x
35M keys -> warning (1.6x); actual: 1.81% FP rate, 1.55x
At 5 bits/key:
150K keys -> no warning; actual: 9.32% FP rate (baseline)
25M keys -> no warning; actual: 9.62% FP rate, 1.03x
200M keys -> no warning; actual: 12.2% FP rate, 1.31x
250M keys -> warning (1.5x); actual: 12.8% FP rate, 1.37x
300M keys -> warning (1.6x); actual: 13.4% FP rate, 1.43x
The reason for the modest inaccuracy at low bits/key is that the assumption of independence between a collision between 32-hash values feeding the filter and an FP in the filter is not quite true for implementations using "simple" logic to compute indices from the stock hash result. There's math on this in my dissertation, but I don't think it's worth the effort just for these extreme cases (> 100 million keys and low-ish bits/key).
Differential Revision: D19471715
Pulled By: pdillinger
fbshipit-source-id: f80c96893a09bf1152630ff0b964e5cdd7e35c68
2020-01-21 06:30:22 +01:00
|
|
|
#ifndef NDEBUG
|
|
|
|
// This could affect build times enough that we should not include it for
|
|
|
|
// accurate speed tests
|
|
|
|
#define PREDICT_FP_RATE
|
|
|
|
#endif
|
|
|
|
|
2020-02-20 21:07:53 +01:00
|
|
|
using ROCKSDB_NAMESPACE::Arena;
|
|
|
|
using ROCKSDB_NAMESPACE::BlockContents;
|
|
|
|
using ROCKSDB_NAMESPACE::BloomFilterPolicy;
|
|
|
|
using ROCKSDB_NAMESPACE::BloomHash;
|
|
|
|
using ROCKSDB_NAMESPACE::BuiltinFilterBitsBuilder;
|
|
|
|
using ROCKSDB_NAMESPACE::CachableEntry;
|
|
|
|
using ROCKSDB_NAMESPACE::EncodeFixed32;
|
2020-09-28 20:33:31 +02:00
|
|
|
using ROCKSDB_NAMESPACE::FastRange32;
|
2020-02-20 21:07:53 +01:00
|
|
|
using ROCKSDB_NAMESPACE::FilterBitsReader;
|
|
|
|
using ROCKSDB_NAMESPACE::FilterBuildingContext;
|
|
|
|
using ROCKSDB_NAMESPACE::FullFilterBlockReader;
|
|
|
|
using ROCKSDB_NAMESPACE::GetSliceHash;
|
|
|
|
using ROCKSDB_NAMESPACE::GetSliceHash64;
|
|
|
|
using ROCKSDB_NAMESPACE::Lower32of64;
|
|
|
|
using ROCKSDB_NAMESPACE::ParsedFullFilterBlock;
|
|
|
|
using ROCKSDB_NAMESPACE::PlainTableBloomV1;
|
|
|
|
using ROCKSDB_NAMESPACE::Random32;
|
|
|
|
using ROCKSDB_NAMESPACE::Slice;
|
2020-04-29 22:06:27 +02:00
|
|
|
using ROCKSDB_NAMESPACE::static_cast_with_check;
|
2020-02-20 21:07:53 +01:00
|
|
|
using ROCKSDB_NAMESPACE::StderrLogger;
|
|
|
|
using ROCKSDB_NAMESPACE::mock::MockBlockBasedTableTester;
|
2019-10-08 05:09:27 +02:00
|
|
|
|
|
|
|
struct KeyMaker {
|
2019-10-24 22:07:09 +02:00
|
|
|
KeyMaker(size_t avg_size)
|
|
|
|
: smallest_size_(avg_size -
|
|
|
|
(FLAGS_vary_key_size_log2_interval >= 30 ? 2 : 0)),
|
|
|
|
buf_size_(avg_size + 11), // pad to vary key size and alignment
|
|
|
|
buf_(new char[buf_size_]) {
|
|
|
|
memset(buf_.get(), 0, buf_size_);
|
|
|
|
assert(smallest_size_ > 8);
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
2019-10-24 22:07:09 +02:00
|
|
|
size_t smallest_size_;
|
|
|
|
size_t buf_size_;
|
|
|
|
std::unique_ptr<char[]> buf_;
|
2019-10-08 05:09:27 +02:00
|
|
|
|
2019-10-24 22:07:09 +02:00
|
|
|
// Returns a unique(-ish) key based on the given parameter values. Each
|
|
|
|
// call returns a Slice from the same buffer so previously returned
|
|
|
|
// Slices should be considered invalidated.
|
2019-10-08 05:09:27 +02:00
|
|
|
Slice Get(uint32_t filter_num, uint32_t val_num) {
|
2019-10-24 22:07:09 +02:00
|
|
|
size_t start = FLAGS_vary_key_alignment ? val_num % 4 : 0;
|
|
|
|
size_t len = smallest_size_;
|
|
|
|
if (FLAGS_vary_key_size_log2_interval < 30) {
|
|
|
|
// To get range [avg_size - 2, avg_size + 2]
|
|
|
|
// use range [smallest_size, smallest_size + 4]
|
2020-09-28 20:33:31 +02:00
|
|
|
len += FastRange32(
|
2019-10-24 22:07:09 +02:00
|
|
|
(val_num >> FLAGS_vary_key_size_log2_interval) * 1234567891, 5);
|
|
|
|
}
|
|
|
|
char * data = buf_.get() + start;
|
|
|
|
// Populate key data such that all data makes it into a key of at
|
|
|
|
// least 8 bytes. We also don't want all the within-filter key
|
|
|
|
// variance confined to a contiguous 32 bits, because then a 32 bit
|
|
|
|
// hash function can "cheat" the false positive rate by
|
|
|
|
// approximating a perfect hash.
|
|
|
|
EncodeFixed32(data, val_num);
|
|
|
|
EncodeFixed32(data + 4, filter_num + val_num);
|
|
|
|
// ensure clearing leftovers from different alignment
|
|
|
|
EncodeFixed32(data + 8, 0);
|
|
|
|
return Slice(data, len);
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
};
|
|
|
|
|
|
|
|
void PrintWarnings() {
|
|
|
|
#if defined(__GNUC__) && !defined(__OPTIMIZE__)
|
|
|
|
fprintf(stdout,
|
|
|
|
"WARNING: Optimization is disabled: benchmarks unnecessarily slow\n");
|
|
|
|
#endif
|
|
|
|
#ifndef NDEBUG
|
|
|
|
fprintf(stdout,
|
|
|
|
"WARNING: Assertions are enabled; benchmarks unnecessarily slow\n");
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
|
|
|
struct FilterInfo {
|
|
|
|
uint32_t filter_id_ = 0;
|
|
|
|
std::unique_ptr<const char[]> owner_;
|
|
|
|
Slice filter_;
|
|
|
|
uint32_t keys_added_ = 0;
|
|
|
|
std::unique_ptr<FilterBitsReader> reader_;
|
|
|
|
std::unique_ptr<FullFilterBlockReader> full_block_reader_;
|
2019-10-25 22:25:28 +02:00
|
|
|
std::unique_ptr<PlainTableBloomV1> plain_table_bloom_;
|
2019-10-08 05:09:27 +02:00
|
|
|
uint64_t outside_queries_ = 0;
|
|
|
|
uint64_t false_positives_ = 0;
|
|
|
|
};
|
|
|
|
|
|
|
|
enum TestMode {
|
|
|
|
kSingleFilter,
|
|
|
|
kBatchPrepared,
|
|
|
|
kBatchUnprepared,
|
|
|
|
kFiftyOneFilter,
|
|
|
|
kEightyTwentyFilter,
|
|
|
|
kRandomFilter,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const std::vector<TestMode> allTestModes = {
|
|
|
|
kSingleFilter, kBatchPrepared, kBatchUnprepared,
|
|
|
|
kFiftyOneFilter, kEightyTwentyFilter, kRandomFilter,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const std::vector<TestMode> quickTestModes = {
|
|
|
|
kSingleFilter,
|
|
|
|
kRandomFilter,
|
|
|
|
};
|
|
|
|
|
2019-10-25 22:25:28 +02:00
|
|
|
static const std::vector<TestMode> bestCaseTestModes = {
|
|
|
|
kSingleFilter,
|
|
|
|
};
|
|
|
|
|
2019-10-08 05:09:27 +02:00
|
|
|
const char *TestModeToString(TestMode tm) {
|
|
|
|
switch (tm) {
|
|
|
|
case kSingleFilter:
|
|
|
|
return "Single filter";
|
|
|
|
case kBatchPrepared:
|
|
|
|
return "Batched, prepared";
|
|
|
|
case kBatchUnprepared:
|
|
|
|
return "Batched, unprepared";
|
|
|
|
case kFiftyOneFilter:
|
|
|
|
return "Skewed 50% in 1%";
|
|
|
|
case kEightyTwentyFilter:
|
|
|
|
return "Skewed 80% in 20%";
|
|
|
|
case kRandomFilter:
|
|
|
|
return "Random filter";
|
|
|
|
}
|
|
|
|
return "Bad TestMode";
|
|
|
|
}
|
|
|
|
|
2019-10-25 22:25:28 +02:00
|
|
|
// Do just enough to keep some data dependence for the
|
|
|
|
// compiler / CPU
|
Add new persistent 64-bit hash (#5984)
Summary:
For upcoming new SST filter implementations, we will use a new
64-bit hash function (XXH3 preview, slightly modified). This change
updates hash.{h,cc} for that change, adds unit tests, and out-of-lines
the implementations to keep hash.h as clean/small as possible.
In developing the unit tests, I discovered that the XXH3 preview always
returns zero for the empty string. Zero is problematic for some
algorithms (including an upcoming SST filter implementation) if it
occurs more often than at the "natural" rate, so it should not be
returned from trivial values using trivial seeds. I modified our fork
of XXH3 to return a modest hash of the seed for the empty string.
With hash function details out-of-lines in hash.h, it makes sense to
enable XXH_INLINE_ALL, so that direct calls to XXH64/XXH32/XXH3p
are inlined. To fix array-bounds warnings on some inline calls, I
injected some casts to uintptr_t in xxhash.cc. (Issue reported to Yann.)
Revised: Reverted using XXH_INLINE_ALL for now. Some Facebook
checks are unhappy about #include on xxhash.cc file. I would
fix that by rename to xxhash_cc.h, but to best preserve history I want
to do that in a separate commit (PR) from the uintptr casts.
Also updated filter_bench for this change, improving the performance
predictability of dry run hashing and adding support for 64-bit hash
(for upcoming new SST filter implementations, minor dead code in the
tool for now).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5984
Differential Revision: D18246567
Pulled By: pdillinger
fbshipit-source-id: 6162fbf6381d63c8cc611dd7ec70e1ddc883fbb8
2019-11-01 00:34:51 +01:00
|
|
|
static uint32_t DryRunNoHash(Slice &s) {
|
2019-10-25 22:25:28 +02:00
|
|
|
uint32_t sz = static_cast<uint32_t>(s.size());
|
|
|
|
if (sz >= 4) {
|
|
|
|
return sz + s.data()[3];
|
|
|
|
} else {
|
|
|
|
return sz;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Add new persistent 64-bit hash (#5984)
Summary:
For upcoming new SST filter implementations, we will use a new
64-bit hash function (XXH3 preview, slightly modified). This change
updates hash.{h,cc} for that change, adds unit tests, and out-of-lines
the implementations to keep hash.h as clean/small as possible.
In developing the unit tests, I discovered that the XXH3 preview always
returns zero for the empty string. Zero is problematic for some
algorithms (including an upcoming SST filter implementation) if it
occurs more often than at the "natural" rate, so it should not be
returned from trivial values using trivial seeds. I modified our fork
of XXH3 to return a modest hash of the seed for the empty string.
With hash function details out-of-lines in hash.h, it makes sense to
enable XXH_INLINE_ALL, so that direct calls to XXH64/XXH32/XXH3p
are inlined. To fix array-bounds warnings on some inline calls, I
injected some casts to uintptr_t in xxhash.cc. (Issue reported to Yann.)
Revised: Reverted using XXH_INLINE_ALL for now. Some Facebook
checks are unhappy about #include on xxhash.cc file. I would
fix that by rename to xxhash_cc.h, but to best preserve history I want
to do that in a separate commit (PR) from the uintptr casts.
Also updated filter_bench for this change, improving the performance
predictability of dry run hashing and adding support for 64-bit hash
(for upcoming new SST filter implementations, minor dead code in the
tool for now).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5984
Differential Revision: D18246567
Pulled By: pdillinger
fbshipit-source-id: 6162fbf6381d63c8cc611dd7ec70e1ddc883fbb8
2019-11-01 00:34:51 +01:00
|
|
|
static uint32_t DryRunHash32(Slice &s) {
|
|
|
|
// Same perf characteristics as GetSliceHash()
|
|
|
|
return BloomHash(s);
|
|
|
|
}
|
|
|
|
|
|
|
|
static uint32_t DryRunHash64(Slice &s) {
|
|
|
|
return Lower32of64(GetSliceHash64(s));
|
|
|
|
}
|
|
|
|
|
2019-10-08 05:09:27 +02:00
|
|
|
struct FilterBench : public MockBlockBasedTableTester {
|
|
|
|
std::vector<KeyMaker> kms_;
|
|
|
|
std::vector<FilterInfo> infos_;
|
2019-10-09 04:19:43 +02:00
|
|
|
Random32 random_;
|
2019-10-24 22:07:09 +02:00
|
|
|
std::ostringstream fp_rate_report_;
|
2019-10-25 22:25:28 +02:00
|
|
|
Arena arena_;
|
2020-01-18 04:36:09 +01:00
|
|
|
StderrLogger stderr_logger_;
|
2020-02-21 22:29:41 +01:00
|
|
|
double m_queries_;
|
2019-10-08 05:09:27 +02:00
|
|
|
|
|
|
|
FilterBench()
|
New Bloom filter implementation for full and partitioned filters (#6007)
Summary:
Adds an improved, replacement Bloom filter implementation (FastLocalBloom) for full and partitioned filters in the block-based table. This replacement is faster and more accurate, especially for high bits per key or millions of keys in a single filter.
Speed
The improved speed, at least on recent x86_64, comes from
* Using fastrange instead of modulo (%)
* Using our new hash function (XXH3 preview, added in a previous commit), which is much faster for large keys and only *slightly* slower on keys around 12 bytes if hashing the same size many thousands of times in a row.
* Optimizing the Bloom filter queries with AVX2 SIMD operations. (Added AVX2 to the USE_SSE=1 build.) Careful design was required to support (a) SIMD-optimized queries, (b) compatible non-SIMD code that's simple and efficient, (c) flexible choice of number of probes, and (d) essentially maximized accuracy for a cache-local Bloom filter. Probes are made eight at a time, so any number of probes up to 8 is the same speed, then up to 16, etc.
* Prefetching cache lines when building the filter. Although this optimization could be applied to the old structure as well, it seems to balance out the small added cost of accumulating 64 bit hashes for adding to the filter rather than 32 bit hashes.
Here's nominal speed data from filter_bench (200MB in filters, about 10k keys each, 10 bits filter data / key, 6 probes, avg key size 24 bytes, includes hashing time) on Skylake DE (relatively low clock speed):
$ ./filter_bench -quick -impl=2 -net_includes_hashing # New Bloom filter
Build avg ns/key: 47.7135
Mixed inside/outside queries...
Single filter net ns/op: 26.2825
Random filter net ns/op: 150.459
Average FP rate %: 0.954651
$ ./filter_bench -quick -impl=0 -net_includes_hashing # Old Bloom filter
Build avg ns/key: 47.2245
Mixed inside/outside queries...
Single filter net ns/op: 63.2978
Random filter net ns/op: 188.038
Average FP rate %: 1.13823
Similar build time but dramatically faster query times on hot data (63 ns to 26 ns), and somewhat faster on stale data (188 ns to 150 ns). Performance differences on batched and skewed query loads are between these extremes as expected.
The only other interesting thing about speed is "inside" (query key was added to filter) vs. "outside" (query key was not added to filter) query times. The non-SIMD implementations are substantially slower when most queries are "outside" vs. "inside". This goes against what one might expect or would have observed years ago, as "outside" queries only need about two probes on average, due to short-circuiting, while "inside" always have num_probes (say 6). The problem is probably the nastily unpredictable branch. The SIMD implementation has few branches (very predictable) and has pretty consistent running time regardless of query outcome.
Accuracy
The generally improved accuracy (re: Issue https://github.com/facebook/rocksdb/issues/5857) comes from a better design for probing indices
within a cache line (re: Issue https://github.com/facebook/rocksdb/issues/4120) and improved accuracy for millions of keys in a single filter from using a 64-bit hash function (XXH3p). Design details in code comments.
Accuracy data (generalizes, except old impl gets worse with millions of keys):
Memory bits per key: FP rate percent old impl -> FP rate percent new impl
6: 5.70953 -> 5.69888
8: 2.45766 -> 2.29709
10: 1.13977 -> 0.959254
12: 0.662498 -> 0.411593
16: 0.353023 -> 0.0873754
24: 0.261552 -> 0.0060971
50: 0.225453 -> ~0.00003 (less than 1 in a million queries are FP)
Fixes https://github.com/facebook/rocksdb/issues/5857
Fixes https://github.com/facebook/rocksdb/issues/4120
Unlike the old implementation, this implementation has a fixed cache line size (64 bytes). At 10 bits per key, the accuracy of this new implementation is very close to the old implementation with 128-byte cache line size. If there's sufficient demand, this implementation could be generalized.
Compatibility
Although old releases would see the new structure as corrupt filter data and read the table as if there's no filter, we've decided only to enable the new Bloom filter with new format_version=5. This provides a smooth path for automatic adoption over time, with an option for early opt-in.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6007
Test Plan: filter_bench has been used thoroughly to validate speed, accuracy, and correctness. Unit tests have been carefully updated to exercise new and old implementations, as well as the logic to select an implementation based on context (format_version).
Differential Revision: D18294749
Pulled By: pdillinger
fbshipit-source-id: d44c9db3696e4d0a17caaec47075b7755c262c5f
2019-11-14 01:31:26 +01:00
|
|
|
: MockBlockBasedTableTester(new BloomFilterPolicy(
|
|
|
|
FLAGS_bits_per_key,
|
|
|
|
static_cast<BloomFilterPolicy::Mode>(FLAGS_impl))),
|
2020-02-21 22:29:41 +01:00
|
|
|
random_(FLAGS_seed),
|
|
|
|
m_queries_(0) {
|
2019-10-08 05:09:27 +02:00
|
|
|
for (uint32_t i = 0; i < FLAGS_batch_size; ++i) {
|
|
|
|
kms_.emplace_back(FLAGS_key_size < 8 ? 8 : FLAGS_key_size);
|
|
|
|
}
|
2020-01-18 04:36:09 +01:00
|
|
|
ioptions_.info_log = &stderr_logger_;
|
Minimize memory internal fragmentation for Bloom filters (#6427)
Summary:
New experimental option BBTO::optimize_filters_for_memory builds
filters that maximize their use of "usable size" from malloc_usable_size,
which is also used to compute block cache charges.
Rather than always "rounding up," we track state in the
BloomFilterPolicy object to mix essentially "rounding down" and
"rounding up" so that the average FP rate of all generated filters is
the same as without the option. (YMMV as heavily accessed filters might
be unluckily lower accuracy.)
Thus, the option near-minimizes what the block cache considers as
"memory used" for a given target Bloom filter false positive rate and
Bloom filter implementation. There are no forward or backward
compatibility issues with this change, though it only works on the
format_version=5 Bloom filter.
With Jemalloc, we see about 10% reduction in memory footprint (and block
cache charge) for Bloom filters, but 1-2% increase in storage footprint,
due to encoding efficiency losses (FP rate is non-linear with bits/key).
Why not weighted random round up/down rather than state tracking? By
only requiring malloc_usable_size, we don't actually know what the next
larger and next smaller usable sizes for the allocator are. We pick a
requested size, accept and use whatever usable size it has, and use the
difference to inform our next choice. This allows us to narrow in on the
right balance without tracking/predicting usable sizes.
Why not weight history of generated filter false positive rates by
number of keys? This could lead to excess skew in small filters after
generating a large filter.
Results from filter_bench with jemalloc (irrelevant details omitted):
(normal keys/filter, but high variance)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.6278
Number of filters: 5516
Total size (MB): 200.046
Reported total allocated memory (MB): 220.597
Reported internal fragmentation: 10.2732%
Bits/key stored: 10.0097
Average FP rate %: 0.965228
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.5104
Number of filters: 5464
Total size (MB): 200.015
Reported total allocated memory (MB): 200.322
Reported internal fragmentation: 0.153709%
Bits/key stored: 10.1011
Average FP rate %: 0.966313
(very few keys / filter, optimization not as effective due to ~59 byte
internal fragmentation in blocked Bloom filter representation)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.5649
Number of filters: 162950
Total size (MB): 200.001
Reported total allocated memory (MB): 224.624
Reported internal fragmentation: 12.3117%
Bits/key stored: 10.2951
Average FP rate %: 0.821534
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 31.8057
Number of filters: 159849
Total size (MB): 200
Reported total allocated memory (MB): 208.846
Reported internal fragmentation: 4.42297%
Bits/key stored: 10.4948
Average FP rate %: 0.811006
(high keys/filter)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.7017
Number of filters: 164
Total size (MB): 200.352
Reported total allocated memory (MB): 221.5
Reported internal fragmentation: 10.5552%
Bits/key stored: 10.0003
Average FP rate %: 0.969358
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.7131
Number of filters: 160
Total size (MB): 200.928
Reported total allocated memory (MB): 200.938
Reported internal fragmentation: 0.00448054%
Bits/key stored: 10.1852
Average FP rate %: 0.963387
And from db_bench (block cache) with jemalloc:
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ ./db_bench -db=/dev/shm/dbbench -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -optimize_filters_for_memory -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ (for FILE in /dev/shm/dbbench.no_optimize/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17063835
$ (for FILE in /dev/shm/dbbench/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17430747
$ #^ 2.1% additional filter storage
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8440400
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 21087528
rocksdb.bloom.filter.useful COUNT : 4963889
rocksdb.bloom.filter.full.positive COUNT : 1214081
rocksdb.bloom.filter.full.true.positive COUNT : 1161999
$ #^ 1.04 % observed FP rate
$ ./db_bench -db=/dev/shm/dbbench -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -optimize_filters_for_memory -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8448592
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 18220328
rocksdb.bloom.filter.useful COUNT : 5360933
rocksdb.bloom.filter.full.positive COUNT : 1321315
rocksdb.bloom.filter.full.true.positive COUNT : 1262999
$ #^ 1.08 % observed FP rate, 13.6% less memory usage for filters
(Due to specific key density, this example tends to generate filters that are "worse than average" for internal fragmentation. "Better than average" cases can show little or no improvement.)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6427
Test Plan: unit test added, 'make check' with gcc, clang and valgrind
Reviewed By: siying
Differential Revision: D22124374
Pulled By: pdillinger
fbshipit-source-id: f3e3aa152f9043ddf4fae25799e76341d0d8714e
2020-06-22 22:30:57 +02:00
|
|
|
table_options_.optimize_filters_for_memory =
|
|
|
|
FLAGS_optimize_filters_for_memory;
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
void Go();
|
|
|
|
|
2019-10-25 22:25:28 +02:00
|
|
|
double RandomQueryTest(uint32_t inside_threshold, bool dry_run,
|
|
|
|
TestMode mode);
|
2019-10-08 05:09:27 +02:00
|
|
|
};
|
|
|
|
|
|
|
|
void FilterBench::Go() {
|
2019-10-25 22:25:28 +02:00
|
|
|
if (FLAGS_use_plain_table_bloom && FLAGS_use_full_block_reader) {
|
|
|
|
throw std::runtime_error(
|
|
|
|
"Can't combine -use_plain_table_bloom and -use_full_block_reader");
|
|
|
|
}
|
New Bloom filter implementation for full and partitioned filters (#6007)
Summary:
Adds an improved, replacement Bloom filter implementation (FastLocalBloom) for full and partitioned filters in the block-based table. This replacement is faster and more accurate, especially for high bits per key or millions of keys in a single filter.
Speed
The improved speed, at least on recent x86_64, comes from
* Using fastrange instead of modulo (%)
* Using our new hash function (XXH3 preview, added in a previous commit), which is much faster for large keys and only *slightly* slower on keys around 12 bytes if hashing the same size many thousands of times in a row.
* Optimizing the Bloom filter queries with AVX2 SIMD operations. (Added AVX2 to the USE_SSE=1 build.) Careful design was required to support (a) SIMD-optimized queries, (b) compatible non-SIMD code that's simple and efficient, (c) flexible choice of number of probes, and (d) essentially maximized accuracy for a cache-local Bloom filter. Probes are made eight at a time, so any number of probes up to 8 is the same speed, then up to 16, etc.
* Prefetching cache lines when building the filter. Although this optimization could be applied to the old structure as well, it seems to balance out the small added cost of accumulating 64 bit hashes for adding to the filter rather than 32 bit hashes.
Here's nominal speed data from filter_bench (200MB in filters, about 10k keys each, 10 bits filter data / key, 6 probes, avg key size 24 bytes, includes hashing time) on Skylake DE (relatively low clock speed):
$ ./filter_bench -quick -impl=2 -net_includes_hashing # New Bloom filter
Build avg ns/key: 47.7135
Mixed inside/outside queries...
Single filter net ns/op: 26.2825
Random filter net ns/op: 150.459
Average FP rate %: 0.954651
$ ./filter_bench -quick -impl=0 -net_includes_hashing # Old Bloom filter
Build avg ns/key: 47.2245
Mixed inside/outside queries...
Single filter net ns/op: 63.2978
Random filter net ns/op: 188.038
Average FP rate %: 1.13823
Similar build time but dramatically faster query times on hot data (63 ns to 26 ns), and somewhat faster on stale data (188 ns to 150 ns). Performance differences on batched and skewed query loads are between these extremes as expected.
The only other interesting thing about speed is "inside" (query key was added to filter) vs. "outside" (query key was not added to filter) query times. The non-SIMD implementations are substantially slower when most queries are "outside" vs. "inside". This goes against what one might expect or would have observed years ago, as "outside" queries only need about two probes on average, due to short-circuiting, while "inside" always have num_probes (say 6). The problem is probably the nastily unpredictable branch. The SIMD implementation has few branches (very predictable) and has pretty consistent running time regardless of query outcome.
Accuracy
The generally improved accuracy (re: Issue https://github.com/facebook/rocksdb/issues/5857) comes from a better design for probing indices
within a cache line (re: Issue https://github.com/facebook/rocksdb/issues/4120) and improved accuracy for millions of keys in a single filter from using a 64-bit hash function (XXH3p). Design details in code comments.
Accuracy data (generalizes, except old impl gets worse with millions of keys):
Memory bits per key: FP rate percent old impl -> FP rate percent new impl
6: 5.70953 -> 5.69888
8: 2.45766 -> 2.29709
10: 1.13977 -> 0.959254
12: 0.662498 -> 0.411593
16: 0.353023 -> 0.0873754
24: 0.261552 -> 0.0060971
50: 0.225453 -> ~0.00003 (less than 1 in a million queries are FP)
Fixes https://github.com/facebook/rocksdb/issues/5857
Fixes https://github.com/facebook/rocksdb/issues/4120
Unlike the old implementation, this implementation has a fixed cache line size (64 bytes). At 10 bits per key, the accuracy of this new implementation is very close to the old implementation with 128-byte cache line size. If there's sufficient demand, this implementation could be generalized.
Compatibility
Although old releases would see the new structure as corrupt filter data and read the table as if there's no filter, we've decided only to enable the new Bloom filter with new format_version=5. This provides a smooth path for automatic adoption over time, with an option for early opt-in.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6007
Test Plan: filter_bench has been used thoroughly to validate speed, accuracy, and correctness. Unit tests have been carefully updated to exercise new and old implementations, as well as the logic to select an implementation based on context (format_version).
Differential Revision: D18294749
Pulled By: pdillinger
fbshipit-source-id: d44c9db3696e4d0a17caaec47075b7755c262c5f
2019-11-14 01:31:26 +01:00
|
|
|
if (FLAGS_use_plain_table_bloom) {
|
|
|
|
if (FLAGS_impl > 1) {
|
|
|
|
throw std::runtime_error(
|
|
|
|
"-impl must currently be >= 0 and <= 1 for Plain table");
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if (FLAGS_impl == 1) {
|
|
|
|
throw std::runtime_error(
|
|
|
|
"Block-based filter not currently supported by filter_bench");
|
|
|
|
}
|
Experimental (production candidate) SST schema for Ribbon filter (#7658)
Summary:
Added experimental public API for Ribbon filter:
NewExperimentalRibbonFilterPolicy(). This experimental API will
take a "Bloom equivalent" bits per key, and configure the Ribbon
filter for the same FP rate as Bloom would have but ~30% space
savings. (Note: optimize_filters_for_memory is not yet implemented
for Ribbon filter. That can be added with no effect on schema.)
Internally, the Ribbon filter is configured using a "one_in_fp_rate"
value, which is 1 over desired FP rate. For example, use 100 for 1%
FP rate. I'm expecting this will be used in the future for configuring
Bloom-like filters, as I expect people to more commonly hold constant
the filter accuracy and change the space vs. time trade-off, rather than
hold constant the space (per key) and change the accuracy vs. time
trade-off, though we might make that available.
### Benchmarking
```
$ ./filter_bench -impl=2 -quick -m_keys_total_max=200 -average_keys_per_filter=100000 -net_includes_hashing
Building...
Build avg ns/key: 34.1341
Number of filters: 1993
Total size (MB): 238.488
Reported total allocated memory (MB): 262.875
Reported internal fragmentation: 10.2255%
Bits/key stored: 10.0029
----------------------------
Mixed inside/outside queries...
Single filter net ns/op: 18.7508
Random filter net ns/op: 258.246
Average FP rate %: 0.968672
----------------------------
Done. (For more info, run with -legend or -help.)
$ ./filter_bench -impl=3 -quick -m_keys_total_max=200 -average_keys_per_filter=100000 -net_includes_hashing
Building...
Build avg ns/key: 130.851
Number of filters: 1993
Total size (MB): 168.166
Reported total allocated memory (MB): 183.211
Reported internal fragmentation: 8.94626%
Bits/key stored: 7.05341
----------------------------
Mixed inside/outside queries...
Single filter net ns/op: 58.4523
Random filter net ns/op: 363.717
Average FP rate %: 0.952978
----------------------------
Done. (For more info, run with -legend or -help.)
```
168.166 / 238.488 = 0.705 -> 29.5% space reduction
130.851 / 34.1341 = 3.83x construction time for this Ribbon filter vs. lastest Bloom filter (could make that as little as about 2.5x for less space reduction)
### Working around a hashing "flaw"
bloom_test discovered a flaw in the simple hashing applied in
StandardHasher when num_starts == 1 (num_slots == 128), showing an
excessively high FP rate. The problem is that when many entries, on the
order of number of hash bits or kCoeffBits, are associated with the same
start location, the correlation between the CoeffRow and ResultRow (for
efficiency) can lead to a solution that is "universal," or nearly so, for
entries mapping to that start location. (Normally, variance in start
location breaks the effective association between CoeffRow and
ResultRow; the same value for CoeffRow is effectively different if start
locations are different.) Without kUseSmash and with num_starts > 1 (thus
num_starts ~= num_slots), this flaw should be completely irrelevant. Even
with 10M slots, the chances of a single slot having just 16 (or more)
entries map to it--not enough to cause an FP problem, which would be local
to that slot if it happened--is 1 in millions. This spreadsheet formula
shows that: =1/(10000000*(1 - POISSON(15, 1, TRUE)))
As kUseSmash==false (the setting for Standard128RibbonBitsBuilder) is
intended for CPU efficiency of filters with many more entries/slots than
kCoeffBits, a very reasonable work-around is to disallow num_starts==1
when !kUseSmash, by making the minimum non-zero number of slots
2*kCoeffBits. This is the work-around I've applied. This also means that
the new Ribbon filter schema (Standard128RibbonBitsBuilder) is not
space-efficient for less than a few hundred entries. Because of this, I
have made it fall back on constructing a Bloom filter, under existing
schema, when that is more space efficient for small filters. (We can
change this in the future if we want.)
TODO: better unit tests for this case in ribbon_test, and probably
update StandardHasher for kUseSmash case so that it can scale nicely to
small filters.
### Other related changes
* Add Ribbon filter to stress/crash test
* Add Ribbon filter to filter_bench as -impl=3
* Add option string support, as in "filter_policy=experimental_ribbon:5.678;"
where 5.678 is the Bloom equivalent bits per key.
* Rename internal mode BloomFilterPolicy::kAuto to kAutoBloom
* Add a general BuiltinFilterBitsBuilder::CalculateNumEntry based on
binary searching CalculateSpace (inefficient), so that subclasses
(especially experimental ones) don't have to provide an efficient
implementation inverting CalculateSpace.
* Minor refactor FastLocalBloomBitsBuilder for new base class
XXH3pFilterBitsBuilder shared with new Standard128RibbonBitsBuilder,
which allows the latter to fall back on Bloom construction in some
extreme cases.
* Mostly updated bloom_test for Ribbon filter, though a test like
FullBloomTest::Schema is a next TODO to ensure schema stability
(in case this becomes production-ready schema as it is).
* Add some APIs to ribbon_impl.h for configuring Ribbon filters.
Although these are reasonably covered by bloom_test, TODO more unit
tests in ribbon_test
* Added a "tool" FindOccupancyForSuccessRate to ribbon_test to get data
for constructing the linear approximations in GetNumSlotsFor95PctSuccess.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/7658
Test Plan:
Some unit tests updated but other testing is left TODO. This
is considered experimental but laying down schema compatibility as early
as possible in case it proves production-quality. Also tested in
stress/crash test.
Reviewed By: jay-zhuang
Differential Revision: D24899349
Pulled By: pdillinger
fbshipit-source-id: 9715f3e6371c959d923aea8077c9423c7a9f82b8
2020-11-13 05:45:02 +01:00
|
|
|
if (FLAGS_impl > 3) {
|
New Bloom filter implementation for full and partitioned filters (#6007)
Summary:
Adds an improved, replacement Bloom filter implementation (FastLocalBloom) for full and partitioned filters in the block-based table. This replacement is faster and more accurate, especially for high bits per key or millions of keys in a single filter.
Speed
The improved speed, at least on recent x86_64, comes from
* Using fastrange instead of modulo (%)
* Using our new hash function (XXH3 preview, added in a previous commit), which is much faster for large keys and only *slightly* slower on keys around 12 bytes if hashing the same size many thousands of times in a row.
* Optimizing the Bloom filter queries with AVX2 SIMD operations. (Added AVX2 to the USE_SSE=1 build.) Careful design was required to support (a) SIMD-optimized queries, (b) compatible non-SIMD code that's simple and efficient, (c) flexible choice of number of probes, and (d) essentially maximized accuracy for a cache-local Bloom filter. Probes are made eight at a time, so any number of probes up to 8 is the same speed, then up to 16, etc.
* Prefetching cache lines when building the filter. Although this optimization could be applied to the old structure as well, it seems to balance out the small added cost of accumulating 64 bit hashes for adding to the filter rather than 32 bit hashes.
Here's nominal speed data from filter_bench (200MB in filters, about 10k keys each, 10 bits filter data / key, 6 probes, avg key size 24 bytes, includes hashing time) on Skylake DE (relatively low clock speed):
$ ./filter_bench -quick -impl=2 -net_includes_hashing # New Bloom filter
Build avg ns/key: 47.7135
Mixed inside/outside queries...
Single filter net ns/op: 26.2825
Random filter net ns/op: 150.459
Average FP rate %: 0.954651
$ ./filter_bench -quick -impl=0 -net_includes_hashing # Old Bloom filter
Build avg ns/key: 47.2245
Mixed inside/outside queries...
Single filter net ns/op: 63.2978
Random filter net ns/op: 188.038
Average FP rate %: 1.13823
Similar build time but dramatically faster query times on hot data (63 ns to 26 ns), and somewhat faster on stale data (188 ns to 150 ns). Performance differences on batched and skewed query loads are between these extremes as expected.
The only other interesting thing about speed is "inside" (query key was added to filter) vs. "outside" (query key was not added to filter) query times. The non-SIMD implementations are substantially slower when most queries are "outside" vs. "inside". This goes against what one might expect or would have observed years ago, as "outside" queries only need about two probes on average, due to short-circuiting, while "inside" always have num_probes (say 6). The problem is probably the nastily unpredictable branch. The SIMD implementation has few branches (very predictable) and has pretty consistent running time regardless of query outcome.
Accuracy
The generally improved accuracy (re: Issue https://github.com/facebook/rocksdb/issues/5857) comes from a better design for probing indices
within a cache line (re: Issue https://github.com/facebook/rocksdb/issues/4120) and improved accuracy for millions of keys in a single filter from using a 64-bit hash function (XXH3p). Design details in code comments.
Accuracy data (generalizes, except old impl gets worse with millions of keys):
Memory bits per key: FP rate percent old impl -> FP rate percent new impl
6: 5.70953 -> 5.69888
8: 2.45766 -> 2.29709
10: 1.13977 -> 0.959254
12: 0.662498 -> 0.411593
16: 0.353023 -> 0.0873754
24: 0.261552 -> 0.0060971
50: 0.225453 -> ~0.00003 (less than 1 in a million queries are FP)
Fixes https://github.com/facebook/rocksdb/issues/5857
Fixes https://github.com/facebook/rocksdb/issues/4120
Unlike the old implementation, this implementation has a fixed cache line size (64 bytes). At 10 bits per key, the accuracy of this new implementation is very close to the old implementation with 128-byte cache line size. If there's sufficient demand, this implementation could be generalized.
Compatibility
Although old releases would see the new structure as corrupt filter data and read the table as if there's no filter, we've decided only to enable the new Bloom filter with new format_version=5. This provides a smooth path for automatic adoption over time, with an option for early opt-in.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6007
Test Plan: filter_bench has been used thoroughly to validate speed, accuracy, and correctness. Unit tests have been carefully updated to exercise new and old implementations, as well as the logic to select an implementation based on context (format_version).
Differential Revision: D18294749
Pulled By: pdillinger
fbshipit-source-id: d44c9db3696e4d0a17caaec47075b7755c262c5f
2019-11-14 01:31:26 +01:00
|
|
|
throw std::runtime_error(
|
Experimental (production candidate) SST schema for Ribbon filter (#7658)
Summary:
Added experimental public API for Ribbon filter:
NewExperimentalRibbonFilterPolicy(). This experimental API will
take a "Bloom equivalent" bits per key, and configure the Ribbon
filter for the same FP rate as Bloom would have but ~30% space
savings. (Note: optimize_filters_for_memory is not yet implemented
for Ribbon filter. That can be added with no effect on schema.)
Internally, the Ribbon filter is configured using a "one_in_fp_rate"
value, which is 1 over desired FP rate. For example, use 100 for 1%
FP rate. I'm expecting this will be used in the future for configuring
Bloom-like filters, as I expect people to more commonly hold constant
the filter accuracy and change the space vs. time trade-off, rather than
hold constant the space (per key) and change the accuracy vs. time
trade-off, though we might make that available.
### Benchmarking
```
$ ./filter_bench -impl=2 -quick -m_keys_total_max=200 -average_keys_per_filter=100000 -net_includes_hashing
Building...
Build avg ns/key: 34.1341
Number of filters: 1993
Total size (MB): 238.488
Reported total allocated memory (MB): 262.875
Reported internal fragmentation: 10.2255%
Bits/key stored: 10.0029
----------------------------
Mixed inside/outside queries...
Single filter net ns/op: 18.7508
Random filter net ns/op: 258.246
Average FP rate %: 0.968672
----------------------------
Done. (For more info, run with -legend or -help.)
$ ./filter_bench -impl=3 -quick -m_keys_total_max=200 -average_keys_per_filter=100000 -net_includes_hashing
Building...
Build avg ns/key: 130.851
Number of filters: 1993
Total size (MB): 168.166
Reported total allocated memory (MB): 183.211
Reported internal fragmentation: 8.94626%
Bits/key stored: 7.05341
----------------------------
Mixed inside/outside queries...
Single filter net ns/op: 58.4523
Random filter net ns/op: 363.717
Average FP rate %: 0.952978
----------------------------
Done. (For more info, run with -legend or -help.)
```
168.166 / 238.488 = 0.705 -> 29.5% space reduction
130.851 / 34.1341 = 3.83x construction time for this Ribbon filter vs. lastest Bloom filter (could make that as little as about 2.5x for less space reduction)
### Working around a hashing "flaw"
bloom_test discovered a flaw in the simple hashing applied in
StandardHasher when num_starts == 1 (num_slots == 128), showing an
excessively high FP rate. The problem is that when many entries, on the
order of number of hash bits or kCoeffBits, are associated with the same
start location, the correlation between the CoeffRow and ResultRow (for
efficiency) can lead to a solution that is "universal," or nearly so, for
entries mapping to that start location. (Normally, variance in start
location breaks the effective association between CoeffRow and
ResultRow; the same value for CoeffRow is effectively different if start
locations are different.) Without kUseSmash and with num_starts > 1 (thus
num_starts ~= num_slots), this flaw should be completely irrelevant. Even
with 10M slots, the chances of a single slot having just 16 (or more)
entries map to it--not enough to cause an FP problem, which would be local
to that slot if it happened--is 1 in millions. This spreadsheet formula
shows that: =1/(10000000*(1 - POISSON(15, 1, TRUE)))
As kUseSmash==false (the setting for Standard128RibbonBitsBuilder) is
intended for CPU efficiency of filters with many more entries/slots than
kCoeffBits, a very reasonable work-around is to disallow num_starts==1
when !kUseSmash, by making the minimum non-zero number of slots
2*kCoeffBits. This is the work-around I've applied. This also means that
the new Ribbon filter schema (Standard128RibbonBitsBuilder) is not
space-efficient for less than a few hundred entries. Because of this, I
have made it fall back on constructing a Bloom filter, under existing
schema, when that is more space efficient for small filters. (We can
change this in the future if we want.)
TODO: better unit tests for this case in ribbon_test, and probably
update StandardHasher for kUseSmash case so that it can scale nicely to
small filters.
### Other related changes
* Add Ribbon filter to stress/crash test
* Add Ribbon filter to filter_bench as -impl=3
* Add option string support, as in "filter_policy=experimental_ribbon:5.678;"
where 5.678 is the Bloom equivalent bits per key.
* Rename internal mode BloomFilterPolicy::kAuto to kAutoBloom
* Add a general BuiltinFilterBitsBuilder::CalculateNumEntry based on
binary searching CalculateSpace (inefficient), so that subclasses
(especially experimental ones) don't have to provide an efficient
implementation inverting CalculateSpace.
* Minor refactor FastLocalBloomBitsBuilder for new base class
XXH3pFilterBitsBuilder shared with new Standard128RibbonBitsBuilder,
which allows the latter to fall back on Bloom construction in some
extreme cases.
* Mostly updated bloom_test for Ribbon filter, though a test like
FullBloomTest::Schema is a next TODO to ensure schema stability
(in case this becomes production-ready schema as it is).
* Add some APIs to ribbon_impl.h for configuring Ribbon filters.
Although these are reasonably covered by bloom_test, TODO more unit
tests in ribbon_test
* Added a "tool" FindOccupancyForSuccessRate to ribbon_test to get data
for constructing the linear approximations in GetNumSlotsFor95PctSuccess.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/7658
Test Plan:
Some unit tests updated but other testing is left TODO. This
is considered experimental but laying down schema compatibility as early
as possible in case it proves production-quality. Also tested in
stress/crash test.
Reviewed By: jay-zhuang
Differential Revision: D24899349
Pulled By: pdillinger
fbshipit-source-id: 9715f3e6371c959d923aea8077c9423c7a9f82b8
2020-11-13 05:45:02 +01:00
|
|
|
"-impl must currently be 0, 2, or 3 for Block-based table");
|
New Bloom filter implementation for full and partitioned filters (#6007)
Summary:
Adds an improved, replacement Bloom filter implementation (FastLocalBloom) for full and partitioned filters in the block-based table. This replacement is faster and more accurate, especially for high bits per key or millions of keys in a single filter.
Speed
The improved speed, at least on recent x86_64, comes from
* Using fastrange instead of modulo (%)
* Using our new hash function (XXH3 preview, added in a previous commit), which is much faster for large keys and only *slightly* slower on keys around 12 bytes if hashing the same size many thousands of times in a row.
* Optimizing the Bloom filter queries with AVX2 SIMD operations. (Added AVX2 to the USE_SSE=1 build.) Careful design was required to support (a) SIMD-optimized queries, (b) compatible non-SIMD code that's simple and efficient, (c) flexible choice of number of probes, and (d) essentially maximized accuracy for a cache-local Bloom filter. Probes are made eight at a time, so any number of probes up to 8 is the same speed, then up to 16, etc.
* Prefetching cache lines when building the filter. Although this optimization could be applied to the old structure as well, it seems to balance out the small added cost of accumulating 64 bit hashes for adding to the filter rather than 32 bit hashes.
Here's nominal speed data from filter_bench (200MB in filters, about 10k keys each, 10 bits filter data / key, 6 probes, avg key size 24 bytes, includes hashing time) on Skylake DE (relatively low clock speed):
$ ./filter_bench -quick -impl=2 -net_includes_hashing # New Bloom filter
Build avg ns/key: 47.7135
Mixed inside/outside queries...
Single filter net ns/op: 26.2825
Random filter net ns/op: 150.459
Average FP rate %: 0.954651
$ ./filter_bench -quick -impl=0 -net_includes_hashing # Old Bloom filter
Build avg ns/key: 47.2245
Mixed inside/outside queries...
Single filter net ns/op: 63.2978
Random filter net ns/op: 188.038
Average FP rate %: 1.13823
Similar build time but dramatically faster query times on hot data (63 ns to 26 ns), and somewhat faster on stale data (188 ns to 150 ns). Performance differences on batched and skewed query loads are between these extremes as expected.
The only other interesting thing about speed is "inside" (query key was added to filter) vs. "outside" (query key was not added to filter) query times. The non-SIMD implementations are substantially slower when most queries are "outside" vs. "inside". This goes against what one might expect or would have observed years ago, as "outside" queries only need about two probes on average, due to short-circuiting, while "inside" always have num_probes (say 6). The problem is probably the nastily unpredictable branch. The SIMD implementation has few branches (very predictable) and has pretty consistent running time regardless of query outcome.
Accuracy
The generally improved accuracy (re: Issue https://github.com/facebook/rocksdb/issues/5857) comes from a better design for probing indices
within a cache line (re: Issue https://github.com/facebook/rocksdb/issues/4120) and improved accuracy for millions of keys in a single filter from using a 64-bit hash function (XXH3p). Design details in code comments.
Accuracy data (generalizes, except old impl gets worse with millions of keys):
Memory bits per key: FP rate percent old impl -> FP rate percent new impl
6: 5.70953 -> 5.69888
8: 2.45766 -> 2.29709
10: 1.13977 -> 0.959254
12: 0.662498 -> 0.411593
16: 0.353023 -> 0.0873754
24: 0.261552 -> 0.0060971
50: 0.225453 -> ~0.00003 (less than 1 in a million queries are FP)
Fixes https://github.com/facebook/rocksdb/issues/5857
Fixes https://github.com/facebook/rocksdb/issues/4120
Unlike the old implementation, this implementation has a fixed cache line size (64 bytes). At 10 bits per key, the accuracy of this new implementation is very close to the old implementation with 128-byte cache line size. If there's sufficient demand, this implementation could be generalized.
Compatibility
Although old releases would see the new structure as corrupt filter data and read the table as if there's no filter, we've decided only to enable the new Bloom filter with new format_version=5. This provides a smooth path for automatic adoption over time, with an option for early opt-in.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6007
Test Plan: filter_bench has been used thoroughly to validate speed, accuracy, and correctness. Unit tests have been carefully updated to exercise new and old implementations, as well as the logic to select an implementation based on context (format_version).
Differential Revision: D18294749
Pulled By: pdillinger
fbshipit-source-id: d44c9db3696e4d0a17caaec47075b7755c262c5f
2019-11-14 01:31:26 +01:00
|
|
|
}
|
2019-10-25 22:25:28 +02:00
|
|
|
}
|
|
|
|
|
2020-02-21 22:29:41 +01:00
|
|
|
if (FLAGS_vary_key_count_ratio < 0.0 || FLAGS_vary_key_count_ratio > 1.0) {
|
|
|
|
throw std::runtime_error("-vary_key_count_ratio must be >= 0.0 and <= 1.0");
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
|
2020-02-21 22:29:41 +01:00
|
|
|
// For example, average_keys_per_filter = 100, vary_key_count_ratio = 0.1.
|
|
|
|
// Varys up to +/- 10 keys. variance_range = 21 (generating value 0..20).
|
|
|
|
// variance_offset = 10, so value - offset average value is always 0.
|
|
|
|
const uint32_t variance_range =
|
|
|
|
1 + 2 * static_cast<uint32_t>(FLAGS_vary_key_count_ratio *
|
|
|
|
FLAGS_average_keys_per_filter);
|
|
|
|
const uint32_t variance_offset = variance_range / 2;
|
|
|
|
|
2019-10-08 05:09:27 +02:00
|
|
|
const std::vector<TestMode> &testModes =
|
2019-10-25 22:25:28 +02:00
|
|
|
FLAGS_best_case ? bestCaseTestModes
|
|
|
|
: FLAGS_quick ? quickTestModes : allTestModes;
|
2020-02-21 22:29:41 +01:00
|
|
|
|
|
|
|
m_queries_ = FLAGS_m_queries;
|
|
|
|
double working_mem_size_mb = FLAGS_working_mem_size_mb;
|
2019-10-08 05:09:27 +02:00
|
|
|
if (FLAGS_quick) {
|
2020-02-21 22:29:41 +01:00
|
|
|
m_queries_ /= 7.0;
|
2019-10-25 22:25:28 +02:00
|
|
|
} else if (FLAGS_best_case) {
|
2020-02-21 22:29:41 +01:00
|
|
|
m_queries_ /= 3.0;
|
|
|
|
working_mem_size_mb /= 10.0;
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
std::cout << "Building..." << std::endl;
|
|
|
|
|
Warn on excessive keys for legacy Bloom filter with 32-bit hash (#6317)
Summary:
With many millions of keys, the old Bloom filter implementation
for the block-based table (format_version <= 4) would have excessive FP
rate due to the limitations of feeding the Bloom filter with a 32-bit hash.
This change computes an estimated inflated FP rate due to this effect
and warns in the log whenever an SST filter is constructed (almost
certainly a "full" not "partitioned" filter) that exceeds 1.5x FP rate
due to this effect. The detailed condition is only checked if 3 million
keys or more have been added to a filter, as this should be a lower
bound for common bits/key settings (< 20).
Recommended remedies include smaller SST file size, using
format_version >= 5 (for new Bloom filter), or using partitioned
filters.
This does not change behavior other than generating warnings for some
constructed filters using the old implementation.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6317
Test Plan:
Example with warning, 15M keys @ 15 bits / key: (working_mem_size_mb is just to stop after building one filter if it's large)
$ ./filter_bench -quick -impl=0 -working_mem_size_mb=1 -bits_per_key=15 -average_keys_per_filter=15000000 2>&1 | grep 'FP rate'
[WARN] [/block_based/filter_policy.cc:292] Using legacy SST/BBT Bloom filter with excessive key count (15.0M @ 15bpk), causing estimated 1.8x higher filter FP rate. Consider using new Bloom with format_version>=5, smaller SST file size, or partitioned filters.
Predicted FP rate %: 0.766702
Average FP rate %: 0.66846
Example without warning (150K keys):
$ ./filter_bench -quick -impl=0 -working_mem_size_mb=1 -bits_per_key=15 -average_keys_per_filter=150000 2>&1 | grep 'FP rate'
Predicted FP rate %: 0.422857
Average FP rate %: 0.379301
$
With more samples at 15 bits/key:
150K keys -> no warning; actual: 0.379% FP rate (baseline)
1M keys -> no warning; actual: 0.396% FP rate, 1.045x
9M keys -> no warning; actual: 0.563% FP rate, 1.485x
10M keys -> warning (1.5x); actual: 0.564% FP rate, 1.488x
15M keys -> warning (1.8x); actual: 0.668% FP rate, 1.76x
25M keys -> warning (2.4x); actual: 0.880% FP rate, 2.32x
At 10 bits/key:
150K keys -> no warning; actual: 1.17% FP rate (baseline)
1M keys -> no warning; actual: 1.16% FP rate
10M keys -> no warning; actual: 1.32% FP rate, 1.13x
25M keys -> no warning; actual: 1.63% FP rate, 1.39x
35M keys -> warning (1.6x); actual: 1.81% FP rate, 1.55x
At 5 bits/key:
150K keys -> no warning; actual: 9.32% FP rate (baseline)
25M keys -> no warning; actual: 9.62% FP rate, 1.03x
200M keys -> no warning; actual: 12.2% FP rate, 1.31x
250M keys -> warning (1.5x); actual: 12.8% FP rate, 1.37x
300M keys -> warning (1.6x); actual: 13.4% FP rate, 1.43x
The reason for the modest inaccuracy at low bits/key is that the assumption of independence between a collision between 32-hash values feeding the filter and an FP in the filter is not quite true for implementations using "simple" logic to compute indices from the stock hash result. There's math on this in my dissertation, but I don't think it's worth the effort just for these extreme cases (> 100 million keys and low-ish bits/key).
Differential Revision: D19471715
Pulled By: pdillinger
fbshipit-source-id: f80c96893a09bf1152630ff0b964e5cdd7e35c68
2020-01-21 06:30:22 +01:00
|
|
|
std::unique_ptr<BuiltinFilterBitsBuilder> builder;
|
2019-12-16 06:29:43 +01:00
|
|
|
|
2019-10-08 05:09:27 +02:00
|
|
|
size_t total_memory_used = 0;
|
Minimize memory internal fragmentation for Bloom filters (#6427)
Summary:
New experimental option BBTO::optimize_filters_for_memory builds
filters that maximize their use of "usable size" from malloc_usable_size,
which is also used to compute block cache charges.
Rather than always "rounding up," we track state in the
BloomFilterPolicy object to mix essentially "rounding down" and
"rounding up" so that the average FP rate of all generated filters is
the same as without the option. (YMMV as heavily accessed filters might
be unluckily lower accuracy.)
Thus, the option near-minimizes what the block cache considers as
"memory used" for a given target Bloom filter false positive rate and
Bloom filter implementation. There are no forward or backward
compatibility issues with this change, though it only works on the
format_version=5 Bloom filter.
With Jemalloc, we see about 10% reduction in memory footprint (and block
cache charge) for Bloom filters, but 1-2% increase in storage footprint,
due to encoding efficiency losses (FP rate is non-linear with bits/key).
Why not weighted random round up/down rather than state tracking? By
only requiring malloc_usable_size, we don't actually know what the next
larger and next smaller usable sizes for the allocator are. We pick a
requested size, accept and use whatever usable size it has, and use the
difference to inform our next choice. This allows us to narrow in on the
right balance without tracking/predicting usable sizes.
Why not weight history of generated filter false positive rates by
number of keys? This could lead to excess skew in small filters after
generating a large filter.
Results from filter_bench with jemalloc (irrelevant details omitted):
(normal keys/filter, but high variance)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.6278
Number of filters: 5516
Total size (MB): 200.046
Reported total allocated memory (MB): 220.597
Reported internal fragmentation: 10.2732%
Bits/key stored: 10.0097
Average FP rate %: 0.965228
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.5104
Number of filters: 5464
Total size (MB): 200.015
Reported total allocated memory (MB): 200.322
Reported internal fragmentation: 0.153709%
Bits/key stored: 10.1011
Average FP rate %: 0.966313
(very few keys / filter, optimization not as effective due to ~59 byte
internal fragmentation in blocked Bloom filter representation)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.5649
Number of filters: 162950
Total size (MB): 200.001
Reported total allocated memory (MB): 224.624
Reported internal fragmentation: 12.3117%
Bits/key stored: 10.2951
Average FP rate %: 0.821534
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 31.8057
Number of filters: 159849
Total size (MB): 200
Reported total allocated memory (MB): 208.846
Reported internal fragmentation: 4.42297%
Bits/key stored: 10.4948
Average FP rate %: 0.811006
(high keys/filter)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.7017
Number of filters: 164
Total size (MB): 200.352
Reported total allocated memory (MB): 221.5
Reported internal fragmentation: 10.5552%
Bits/key stored: 10.0003
Average FP rate %: 0.969358
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.7131
Number of filters: 160
Total size (MB): 200.928
Reported total allocated memory (MB): 200.938
Reported internal fragmentation: 0.00448054%
Bits/key stored: 10.1852
Average FP rate %: 0.963387
And from db_bench (block cache) with jemalloc:
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ ./db_bench -db=/dev/shm/dbbench -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -optimize_filters_for_memory -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ (for FILE in /dev/shm/dbbench.no_optimize/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17063835
$ (for FILE in /dev/shm/dbbench/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17430747
$ #^ 2.1% additional filter storage
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8440400
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 21087528
rocksdb.bloom.filter.useful COUNT : 4963889
rocksdb.bloom.filter.full.positive COUNT : 1214081
rocksdb.bloom.filter.full.true.positive COUNT : 1161999
$ #^ 1.04 % observed FP rate
$ ./db_bench -db=/dev/shm/dbbench -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -optimize_filters_for_memory -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8448592
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 18220328
rocksdb.bloom.filter.useful COUNT : 5360933
rocksdb.bloom.filter.full.positive COUNT : 1321315
rocksdb.bloom.filter.full.true.positive COUNT : 1262999
$ #^ 1.08 % observed FP rate, 13.6% less memory usage for filters
(Due to specific key density, this example tends to generate filters that are "worse than average" for internal fragmentation. "Better than average" cases can show little or no improvement.)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6427
Test Plan: unit test added, 'make check' with gcc, clang and valgrind
Reviewed By: siying
Differential Revision: D22124374
Pulled By: pdillinger
fbshipit-source-id: f3e3aa152f9043ddf4fae25799e76341d0d8714e
2020-06-22 22:30:57 +02:00
|
|
|
size_t total_size = 0;
|
2019-10-08 05:09:27 +02:00
|
|
|
size_t total_keys_added = 0;
|
Warn on excessive keys for legacy Bloom filter with 32-bit hash (#6317)
Summary:
With many millions of keys, the old Bloom filter implementation
for the block-based table (format_version <= 4) would have excessive FP
rate due to the limitations of feeding the Bloom filter with a 32-bit hash.
This change computes an estimated inflated FP rate due to this effect
and warns in the log whenever an SST filter is constructed (almost
certainly a "full" not "partitioned" filter) that exceeds 1.5x FP rate
due to this effect. The detailed condition is only checked if 3 million
keys or more have been added to a filter, as this should be a lower
bound for common bits/key settings (< 20).
Recommended remedies include smaller SST file size, using
format_version >= 5 (for new Bloom filter), or using partitioned
filters.
This does not change behavior other than generating warnings for some
constructed filters using the old implementation.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6317
Test Plan:
Example with warning, 15M keys @ 15 bits / key: (working_mem_size_mb is just to stop after building one filter if it's large)
$ ./filter_bench -quick -impl=0 -working_mem_size_mb=1 -bits_per_key=15 -average_keys_per_filter=15000000 2>&1 | grep 'FP rate'
[WARN] [/block_based/filter_policy.cc:292] Using legacy SST/BBT Bloom filter with excessive key count (15.0M @ 15bpk), causing estimated 1.8x higher filter FP rate. Consider using new Bloom with format_version>=5, smaller SST file size, or partitioned filters.
Predicted FP rate %: 0.766702
Average FP rate %: 0.66846
Example without warning (150K keys):
$ ./filter_bench -quick -impl=0 -working_mem_size_mb=1 -bits_per_key=15 -average_keys_per_filter=150000 2>&1 | grep 'FP rate'
Predicted FP rate %: 0.422857
Average FP rate %: 0.379301
$
With more samples at 15 bits/key:
150K keys -> no warning; actual: 0.379% FP rate (baseline)
1M keys -> no warning; actual: 0.396% FP rate, 1.045x
9M keys -> no warning; actual: 0.563% FP rate, 1.485x
10M keys -> warning (1.5x); actual: 0.564% FP rate, 1.488x
15M keys -> warning (1.8x); actual: 0.668% FP rate, 1.76x
25M keys -> warning (2.4x); actual: 0.880% FP rate, 2.32x
At 10 bits/key:
150K keys -> no warning; actual: 1.17% FP rate (baseline)
1M keys -> no warning; actual: 1.16% FP rate
10M keys -> no warning; actual: 1.32% FP rate, 1.13x
25M keys -> no warning; actual: 1.63% FP rate, 1.39x
35M keys -> warning (1.6x); actual: 1.81% FP rate, 1.55x
At 5 bits/key:
150K keys -> no warning; actual: 9.32% FP rate (baseline)
25M keys -> no warning; actual: 9.62% FP rate, 1.03x
200M keys -> no warning; actual: 12.2% FP rate, 1.31x
250M keys -> warning (1.5x); actual: 12.8% FP rate, 1.37x
300M keys -> warning (1.6x); actual: 13.4% FP rate, 1.43x
The reason for the modest inaccuracy at low bits/key is that the assumption of independence between a collision between 32-hash values feeding the filter and an FP in the filter is not quite true for implementations using "simple" logic to compute indices from the stock hash result. There's math on this in my dissertation, but I don't think it's worth the effort just for these extreme cases (> 100 million keys and low-ish bits/key).
Differential Revision: D19471715
Pulled By: pdillinger
fbshipit-source-id: f80c96893a09bf1152630ff0b964e5cdd7e35c68
2020-01-21 06:30:22 +01:00
|
|
|
#ifdef PREDICT_FP_RATE
|
|
|
|
double weighted_predicted_fp_rate = 0.0;
|
|
|
|
#endif
|
2020-02-21 22:29:41 +01:00
|
|
|
size_t max_total_keys;
|
|
|
|
size_t max_mem;
|
|
|
|
if (FLAGS_m_keys_total_max > 0) {
|
|
|
|
max_total_keys = static_cast<size_t>(1000000 * FLAGS_m_keys_total_max);
|
|
|
|
max_mem = SIZE_MAX;
|
|
|
|
} else {
|
|
|
|
max_total_keys = SIZE_MAX;
|
|
|
|
max_mem = static_cast<size_t>(1024 * 1024 * working_mem_size_mb);
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
|
2021-01-26 07:07:26 +01:00
|
|
|
ROCKSDB_NAMESPACE::StopWatchNano timer(
|
2021-03-15 12:32:24 +01:00
|
|
|
ROCKSDB_NAMESPACE::SystemClock::Default().get(), true);
|
2019-10-08 05:09:27 +02:00
|
|
|
|
2020-02-21 22:29:41 +01:00
|
|
|
infos_.clear();
|
Minimize memory internal fragmentation for Bloom filters (#6427)
Summary:
New experimental option BBTO::optimize_filters_for_memory builds
filters that maximize their use of "usable size" from malloc_usable_size,
which is also used to compute block cache charges.
Rather than always "rounding up," we track state in the
BloomFilterPolicy object to mix essentially "rounding down" and
"rounding up" so that the average FP rate of all generated filters is
the same as without the option. (YMMV as heavily accessed filters might
be unluckily lower accuracy.)
Thus, the option near-minimizes what the block cache considers as
"memory used" for a given target Bloom filter false positive rate and
Bloom filter implementation. There are no forward or backward
compatibility issues with this change, though it only works on the
format_version=5 Bloom filter.
With Jemalloc, we see about 10% reduction in memory footprint (and block
cache charge) for Bloom filters, but 1-2% increase in storage footprint,
due to encoding efficiency losses (FP rate is non-linear with bits/key).
Why not weighted random round up/down rather than state tracking? By
only requiring malloc_usable_size, we don't actually know what the next
larger and next smaller usable sizes for the allocator are. We pick a
requested size, accept and use whatever usable size it has, and use the
difference to inform our next choice. This allows us to narrow in on the
right balance without tracking/predicting usable sizes.
Why not weight history of generated filter false positive rates by
number of keys? This could lead to excess skew in small filters after
generating a large filter.
Results from filter_bench with jemalloc (irrelevant details omitted):
(normal keys/filter, but high variance)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.6278
Number of filters: 5516
Total size (MB): 200.046
Reported total allocated memory (MB): 220.597
Reported internal fragmentation: 10.2732%
Bits/key stored: 10.0097
Average FP rate %: 0.965228
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.5104
Number of filters: 5464
Total size (MB): 200.015
Reported total allocated memory (MB): 200.322
Reported internal fragmentation: 0.153709%
Bits/key stored: 10.1011
Average FP rate %: 0.966313
(very few keys / filter, optimization not as effective due to ~59 byte
internal fragmentation in blocked Bloom filter representation)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.5649
Number of filters: 162950
Total size (MB): 200.001
Reported total allocated memory (MB): 224.624
Reported internal fragmentation: 12.3117%
Bits/key stored: 10.2951
Average FP rate %: 0.821534
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 31.8057
Number of filters: 159849
Total size (MB): 200
Reported total allocated memory (MB): 208.846
Reported internal fragmentation: 4.42297%
Bits/key stored: 10.4948
Average FP rate %: 0.811006
(high keys/filter)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.7017
Number of filters: 164
Total size (MB): 200.352
Reported total allocated memory (MB): 221.5
Reported internal fragmentation: 10.5552%
Bits/key stored: 10.0003
Average FP rate %: 0.969358
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.7131
Number of filters: 160
Total size (MB): 200.928
Reported total allocated memory (MB): 200.938
Reported internal fragmentation: 0.00448054%
Bits/key stored: 10.1852
Average FP rate %: 0.963387
And from db_bench (block cache) with jemalloc:
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ ./db_bench -db=/dev/shm/dbbench -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -optimize_filters_for_memory -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ (for FILE in /dev/shm/dbbench.no_optimize/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17063835
$ (for FILE in /dev/shm/dbbench/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17430747
$ #^ 2.1% additional filter storage
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8440400
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 21087528
rocksdb.bloom.filter.useful COUNT : 4963889
rocksdb.bloom.filter.full.positive COUNT : 1214081
rocksdb.bloom.filter.full.true.positive COUNT : 1161999
$ #^ 1.04 % observed FP rate
$ ./db_bench -db=/dev/shm/dbbench -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -optimize_filters_for_memory -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8448592
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 18220328
rocksdb.bloom.filter.useful COUNT : 5360933
rocksdb.bloom.filter.full.positive COUNT : 1321315
rocksdb.bloom.filter.full.true.positive COUNT : 1262999
$ #^ 1.08 % observed FP rate, 13.6% less memory usage for filters
(Due to specific key density, this example tends to generate filters that are "worse than average" for internal fragmentation. "Better than average" cases can show little or no improvement.)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6427
Test Plan: unit test added, 'make check' with gcc, clang and valgrind
Reviewed By: siying
Differential Revision: D22124374
Pulled By: pdillinger
fbshipit-source-id: f3e3aa152f9043ddf4fae25799e76341d0d8714e
2020-06-22 22:30:57 +02:00
|
|
|
while ((working_mem_size_mb == 0 || total_size < max_mem) &&
|
2020-02-21 22:29:41 +01:00
|
|
|
total_keys_added < max_total_keys) {
|
2019-10-09 04:19:43 +02:00
|
|
|
uint32_t filter_id = random_.Next();
|
2019-10-08 05:09:27 +02:00
|
|
|
uint32_t keys_to_add = FLAGS_average_keys_per_filter +
|
2020-09-28 20:33:31 +02:00
|
|
|
FastRange32(random_.Next(), variance_range) -
|
2020-02-21 22:29:41 +01:00
|
|
|
variance_offset;
|
|
|
|
if (max_total_keys - total_keys_added < keys_to_add) {
|
|
|
|
keys_to_add = static_cast<uint32_t>(max_total_keys - total_keys_added);
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
infos_.emplace_back();
|
|
|
|
FilterInfo &info = infos_.back();
|
|
|
|
info.filter_id_ = filter_id;
|
|
|
|
info.keys_added_ = keys_to_add;
|
2019-10-25 22:25:28 +02:00
|
|
|
if (FLAGS_use_plain_table_bloom) {
|
|
|
|
info.plain_table_bloom_.reset(new PlainTableBloomV1());
|
|
|
|
info.plain_table_bloom_->SetTotalBits(
|
2020-01-31 01:05:44 +01:00
|
|
|
&arena_, static_cast<uint32_t>(keys_to_add * FLAGS_bits_per_key),
|
|
|
|
FLAGS_impl, 0 /*huge_page*/, nullptr /*logger*/);
|
2019-10-25 22:25:28 +02:00
|
|
|
for (uint32_t i = 0; i < keys_to_add; ++i) {
|
|
|
|
uint32_t hash = GetSliceHash(kms_[0].Get(filter_id, i));
|
|
|
|
info.plain_table_bloom_->AddHash(hash);
|
|
|
|
}
|
|
|
|
info.filter_ = info.plain_table_bloom_->GetRawData();
|
|
|
|
} else {
|
2019-12-16 06:29:43 +01:00
|
|
|
if (!builder) {
|
2020-04-29 22:06:27 +02:00
|
|
|
builder.reset(
|
|
|
|
static_cast_with_check<BuiltinFilterBitsBuilder>(GetBuilder()));
|
2019-12-16 06:29:43 +01:00
|
|
|
}
|
2019-10-25 22:25:28 +02:00
|
|
|
for (uint32_t i = 0; i < keys_to_add; ++i) {
|
|
|
|
builder->AddKey(kms_[0].Get(filter_id, i));
|
|
|
|
}
|
|
|
|
info.filter_ = builder->Finish(&info.owner_);
|
Warn on excessive keys for legacy Bloom filter with 32-bit hash (#6317)
Summary:
With many millions of keys, the old Bloom filter implementation
for the block-based table (format_version <= 4) would have excessive FP
rate due to the limitations of feeding the Bloom filter with a 32-bit hash.
This change computes an estimated inflated FP rate due to this effect
and warns in the log whenever an SST filter is constructed (almost
certainly a "full" not "partitioned" filter) that exceeds 1.5x FP rate
due to this effect. The detailed condition is only checked if 3 million
keys or more have been added to a filter, as this should be a lower
bound for common bits/key settings (< 20).
Recommended remedies include smaller SST file size, using
format_version >= 5 (for new Bloom filter), or using partitioned
filters.
This does not change behavior other than generating warnings for some
constructed filters using the old implementation.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6317
Test Plan:
Example with warning, 15M keys @ 15 bits / key: (working_mem_size_mb is just to stop after building one filter if it's large)
$ ./filter_bench -quick -impl=0 -working_mem_size_mb=1 -bits_per_key=15 -average_keys_per_filter=15000000 2>&1 | grep 'FP rate'
[WARN] [/block_based/filter_policy.cc:292] Using legacy SST/BBT Bloom filter with excessive key count (15.0M @ 15bpk), causing estimated 1.8x higher filter FP rate. Consider using new Bloom with format_version>=5, smaller SST file size, or partitioned filters.
Predicted FP rate %: 0.766702
Average FP rate %: 0.66846
Example without warning (150K keys):
$ ./filter_bench -quick -impl=0 -working_mem_size_mb=1 -bits_per_key=15 -average_keys_per_filter=150000 2>&1 | grep 'FP rate'
Predicted FP rate %: 0.422857
Average FP rate %: 0.379301
$
With more samples at 15 bits/key:
150K keys -> no warning; actual: 0.379% FP rate (baseline)
1M keys -> no warning; actual: 0.396% FP rate, 1.045x
9M keys -> no warning; actual: 0.563% FP rate, 1.485x
10M keys -> warning (1.5x); actual: 0.564% FP rate, 1.488x
15M keys -> warning (1.8x); actual: 0.668% FP rate, 1.76x
25M keys -> warning (2.4x); actual: 0.880% FP rate, 2.32x
At 10 bits/key:
150K keys -> no warning; actual: 1.17% FP rate (baseline)
1M keys -> no warning; actual: 1.16% FP rate
10M keys -> no warning; actual: 1.32% FP rate, 1.13x
25M keys -> no warning; actual: 1.63% FP rate, 1.39x
35M keys -> warning (1.6x); actual: 1.81% FP rate, 1.55x
At 5 bits/key:
150K keys -> no warning; actual: 9.32% FP rate (baseline)
25M keys -> no warning; actual: 9.62% FP rate, 1.03x
200M keys -> no warning; actual: 12.2% FP rate, 1.31x
250M keys -> warning (1.5x); actual: 12.8% FP rate, 1.37x
300M keys -> warning (1.6x); actual: 13.4% FP rate, 1.43x
The reason for the modest inaccuracy at low bits/key is that the assumption of independence between a collision between 32-hash values feeding the filter and an FP in the filter is not quite true for implementations using "simple" logic to compute indices from the stock hash result. There's math on this in my dissertation, but I don't think it's worth the effort just for these extreme cases (> 100 million keys and low-ish bits/key).
Differential Revision: D19471715
Pulled By: pdillinger
fbshipit-source-id: f80c96893a09bf1152630ff0b964e5cdd7e35c68
2020-01-21 06:30:22 +01:00
|
|
|
#ifdef PREDICT_FP_RATE
|
|
|
|
weighted_predicted_fp_rate +=
|
|
|
|
keys_to_add *
|
|
|
|
builder->EstimatedFpRate(keys_to_add, info.filter_.size());
|
|
|
|
#endif
|
2019-12-16 06:29:43 +01:00
|
|
|
if (FLAGS_new_builder) {
|
|
|
|
builder.reset();
|
|
|
|
}
|
2019-10-25 22:25:28 +02:00
|
|
|
info.reader_.reset(
|
|
|
|
table_options_.filter_policy->GetFilterBitsReader(info.filter_));
|
|
|
|
CachableEntry<ParsedFullFilterBlock> block(
|
|
|
|
new ParsedFullFilterBlock(table_options_.filter_policy.get(),
|
|
|
|
BlockContents(info.filter_)),
|
|
|
|
nullptr /* cache */, nullptr /* cache_handle */,
|
|
|
|
true /* own_value */);
|
|
|
|
info.full_block_reader_.reset(
|
|
|
|
new FullFilterBlockReader(table_.get(), std::move(block)));
|
|
|
|
}
|
Minimize memory internal fragmentation for Bloom filters (#6427)
Summary:
New experimental option BBTO::optimize_filters_for_memory builds
filters that maximize their use of "usable size" from malloc_usable_size,
which is also used to compute block cache charges.
Rather than always "rounding up," we track state in the
BloomFilterPolicy object to mix essentially "rounding down" and
"rounding up" so that the average FP rate of all generated filters is
the same as without the option. (YMMV as heavily accessed filters might
be unluckily lower accuracy.)
Thus, the option near-minimizes what the block cache considers as
"memory used" for a given target Bloom filter false positive rate and
Bloom filter implementation. There are no forward or backward
compatibility issues with this change, though it only works on the
format_version=5 Bloom filter.
With Jemalloc, we see about 10% reduction in memory footprint (and block
cache charge) for Bloom filters, but 1-2% increase in storage footprint,
due to encoding efficiency losses (FP rate is non-linear with bits/key).
Why not weighted random round up/down rather than state tracking? By
only requiring malloc_usable_size, we don't actually know what the next
larger and next smaller usable sizes for the allocator are. We pick a
requested size, accept and use whatever usable size it has, and use the
difference to inform our next choice. This allows us to narrow in on the
right balance without tracking/predicting usable sizes.
Why not weight history of generated filter false positive rates by
number of keys? This could lead to excess skew in small filters after
generating a large filter.
Results from filter_bench with jemalloc (irrelevant details omitted):
(normal keys/filter, but high variance)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.6278
Number of filters: 5516
Total size (MB): 200.046
Reported total allocated memory (MB): 220.597
Reported internal fragmentation: 10.2732%
Bits/key stored: 10.0097
Average FP rate %: 0.965228
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.5104
Number of filters: 5464
Total size (MB): 200.015
Reported total allocated memory (MB): 200.322
Reported internal fragmentation: 0.153709%
Bits/key stored: 10.1011
Average FP rate %: 0.966313
(very few keys / filter, optimization not as effective due to ~59 byte
internal fragmentation in blocked Bloom filter representation)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.5649
Number of filters: 162950
Total size (MB): 200.001
Reported total allocated memory (MB): 224.624
Reported internal fragmentation: 12.3117%
Bits/key stored: 10.2951
Average FP rate %: 0.821534
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 31.8057
Number of filters: 159849
Total size (MB): 200
Reported total allocated memory (MB): 208.846
Reported internal fragmentation: 4.42297%
Bits/key stored: 10.4948
Average FP rate %: 0.811006
(high keys/filter)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.7017
Number of filters: 164
Total size (MB): 200.352
Reported total allocated memory (MB): 221.5
Reported internal fragmentation: 10.5552%
Bits/key stored: 10.0003
Average FP rate %: 0.969358
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.7131
Number of filters: 160
Total size (MB): 200.928
Reported total allocated memory (MB): 200.938
Reported internal fragmentation: 0.00448054%
Bits/key stored: 10.1852
Average FP rate %: 0.963387
And from db_bench (block cache) with jemalloc:
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ ./db_bench -db=/dev/shm/dbbench -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -optimize_filters_for_memory -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ (for FILE in /dev/shm/dbbench.no_optimize/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17063835
$ (for FILE in /dev/shm/dbbench/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17430747
$ #^ 2.1% additional filter storage
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8440400
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 21087528
rocksdb.bloom.filter.useful COUNT : 4963889
rocksdb.bloom.filter.full.positive COUNT : 1214081
rocksdb.bloom.filter.full.true.positive COUNT : 1161999
$ #^ 1.04 % observed FP rate
$ ./db_bench -db=/dev/shm/dbbench -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -optimize_filters_for_memory -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8448592
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 18220328
rocksdb.bloom.filter.useful COUNT : 5360933
rocksdb.bloom.filter.full.positive COUNT : 1321315
rocksdb.bloom.filter.full.true.positive COUNT : 1262999
$ #^ 1.08 % observed FP rate, 13.6% less memory usage for filters
(Due to specific key density, this example tends to generate filters that are "worse than average" for internal fragmentation. "Better than average" cases can show little or no improvement.)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6427
Test Plan: unit test added, 'make check' with gcc, clang and valgrind
Reviewed By: siying
Differential Revision: D22124374
Pulled By: pdillinger
fbshipit-source-id: f3e3aa152f9043ddf4fae25799e76341d0d8714e
2020-06-22 22:30:57 +02:00
|
|
|
total_size += info.filter_.size();
|
|
|
|
#ifdef ROCKSDB_MALLOC_USABLE_SIZE
|
|
|
|
total_memory_used +=
|
|
|
|
malloc_usable_size(const_cast<char *>(info.filter_.data()));
|
|
|
|
#endif // ROCKSDB_MALLOC_USABLE_SIZE
|
2019-10-08 05:09:27 +02:00
|
|
|
total_keys_added += keys_to_add;
|
|
|
|
}
|
|
|
|
|
|
|
|
uint64_t elapsed_nanos = timer.ElapsedNanos();
|
|
|
|
double ns = double(elapsed_nanos) / total_keys_added;
|
|
|
|
std::cout << "Build avg ns/key: " << ns << std::endl;
|
|
|
|
std::cout << "Number of filters: " << infos_.size() << std::endl;
|
Minimize memory internal fragmentation for Bloom filters (#6427)
Summary:
New experimental option BBTO::optimize_filters_for_memory builds
filters that maximize their use of "usable size" from malloc_usable_size,
which is also used to compute block cache charges.
Rather than always "rounding up," we track state in the
BloomFilterPolicy object to mix essentially "rounding down" and
"rounding up" so that the average FP rate of all generated filters is
the same as without the option. (YMMV as heavily accessed filters might
be unluckily lower accuracy.)
Thus, the option near-minimizes what the block cache considers as
"memory used" for a given target Bloom filter false positive rate and
Bloom filter implementation. There are no forward or backward
compatibility issues with this change, though it only works on the
format_version=5 Bloom filter.
With Jemalloc, we see about 10% reduction in memory footprint (and block
cache charge) for Bloom filters, but 1-2% increase in storage footprint,
due to encoding efficiency losses (FP rate is non-linear with bits/key).
Why not weighted random round up/down rather than state tracking? By
only requiring malloc_usable_size, we don't actually know what the next
larger and next smaller usable sizes for the allocator are. We pick a
requested size, accept and use whatever usable size it has, and use the
difference to inform our next choice. This allows us to narrow in on the
right balance without tracking/predicting usable sizes.
Why not weight history of generated filter false positive rates by
number of keys? This could lead to excess skew in small filters after
generating a large filter.
Results from filter_bench with jemalloc (irrelevant details omitted):
(normal keys/filter, but high variance)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.6278
Number of filters: 5516
Total size (MB): 200.046
Reported total allocated memory (MB): 220.597
Reported internal fragmentation: 10.2732%
Bits/key stored: 10.0097
Average FP rate %: 0.965228
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.5104
Number of filters: 5464
Total size (MB): 200.015
Reported total allocated memory (MB): 200.322
Reported internal fragmentation: 0.153709%
Bits/key stored: 10.1011
Average FP rate %: 0.966313
(very few keys / filter, optimization not as effective due to ~59 byte
internal fragmentation in blocked Bloom filter representation)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.5649
Number of filters: 162950
Total size (MB): 200.001
Reported total allocated memory (MB): 224.624
Reported internal fragmentation: 12.3117%
Bits/key stored: 10.2951
Average FP rate %: 0.821534
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 31.8057
Number of filters: 159849
Total size (MB): 200
Reported total allocated memory (MB): 208.846
Reported internal fragmentation: 4.42297%
Bits/key stored: 10.4948
Average FP rate %: 0.811006
(high keys/filter)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.7017
Number of filters: 164
Total size (MB): 200.352
Reported total allocated memory (MB): 221.5
Reported internal fragmentation: 10.5552%
Bits/key stored: 10.0003
Average FP rate %: 0.969358
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.7131
Number of filters: 160
Total size (MB): 200.928
Reported total allocated memory (MB): 200.938
Reported internal fragmentation: 0.00448054%
Bits/key stored: 10.1852
Average FP rate %: 0.963387
And from db_bench (block cache) with jemalloc:
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ ./db_bench -db=/dev/shm/dbbench -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -optimize_filters_for_memory -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ (for FILE in /dev/shm/dbbench.no_optimize/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17063835
$ (for FILE in /dev/shm/dbbench/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17430747
$ #^ 2.1% additional filter storage
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8440400
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 21087528
rocksdb.bloom.filter.useful COUNT : 4963889
rocksdb.bloom.filter.full.positive COUNT : 1214081
rocksdb.bloom.filter.full.true.positive COUNT : 1161999
$ #^ 1.04 % observed FP rate
$ ./db_bench -db=/dev/shm/dbbench -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -optimize_filters_for_memory -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8448592
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 18220328
rocksdb.bloom.filter.useful COUNT : 5360933
rocksdb.bloom.filter.full.positive COUNT : 1321315
rocksdb.bloom.filter.full.true.positive COUNT : 1262999
$ #^ 1.08 % observed FP rate, 13.6% less memory usage for filters
(Due to specific key density, this example tends to generate filters that are "worse than average" for internal fragmentation. "Better than average" cases can show little or no improvement.)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6427
Test Plan: unit test added, 'make check' with gcc, clang and valgrind
Reviewed By: siying
Differential Revision: D22124374
Pulled By: pdillinger
fbshipit-source-id: f3e3aa152f9043ddf4fae25799e76341d0d8714e
2020-06-22 22:30:57 +02:00
|
|
|
std::cout << "Total size (MB): " << total_size / 1024.0 / 1024.0 << std::endl;
|
|
|
|
if (total_memory_used > 0) {
|
|
|
|
std::cout << "Reported total allocated memory (MB): "
|
|
|
|
<< total_memory_used / 1024.0 / 1024.0 << std::endl;
|
|
|
|
std::cout << "Reported internal fragmentation: "
|
|
|
|
<< (total_memory_used - total_size) * 100.0 / total_size << "%"
|
|
|
|
<< std::endl;
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
|
Minimize memory internal fragmentation for Bloom filters (#6427)
Summary:
New experimental option BBTO::optimize_filters_for_memory builds
filters that maximize their use of "usable size" from malloc_usable_size,
which is also used to compute block cache charges.
Rather than always "rounding up," we track state in the
BloomFilterPolicy object to mix essentially "rounding down" and
"rounding up" so that the average FP rate of all generated filters is
the same as without the option. (YMMV as heavily accessed filters might
be unluckily lower accuracy.)
Thus, the option near-minimizes what the block cache considers as
"memory used" for a given target Bloom filter false positive rate and
Bloom filter implementation. There are no forward or backward
compatibility issues with this change, though it only works on the
format_version=5 Bloom filter.
With Jemalloc, we see about 10% reduction in memory footprint (and block
cache charge) for Bloom filters, but 1-2% increase in storage footprint,
due to encoding efficiency losses (FP rate is non-linear with bits/key).
Why not weighted random round up/down rather than state tracking? By
only requiring malloc_usable_size, we don't actually know what the next
larger and next smaller usable sizes for the allocator are. We pick a
requested size, accept and use whatever usable size it has, and use the
difference to inform our next choice. This allows us to narrow in on the
right balance without tracking/predicting usable sizes.
Why not weight history of generated filter false positive rates by
number of keys? This could lead to excess skew in small filters after
generating a large filter.
Results from filter_bench with jemalloc (irrelevant details omitted):
(normal keys/filter, but high variance)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.6278
Number of filters: 5516
Total size (MB): 200.046
Reported total allocated memory (MB): 220.597
Reported internal fragmentation: 10.2732%
Bits/key stored: 10.0097
Average FP rate %: 0.965228
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=30000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.5104
Number of filters: 5464
Total size (MB): 200.015
Reported total allocated memory (MB): 200.322
Reported internal fragmentation: 0.153709%
Bits/key stored: 10.1011
Average FP rate %: 0.966313
(very few keys / filter, optimization not as effective due to ~59 byte
internal fragmentation in blocked Bloom filter representation)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.5649
Number of filters: 162950
Total size (MB): 200.001
Reported total allocated memory (MB): 224.624
Reported internal fragmentation: 12.3117%
Bits/key stored: 10.2951
Average FP rate %: 0.821534
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 31.8057
Number of filters: 159849
Total size (MB): 200
Reported total allocated memory (MB): 208.846
Reported internal fragmentation: 4.42297%
Bits/key stored: 10.4948
Average FP rate %: 0.811006
(high keys/filter)
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9
Build avg ns/key: 29.7017
Number of filters: 164
Total size (MB): 200.352
Reported total allocated memory (MB): 221.5
Reported internal fragmentation: 10.5552%
Bits/key stored: 10.0003
Average FP rate %: 0.969358
$ ./filter_bench -quick -impl=2 -average_keys_per_filter=1000000 -vary_key_count_ratio=0.9 -optimize_filters_for_memory
Build avg ns/key: 30.7131
Number of filters: 160
Total size (MB): 200.928
Reported total allocated memory (MB): 200.938
Reported internal fragmentation: 0.00448054%
Bits/key stored: 10.1852
Average FP rate %: 0.963387
And from db_bench (block cache) with jemalloc:
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ ./db_bench -db=/dev/shm/dbbench -benchmarks=fillrandom -format_version=5 -value_size=90 -bloom_bits=10 -num=2000000 -threads=8 -optimize_filters_for_memory -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false
$ (for FILE in /dev/shm/dbbench.no_optimize/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17063835
$ (for FILE in /dev/shm/dbbench/*.sst; do ./sst_dump --file=$FILE --show_properties | grep 'filter block' ; done) | awk '{ t += $4; } END { print t; }'
17430747
$ #^ 2.1% additional filter storage
$ ./db_bench -db=/dev/shm/dbbench.no_optimize -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8440400
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 21087528
rocksdb.bloom.filter.useful COUNT : 4963889
rocksdb.bloom.filter.full.positive COUNT : 1214081
rocksdb.bloom.filter.full.true.positive COUNT : 1161999
$ #^ 1.04 % observed FP rate
$ ./db_bench -db=/dev/shm/dbbench -use_existing_db -benchmarks=readrandom,stats -statistics -bloom_bits=10 -num=2000000 -compaction_style=2 -fifo_compaction_max_table_files_size_mb=10000 -fifo_compaction_allow_compaction=false -optimize_filters_for_memory -duration=10 -cache_index_and_filter_blocks -cache_size=1000000000
rocksdb.block.cache.index.add COUNT : 33
rocksdb.block.cache.index.bytes.insert COUNT : 8448592
rocksdb.block.cache.filter.add COUNT : 33
rocksdb.block.cache.filter.bytes.insert COUNT : 18220328
rocksdb.bloom.filter.useful COUNT : 5360933
rocksdb.bloom.filter.full.positive COUNT : 1321315
rocksdb.bloom.filter.full.true.positive COUNT : 1262999
$ #^ 1.08 % observed FP rate, 13.6% less memory usage for filters
(Due to specific key density, this example tends to generate filters that are "worse than average" for internal fragmentation. "Better than average" cases can show little or no improvement.)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6427
Test Plan: unit test added, 'make check' with gcc, clang and valgrind
Reviewed By: siying
Differential Revision: D22124374
Pulled By: pdillinger
fbshipit-source-id: f3e3aa152f9043ddf4fae25799e76341d0d8714e
2020-06-22 22:30:57 +02:00
|
|
|
double bpk = total_size * 8.0 / total_keys_added;
|
|
|
|
std::cout << "Bits/key stored: " << bpk << std::endl;
|
Warn on excessive keys for legacy Bloom filter with 32-bit hash (#6317)
Summary:
With many millions of keys, the old Bloom filter implementation
for the block-based table (format_version <= 4) would have excessive FP
rate due to the limitations of feeding the Bloom filter with a 32-bit hash.
This change computes an estimated inflated FP rate due to this effect
and warns in the log whenever an SST filter is constructed (almost
certainly a "full" not "partitioned" filter) that exceeds 1.5x FP rate
due to this effect. The detailed condition is only checked if 3 million
keys or more have been added to a filter, as this should be a lower
bound for common bits/key settings (< 20).
Recommended remedies include smaller SST file size, using
format_version >= 5 (for new Bloom filter), or using partitioned
filters.
This does not change behavior other than generating warnings for some
constructed filters using the old implementation.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6317
Test Plan:
Example with warning, 15M keys @ 15 bits / key: (working_mem_size_mb is just to stop after building one filter if it's large)
$ ./filter_bench -quick -impl=0 -working_mem_size_mb=1 -bits_per_key=15 -average_keys_per_filter=15000000 2>&1 | grep 'FP rate'
[WARN] [/block_based/filter_policy.cc:292] Using legacy SST/BBT Bloom filter with excessive key count (15.0M @ 15bpk), causing estimated 1.8x higher filter FP rate. Consider using new Bloom with format_version>=5, smaller SST file size, or partitioned filters.
Predicted FP rate %: 0.766702
Average FP rate %: 0.66846
Example without warning (150K keys):
$ ./filter_bench -quick -impl=0 -working_mem_size_mb=1 -bits_per_key=15 -average_keys_per_filter=150000 2>&1 | grep 'FP rate'
Predicted FP rate %: 0.422857
Average FP rate %: 0.379301
$
With more samples at 15 bits/key:
150K keys -> no warning; actual: 0.379% FP rate (baseline)
1M keys -> no warning; actual: 0.396% FP rate, 1.045x
9M keys -> no warning; actual: 0.563% FP rate, 1.485x
10M keys -> warning (1.5x); actual: 0.564% FP rate, 1.488x
15M keys -> warning (1.8x); actual: 0.668% FP rate, 1.76x
25M keys -> warning (2.4x); actual: 0.880% FP rate, 2.32x
At 10 bits/key:
150K keys -> no warning; actual: 1.17% FP rate (baseline)
1M keys -> no warning; actual: 1.16% FP rate
10M keys -> no warning; actual: 1.32% FP rate, 1.13x
25M keys -> no warning; actual: 1.63% FP rate, 1.39x
35M keys -> warning (1.6x); actual: 1.81% FP rate, 1.55x
At 5 bits/key:
150K keys -> no warning; actual: 9.32% FP rate (baseline)
25M keys -> no warning; actual: 9.62% FP rate, 1.03x
200M keys -> no warning; actual: 12.2% FP rate, 1.31x
250M keys -> warning (1.5x); actual: 12.8% FP rate, 1.37x
300M keys -> warning (1.6x); actual: 13.4% FP rate, 1.43x
The reason for the modest inaccuracy at low bits/key is that the assumption of independence between a collision between 32-hash values feeding the filter and an FP in the filter is not quite true for implementations using "simple" logic to compute indices from the stock hash result. There's math on this in my dissertation, but I don't think it's worth the effort just for these extreme cases (> 100 million keys and low-ish bits/key).
Differential Revision: D19471715
Pulled By: pdillinger
fbshipit-source-id: f80c96893a09bf1152630ff0b964e5cdd7e35c68
2020-01-21 06:30:22 +01:00
|
|
|
#ifdef PREDICT_FP_RATE
|
|
|
|
std::cout << "Predicted FP rate %: "
|
|
|
|
<< 100.0 * (weighted_predicted_fp_rate / total_keys_added)
|
|
|
|
<< std::endl;
|
|
|
|
#endif
|
2019-10-25 22:25:28 +02:00
|
|
|
if (!FLAGS_quick && !FLAGS_best_case) {
|
2019-10-08 05:09:27 +02:00
|
|
|
double tolerable_rate = std::pow(2.0, -(bpk - 1.0) / (1.4 + bpk / 50.0));
|
|
|
|
std::cout << "Best possible FP rate %: " << 100.0 * std::pow(2.0, -bpk)
|
|
|
|
<< std::endl;
|
|
|
|
std::cout << "Tolerable FP rate %: " << 100.0 * tolerable_rate << std::endl;
|
|
|
|
|
|
|
|
std::cout << "----------------------------" << std::endl;
|
|
|
|
std::cout << "Verifying..." << std::endl;
|
|
|
|
|
New Bloom filter implementation for full and partitioned filters (#6007)
Summary:
Adds an improved, replacement Bloom filter implementation (FastLocalBloom) for full and partitioned filters in the block-based table. This replacement is faster and more accurate, especially for high bits per key or millions of keys in a single filter.
Speed
The improved speed, at least on recent x86_64, comes from
* Using fastrange instead of modulo (%)
* Using our new hash function (XXH3 preview, added in a previous commit), which is much faster for large keys and only *slightly* slower on keys around 12 bytes if hashing the same size many thousands of times in a row.
* Optimizing the Bloom filter queries with AVX2 SIMD operations. (Added AVX2 to the USE_SSE=1 build.) Careful design was required to support (a) SIMD-optimized queries, (b) compatible non-SIMD code that's simple and efficient, (c) flexible choice of number of probes, and (d) essentially maximized accuracy for a cache-local Bloom filter. Probes are made eight at a time, so any number of probes up to 8 is the same speed, then up to 16, etc.
* Prefetching cache lines when building the filter. Although this optimization could be applied to the old structure as well, it seems to balance out the small added cost of accumulating 64 bit hashes for adding to the filter rather than 32 bit hashes.
Here's nominal speed data from filter_bench (200MB in filters, about 10k keys each, 10 bits filter data / key, 6 probes, avg key size 24 bytes, includes hashing time) on Skylake DE (relatively low clock speed):
$ ./filter_bench -quick -impl=2 -net_includes_hashing # New Bloom filter
Build avg ns/key: 47.7135
Mixed inside/outside queries...
Single filter net ns/op: 26.2825
Random filter net ns/op: 150.459
Average FP rate %: 0.954651
$ ./filter_bench -quick -impl=0 -net_includes_hashing # Old Bloom filter
Build avg ns/key: 47.2245
Mixed inside/outside queries...
Single filter net ns/op: 63.2978
Random filter net ns/op: 188.038
Average FP rate %: 1.13823
Similar build time but dramatically faster query times on hot data (63 ns to 26 ns), and somewhat faster on stale data (188 ns to 150 ns). Performance differences on batched and skewed query loads are between these extremes as expected.
The only other interesting thing about speed is "inside" (query key was added to filter) vs. "outside" (query key was not added to filter) query times. The non-SIMD implementations are substantially slower when most queries are "outside" vs. "inside". This goes against what one might expect or would have observed years ago, as "outside" queries only need about two probes on average, due to short-circuiting, while "inside" always have num_probes (say 6). The problem is probably the nastily unpredictable branch. The SIMD implementation has few branches (very predictable) and has pretty consistent running time regardless of query outcome.
Accuracy
The generally improved accuracy (re: Issue https://github.com/facebook/rocksdb/issues/5857) comes from a better design for probing indices
within a cache line (re: Issue https://github.com/facebook/rocksdb/issues/4120) and improved accuracy for millions of keys in a single filter from using a 64-bit hash function (XXH3p). Design details in code comments.
Accuracy data (generalizes, except old impl gets worse with millions of keys):
Memory bits per key: FP rate percent old impl -> FP rate percent new impl
6: 5.70953 -> 5.69888
8: 2.45766 -> 2.29709
10: 1.13977 -> 0.959254
12: 0.662498 -> 0.411593
16: 0.353023 -> 0.0873754
24: 0.261552 -> 0.0060971
50: 0.225453 -> ~0.00003 (less than 1 in a million queries are FP)
Fixes https://github.com/facebook/rocksdb/issues/5857
Fixes https://github.com/facebook/rocksdb/issues/4120
Unlike the old implementation, this implementation has a fixed cache line size (64 bytes). At 10 bits per key, the accuracy of this new implementation is very close to the old implementation with 128-byte cache line size. If there's sufficient demand, this implementation could be generalized.
Compatibility
Although old releases would see the new structure as corrupt filter data and read the table as if there's no filter, we've decided only to enable the new Bloom filter with new format_version=5. This provides a smooth path for automatic adoption over time, with an option for early opt-in.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6007
Test Plan: filter_bench has been used thoroughly to validate speed, accuracy, and correctness. Unit tests have been carefully updated to exercise new and old implementations, as well as the logic to select an implementation based on context (format_version).
Differential Revision: D18294749
Pulled By: pdillinger
fbshipit-source-id: d44c9db3696e4d0a17caaec47075b7755c262c5f
2019-11-14 01:31:26 +01:00
|
|
|
uint32_t outside_q_per_f =
|
2020-02-21 22:29:41 +01:00
|
|
|
static_cast<uint32_t>(m_queries_ * 1000000 / infos_.size());
|
2019-10-08 05:09:27 +02:00
|
|
|
uint64_t fps = 0;
|
|
|
|
for (uint32_t i = 0; i < infos_.size(); ++i) {
|
|
|
|
FilterInfo &info = infos_[i];
|
|
|
|
for (uint32_t j = 0; j < info.keys_added_; ++j) {
|
2019-10-25 22:25:28 +02:00
|
|
|
if (FLAGS_use_plain_table_bloom) {
|
|
|
|
uint32_t hash = GetSliceHash(kms_[0].Get(info.filter_id_, j));
|
|
|
|
ALWAYS_ASSERT(info.plain_table_bloom_->MayContainHash(hash));
|
|
|
|
} else {
|
|
|
|
ALWAYS_ASSERT(
|
|
|
|
info.reader_->MayMatch(kms_[0].Get(info.filter_id_, j)));
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
for (uint32_t j = 0; j < outside_q_per_f; ++j) {
|
2019-10-25 22:25:28 +02:00
|
|
|
if (FLAGS_use_plain_table_bloom) {
|
|
|
|
uint32_t hash =
|
|
|
|
GetSliceHash(kms_[0].Get(info.filter_id_, j | 0x80000000));
|
|
|
|
fps += info.plain_table_bloom_->MayContainHash(hash);
|
|
|
|
} else {
|
|
|
|
fps += info.reader_->MayMatch(
|
|
|
|
kms_[0].Get(info.filter_id_, j | 0x80000000));
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
std::cout << " No FNs :)" << std::endl;
|
|
|
|
double prelim_rate = double(fps) / outside_q_per_f / infos_.size();
|
|
|
|
std::cout << " Prelim FP rate %: " << (100.0 * prelim_rate) << std::endl;
|
|
|
|
|
|
|
|
if (!FLAGS_allow_bad_fp_rate) {
|
|
|
|
ALWAYS_ASSERT(prelim_rate < tolerable_rate);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
std::cout << "----------------------------" << std::endl;
|
2019-10-25 22:25:28 +02:00
|
|
|
std::cout << "Mixed inside/outside queries..." << std::endl;
|
|
|
|
// 50% each inside and outside
|
|
|
|
uint32_t inside_threshold = UINT32_MAX / 2;
|
2019-10-08 05:09:27 +02:00
|
|
|
for (TestMode tm : testModes) {
|
2019-10-09 04:19:43 +02:00
|
|
|
random_.Seed(FLAGS_seed + 1);
|
2019-10-25 22:25:28 +02:00
|
|
|
double f = RandomQueryTest(inside_threshold, /*dry_run*/ false, tm);
|
2019-10-24 22:07:09 +02:00
|
|
|
random_.Seed(FLAGS_seed + 1);
|
2019-10-25 22:25:28 +02:00
|
|
|
double d = RandomQueryTest(inside_threshold, /*dry_run*/ true, tm);
|
2019-10-24 22:07:09 +02:00
|
|
|
std::cout << " " << TestModeToString(tm) << " net ns/op: " << (f - d)
|
|
|
|
<< std::endl;
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
|
2019-10-25 22:25:28 +02:00
|
|
|
if (!FLAGS_quick) {
|
|
|
|
std::cout << "----------------------------" << std::endl;
|
|
|
|
std::cout << "Inside queries (mostly)..." << std::endl;
|
|
|
|
// Do about 95% inside queries rather than 100% so that branch predictor
|
|
|
|
// can't give itself an artifically crazy advantage.
|
|
|
|
inside_threshold = UINT32_MAX / 20 * 19;
|
|
|
|
for (TestMode tm : testModes) {
|
|
|
|
random_.Seed(FLAGS_seed + 1);
|
|
|
|
double f = RandomQueryTest(inside_threshold, /*dry_run*/ false, tm);
|
|
|
|
random_.Seed(FLAGS_seed + 1);
|
|
|
|
double d = RandomQueryTest(inside_threshold, /*dry_run*/ true, tm);
|
|
|
|
std::cout << " " << TestModeToString(tm) << " net ns/op: " << (f - d)
|
|
|
|
<< std::endl;
|
|
|
|
}
|
|
|
|
|
|
|
|
std::cout << "----------------------------" << std::endl;
|
|
|
|
std::cout << "Outside queries (mostly)..." << std::endl;
|
|
|
|
// Do about 95% outside queries rather than 100% so that branch predictor
|
|
|
|
// can't give itself an artifically crazy advantage.
|
|
|
|
inside_threshold = UINT32_MAX / 20;
|
|
|
|
for (TestMode tm : testModes) {
|
|
|
|
random_.Seed(FLAGS_seed + 2);
|
|
|
|
double f = RandomQueryTest(inside_threshold, /*dry_run*/ false, tm);
|
|
|
|
random_.Seed(FLAGS_seed + 2);
|
|
|
|
double d = RandomQueryTest(inside_threshold, /*dry_run*/ true, tm);
|
|
|
|
std::cout << " " << TestModeToString(tm) << " net ns/op: " << (f - d)
|
|
|
|
<< std::endl;
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
2019-10-24 22:07:09 +02:00
|
|
|
std::cout << fp_rate_report_.str();
|
2019-10-08 05:09:27 +02:00
|
|
|
|
|
|
|
std::cout << "----------------------------" << std::endl;
|
|
|
|
std::cout << "Done. (For more info, run with -legend or -help.)" << std::endl;
|
|
|
|
}
|
|
|
|
|
2019-10-25 22:25:28 +02:00
|
|
|
double FilterBench::RandomQueryTest(uint32_t inside_threshold, bool dry_run,
|
|
|
|
TestMode mode) {
|
2019-10-08 05:09:27 +02:00
|
|
|
for (auto &info : infos_) {
|
|
|
|
info.outside_queries_ = 0;
|
|
|
|
info.false_positives_ = 0;
|
|
|
|
}
|
|
|
|
|
Add new persistent 64-bit hash (#5984)
Summary:
For upcoming new SST filter implementations, we will use a new
64-bit hash function (XXH3 preview, slightly modified). This change
updates hash.{h,cc} for that change, adds unit tests, and out-of-lines
the implementations to keep hash.h as clean/small as possible.
In developing the unit tests, I discovered that the XXH3 preview always
returns zero for the empty string. Zero is problematic for some
algorithms (including an upcoming SST filter implementation) if it
occurs more often than at the "natural" rate, so it should not be
returned from trivial values using trivial seeds. I modified our fork
of XXH3 to return a modest hash of the seed for the empty string.
With hash function details out-of-lines in hash.h, it makes sense to
enable XXH_INLINE_ALL, so that direct calls to XXH64/XXH32/XXH3p
are inlined. To fix array-bounds warnings on some inline calls, I
injected some casts to uintptr_t in xxhash.cc. (Issue reported to Yann.)
Revised: Reverted using XXH_INLINE_ALL for now. Some Facebook
checks are unhappy about #include on xxhash.cc file. I would
fix that by rename to xxhash_cc.h, but to best preserve history I want
to do that in a separate commit (PR) from the uintptr casts.
Also updated filter_bench for this change, improving the performance
predictability of dry run hashing and adding support for 64-bit hash
(for upcoming new SST filter implementations, minor dead code in the
tool for now).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5984
Differential Revision: D18246567
Pulled By: pdillinger
fbshipit-source-id: 6162fbf6381d63c8cc611dd7ec70e1ddc883fbb8
2019-11-01 00:34:51 +01:00
|
|
|
auto dry_run_hash_fn = DryRunNoHash;
|
|
|
|
if (!FLAGS_net_includes_hashing) {
|
|
|
|
if (FLAGS_impl < 2 || FLAGS_use_plain_table_bloom) {
|
|
|
|
dry_run_hash_fn = DryRunHash32;
|
|
|
|
} else {
|
|
|
|
dry_run_hash_fn = DryRunHash64;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-10-09 04:19:43 +02:00
|
|
|
uint32_t num_infos = static_cast<uint32_t>(infos_.size());
|
2019-10-08 05:09:27 +02:00
|
|
|
uint32_t dry_run_hash = 0;
|
2020-02-21 22:29:41 +01:00
|
|
|
uint64_t max_queries = static_cast<uint64_t>(m_queries_ * 1000000 + 0.50);
|
2019-10-08 05:09:27 +02:00
|
|
|
// Some filters may be considered secondary in order to implement skewed
|
|
|
|
// queries. num_primary_filters is the number that are to be treated as
|
|
|
|
// equal, and any remainder will be treated as secondary.
|
2019-10-09 04:19:43 +02:00
|
|
|
uint32_t num_primary_filters = num_infos;
|
2019-10-08 05:09:27 +02:00
|
|
|
// The proportion (when divided by 2^32 - 1) of filter queries going to
|
|
|
|
// the primary filters (default = all). The remainder of queries are
|
|
|
|
// against secondary filters.
|
|
|
|
uint32_t primary_filter_threshold = 0xffffffff;
|
|
|
|
if (mode == kSingleFilter) {
|
|
|
|
// 100% of queries to 1 filter
|
|
|
|
num_primary_filters = 1;
|
|
|
|
} else if (mode == kFiftyOneFilter) {
|
2020-12-12 07:17:08 +01:00
|
|
|
if (num_infos < 50) {
|
|
|
|
return 0.0; // skip
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
// 50% of queries
|
|
|
|
primary_filter_threshold /= 2;
|
|
|
|
// to 1% of filters
|
|
|
|
num_primary_filters = (num_primary_filters + 99) / 100;
|
|
|
|
} else if (mode == kEightyTwentyFilter) {
|
2020-12-12 07:17:08 +01:00
|
|
|
if (num_infos < 5) {
|
|
|
|
return 0.0; // skip
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
// 80% of queries
|
|
|
|
primary_filter_threshold = primary_filter_threshold / 5 * 4;
|
|
|
|
// to 20% of filters
|
|
|
|
num_primary_filters = (num_primary_filters + 4) / 5;
|
2020-12-12 07:17:08 +01:00
|
|
|
} else if (mode == kRandomFilter) {
|
|
|
|
if (num_infos == 1) {
|
|
|
|
return 0.0; // skip
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
2019-10-09 04:19:43 +02:00
|
|
|
uint32_t batch_size = 1;
|
2019-10-24 22:07:09 +02:00
|
|
|
std::unique_ptr<Slice[]> batch_slices;
|
|
|
|
std::unique_ptr<Slice *[]> batch_slice_ptrs;
|
2019-10-08 05:09:27 +02:00
|
|
|
std::unique_ptr<bool[]> batch_results;
|
|
|
|
if (mode == kBatchPrepared || mode == kBatchUnprepared) {
|
2019-10-09 04:19:43 +02:00
|
|
|
batch_size = static_cast<uint32_t>(kms_.size());
|
2019-10-24 22:07:09 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
batch_slices.reset(new Slice[batch_size]);
|
|
|
|
batch_slice_ptrs.reset(new Slice *[batch_size]);
|
|
|
|
batch_results.reset(new bool[batch_size]);
|
|
|
|
for (uint32_t i = 0; i < batch_size; ++i) {
|
|
|
|
batch_results[i] = false;
|
|
|
|
batch_slice_ptrs[i] = &batch_slices[i];
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
|
2021-01-26 07:07:26 +01:00
|
|
|
ROCKSDB_NAMESPACE::StopWatchNano timer(
|
2021-03-15 12:32:24 +01:00
|
|
|
ROCKSDB_NAMESPACE::SystemClock::Default().get(), true);
|
2019-10-08 05:09:27 +02:00
|
|
|
|
|
|
|
for (uint64_t q = 0; q < max_queries; q += batch_size) {
|
2019-10-25 22:25:28 +02:00
|
|
|
bool inside_this_time = random_.Next() <= inside_threshold;
|
|
|
|
|
2019-10-08 05:09:27 +02:00
|
|
|
uint32_t filter_index;
|
2019-10-09 04:19:43 +02:00
|
|
|
if (random_.Next() <= primary_filter_threshold) {
|
|
|
|
filter_index = random_.Uniformish(num_primary_filters);
|
2019-10-08 05:09:27 +02:00
|
|
|
} else {
|
|
|
|
// secondary
|
2019-10-09 04:19:43 +02:00
|
|
|
filter_index = num_primary_filters +
|
|
|
|
random_.Uniformish(num_infos - num_primary_filters);
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
FilterInfo &info = infos_[filter_index];
|
2019-10-09 04:19:43 +02:00
|
|
|
for (uint32_t i = 0; i < batch_size; ++i) {
|
2019-10-25 22:25:28 +02:00
|
|
|
if (inside_this_time) {
|
2019-10-24 22:07:09 +02:00
|
|
|
batch_slices[i] =
|
|
|
|
kms_[i].Get(info.filter_id_, random_.Uniformish(info.keys_added_));
|
2019-10-08 05:09:27 +02:00
|
|
|
} else {
|
2019-10-24 22:07:09 +02:00
|
|
|
batch_slices[i] =
|
|
|
|
kms_[i].Get(info.filter_id_, random_.Uniformish(info.keys_added_) |
|
|
|
|
uint32_t{0x80000000});
|
2019-10-08 05:09:27 +02:00
|
|
|
info.outside_queries_++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
// TODO: implement batched interface to full block reader
|
2019-10-25 22:25:28 +02:00
|
|
|
// TODO: implement batched interface to plain table bloom
|
|
|
|
if (mode == kBatchPrepared && !FLAGS_use_full_block_reader &&
|
|
|
|
!FLAGS_use_plain_table_bloom) {
|
2019-10-09 04:19:43 +02:00
|
|
|
for (uint32_t i = 0; i < batch_size; ++i) {
|
2019-10-08 05:09:27 +02:00
|
|
|
batch_results[i] = false;
|
|
|
|
}
|
2019-10-25 22:25:28 +02:00
|
|
|
if (dry_run) {
|
|
|
|
for (uint32_t i = 0; i < batch_size; ++i) {
|
|
|
|
batch_results[i] = true;
|
Add new persistent 64-bit hash (#5984)
Summary:
For upcoming new SST filter implementations, we will use a new
64-bit hash function (XXH3 preview, slightly modified). This change
updates hash.{h,cc} for that change, adds unit tests, and out-of-lines
the implementations to keep hash.h as clean/small as possible.
In developing the unit tests, I discovered that the XXH3 preview always
returns zero for the empty string. Zero is problematic for some
algorithms (including an upcoming SST filter implementation) if it
occurs more often than at the "natural" rate, so it should not be
returned from trivial values using trivial seeds. I modified our fork
of XXH3 to return a modest hash of the seed for the empty string.
With hash function details out-of-lines in hash.h, it makes sense to
enable XXH_INLINE_ALL, so that direct calls to XXH64/XXH32/XXH3p
are inlined. To fix array-bounds warnings on some inline calls, I
injected some casts to uintptr_t in xxhash.cc. (Issue reported to Yann.)
Revised: Reverted using XXH_INLINE_ALL for now. Some Facebook
checks are unhappy about #include on xxhash.cc file. I would
fix that by rename to xxhash_cc.h, but to best preserve history I want
to do that in a separate commit (PR) from the uintptr casts.
Also updated filter_bench for this change, improving the performance
predictability of dry run hashing and adding support for 64-bit hash
(for upcoming new SST filter implementations, minor dead code in the
tool for now).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5984
Differential Revision: D18246567
Pulled By: pdillinger
fbshipit-source-id: 6162fbf6381d63c8cc611dd7ec70e1ddc883fbb8
2019-11-01 00:34:51 +01:00
|
|
|
dry_run_hash += dry_run_hash_fn(batch_slices[i]);
|
2019-10-25 22:25:28 +02:00
|
|
|
}
|
|
|
|
} else {
|
|
|
|
info.reader_->MayMatch(batch_size, batch_slice_ptrs.get(),
|
|
|
|
batch_results.get());
|
|
|
|
}
|
2019-10-09 04:19:43 +02:00
|
|
|
for (uint32_t i = 0; i < batch_size; ++i) {
|
2019-10-25 22:25:28 +02:00
|
|
|
if (inside_this_time) {
|
2019-10-08 05:09:27 +02:00
|
|
|
ALWAYS_ASSERT(batch_results[i]);
|
|
|
|
} else {
|
|
|
|
info.false_positives_ += batch_results[i];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} else {
|
2019-10-09 04:19:43 +02:00
|
|
|
for (uint32_t i = 0; i < batch_size; ++i) {
|
2019-10-25 22:25:28 +02:00
|
|
|
bool may_match;
|
|
|
|
if (FLAGS_use_plain_table_bloom) {
|
|
|
|
if (dry_run) {
|
Add new persistent 64-bit hash (#5984)
Summary:
For upcoming new SST filter implementations, we will use a new
64-bit hash function (XXH3 preview, slightly modified). This change
updates hash.{h,cc} for that change, adds unit tests, and out-of-lines
the implementations to keep hash.h as clean/small as possible.
In developing the unit tests, I discovered that the XXH3 preview always
returns zero for the empty string. Zero is problematic for some
algorithms (including an upcoming SST filter implementation) if it
occurs more often than at the "natural" rate, so it should not be
returned from trivial values using trivial seeds. I modified our fork
of XXH3 to return a modest hash of the seed for the empty string.
With hash function details out-of-lines in hash.h, it makes sense to
enable XXH_INLINE_ALL, so that direct calls to XXH64/XXH32/XXH3p
are inlined. To fix array-bounds warnings on some inline calls, I
injected some casts to uintptr_t in xxhash.cc. (Issue reported to Yann.)
Revised: Reverted using XXH_INLINE_ALL for now. Some Facebook
checks are unhappy about #include on xxhash.cc file. I would
fix that by rename to xxhash_cc.h, but to best preserve history I want
to do that in a separate commit (PR) from the uintptr casts.
Also updated filter_bench for this change, improving the performance
predictability of dry run hashing and adding support for 64-bit hash
(for upcoming new SST filter implementations, minor dead code in the
tool for now).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5984
Differential Revision: D18246567
Pulled By: pdillinger
fbshipit-source-id: 6162fbf6381d63c8cc611dd7ec70e1ddc883fbb8
2019-11-01 00:34:51 +01:00
|
|
|
dry_run_hash += dry_run_hash_fn(batch_slices[i]);
|
2019-10-25 22:25:28 +02:00
|
|
|
may_match = true;
|
|
|
|
} else {
|
|
|
|
uint32_t hash = GetSliceHash(batch_slices[i]);
|
|
|
|
may_match = info.plain_table_bloom_->MayContainHash(hash);
|
|
|
|
}
|
|
|
|
} else if (FLAGS_use_full_block_reader) {
|
|
|
|
if (dry_run) {
|
Add new persistent 64-bit hash (#5984)
Summary:
For upcoming new SST filter implementations, we will use a new
64-bit hash function (XXH3 preview, slightly modified). This change
updates hash.{h,cc} for that change, adds unit tests, and out-of-lines
the implementations to keep hash.h as clean/small as possible.
In developing the unit tests, I discovered that the XXH3 preview always
returns zero for the empty string. Zero is problematic for some
algorithms (including an upcoming SST filter implementation) if it
occurs more often than at the "natural" rate, so it should not be
returned from trivial values using trivial seeds. I modified our fork
of XXH3 to return a modest hash of the seed for the empty string.
With hash function details out-of-lines in hash.h, it makes sense to
enable XXH_INLINE_ALL, so that direct calls to XXH64/XXH32/XXH3p
are inlined. To fix array-bounds warnings on some inline calls, I
injected some casts to uintptr_t in xxhash.cc. (Issue reported to Yann.)
Revised: Reverted using XXH_INLINE_ALL for now. Some Facebook
checks are unhappy about #include on xxhash.cc file. I would
fix that by rename to xxhash_cc.h, but to best preserve history I want
to do that in a separate commit (PR) from the uintptr casts.
Also updated filter_bench for this change, improving the performance
predictability of dry run hashing and adding support for 64-bit hash
(for upcoming new SST filter implementations, minor dead code in the
tool for now).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5984
Differential Revision: D18246567
Pulled By: pdillinger
fbshipit-source-id: 6162fbf6381d63c8cc611dd7ec70e1ddc883fbb8
2019-11-01 00:34:51 +01:00
|
|
|
dry_run_hash += dry_run_hash_fn(batch_slices[i]);
|
2019-10-25 22:25:28 +02:00
|
|
|
may_match = true;
|
|
|
|
} else {
|
2019-10-08 05:09:27 +02:00
|
|
|
may_match = info.full_block_reader_->KeyMayMatch(
|
2019-10-24 22:07:09 +02:00
|
|
|
batch_slices[i],
|
2019-10-08 05:09:27 +02:00
|
|
|
/*prefix_extractor=*/nullptr,
|
2020-02-20 21:07:53 +01:00
|
|
|
/*block_offset=*/ROCKSDB_NAMESPACE::kNotValid,
|
2019-10-08 05:09:27 +02:00
|
|
|
/*no_io=*/false, /*const_ikey_ptr=*/nullptr,
|
|
|
|
/*get_context=*/nullptr,
|
|
|
|
/*lookup_context=*/nullptr);
|
|
|
|
}
|
2019-10-25 22:25:28 +02:00
|
|
|
} else {
|
|
|
|
if (dry_run) {
|
Add new persistent 64-bit hash (#5984)
Summary:
For upcoming new SST filter implementations, we will use a new
64-bit hash function (XXH3 preview, slightly modified). This change
updates hash.{h,cc} for that change, adds unit tests, and out-of-lines
the implementations to keep hash.h as clean/small as possible.
In developing the unit tests, I discovered that the XXH3 preview always
returns zero for the empty string. Zero is problematic for some
algorithms (including an upcoming SST filter implementation) if it
occurs more often than at the "natural" rate, so it should not be
returned from trivial values using trivial seeds. I modified our fork
of XXH3 to return a modest hash of the seed for the empty string.
With hash function details out-of-lines in hash.h, it makes sense to
enable XXH_INLINE_ALL, so that direct calls to XXH64/XXH32/XXH3p
are inlined. To fix array-bounds warnings on some inline calls, I
injected some casts to uintptr_t in xxhash.cc. (Issue reported to Yann.)
Revised: Reverted using XXH_INLINE_ALL for now. Some Facebook
checks are unhappy about #include on xxhash.cc file. I would
fix that by rename to xxhash_cc.h, but to best preserve history I want
to do that in a separate commit (PR) from the uintptr casts.
Also updated filter_bench for this change, improving the performance
predictability of dry run hashing and adding support for 64-bit hash
(for upcoming new SST filter implementations, minor dead code in the
tool for now).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5984
Differential Revision: D18246567
Pulled By: pdillinger
fbshipit-source-id: 6162fbf6381d63c8cc611dd7ec70e1ddc883fbb8
2019-11-01 00:34:51 +01:00
|
|
|
dry_run_hash += dry_run_hash_fn(batch_slices[i]);
|
2019-10-25 22:25:28 +02:00
|
|
|
may_match = true;
|
2019-10-08 05:09:27 +02:00
|
|
|
} else {
|
2019-10-25 22:25:28 +02:00
|
|
|
may_match = info.reader_->MayMatch(batch_slices[i]);
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
}
|
2019-10-25 22:25:28 +02:00
|
|
|
if (inside_this_time) {
|
|
|
|
ALWAYS_ASSERT(may_match);
|
|
|
|
} else {
|
|
|
|
info.false_positives_ += may_match;
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
uint64_t elapsed_nanos = timer.ElapsedNanos();
|
|
|
|
double ns = double(elapsed_nanos) / max_queries;
|
|
|
|
|
2019-10-24 22:07:09 +02:00
|
|
|
if (!FLAGS_quick) {
|
|
|
|
if (dry_run) {
|
|
|
|
// Printing part of hash prevents dry run components from being optimized
|
|
|
|
// away by compiler
|
|
|
|
std::cout << " Dry run (" << std::hex << (dry_run_hash & 0xfffff)
|
|
|
|
<< std::dec << ") ";
|
|
|
|
} else {
|
|
|
|
std::cout << " Gross filter ";
|
|
|
|
}
|
|
|
|
std::cout << "ns/op: " << ns << std::endl;
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
|
2019-10-25 22:25:28 +02:00
|
|
|
if (!dry_run) {
|
2019-11-20 00:41:56 +01:00
|
|
|
fp_rate_report_.str("");
|
2019-10-08 05:09:27 +02:00
|
|
|
uint64_t q = 0;
|
|
|
|
uint64_t fp = 0;
|
|
|
|
double worst_fp_rate = 0.0;
|
|
|
|
double best_fp_rate = 1.0;
|
|
|
|
for (auto &info : infos_) {
|
|
|
|
q += info.outside_queries_;
|
|
|
|
fp += info.false_positives_;
|
|
|
|
if (info.outside_queries_ > 0) {
|
|
|
|
double fp_rate = double(info.false_positives_) / info.outside_queries_;
|
|
|
|
worst_fp_rate = std::max(worst_fp_rate, fp_rate);
|
|
|
|
best_fp_rate = std::min(best_fp_rate, fp_rate);
|
|
|
|
}
|
|
|
|
}
|
2019-10-24 22:07:09 +02:00
|
|
|
fp_rate_report_ << " Average FP rate %: " << 100.0 * fp / q << std::endl;
|
2019-10-25 22:25:28 +02:00
|
|
|
if (!FLAGS_quick && !FLAGS_best_case) {
|
2019-10-24 22:07:09 +02:00
|
|
|
fp_rate_report_ << " Worst FP rate %: " << 100.0 * worst_fp_rate
|
|
|
|
<< std::endl;
|
|
|
|
fp_rate_report_ << " Best FP rate %: " << 100.0 * best_fp_rate
|
|
|
|
<< std::endl;
|
|
|
|
fp_rate_report_ << " Best possible bits/key: "
|
|
|
|
<< -std::log(double(fp) / q) / std::log(2.0) << std::endl;
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
}
|
2019-10-24 22:07:09 +02:00
|
|
|
return ns;
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
int main(int argc, char **argv) {
|
2020-02-20 21:07:53 +01:00
|
|
|
ROCKSDB_NAMESPACE::port::InstallStackTraceHandler();
|
2019-10-08 05:09:27 +02:00
|
|
|
SetUsageMessage(std::string("\nUSAGE:\n") + std::string(argv[0]) +
|
|
|
|
" [-quick] [OTHER OPTIONS]...");
|
|
|
|
ParseCommandLineFlags(&argc, &argv, true);
|
|
|
|
|
|
|
|
PrintWarnings();
|
|
|
|
|
|
|
|
if (FLAGS_legend) {
|
|
|
|
std::cout
|
|
|
|
<< "Legend:" << std::endl
|
|
|
|
<< " \"Inside\" - key that was added to filter" << std::endl
|
|
|
|
<< " \"Outside\" - key that was not added to filter" << std::endl
|
|
|
|
<< " \"FN\" - false negative query (must not happen)" << std::endl
|
|
|
|
<< " \"FP\" - false positive query (OK at low rate)" << std::endl
|
2019-10-24 22:07:09 +02:00
|
|
|
<< " \"Dry run\" - cost of testing and hashing overhead." << std::endl
|
|
|
|
<< " \"Gross filter\" - cost of filter queries including testing "
|
|
|
|
<< "\n and hashing overhead." << std::endl
|
|
|
|
<< " \"net\" - best estimate of time in filter operation, without "
|
|
|
|
<< "\n testing and hashing overhead (gross filter - dry run)"
|
|
|
|
<< std::endl
|
|
|
|
<< " \"ns/op\" - nanoseconds per operation (key query or add)"
|
|
|
|
<< std::endl
|
2019-10-08 05:09:27 +02:00
|
|
|
<< " \"Single filter\" - essentially minimum cost, assuming filter"
|
|
|
|
<< "\n fits easily in L1 CPU cache." << std::endl
|
|
|
|
<< " \"Batched, prepared\" - several queries at once against a"
|
|
|
|
<< "\n randomly chosen filter, using multi-query interface."
|
|
|
|
<< std::endl
|
|
|
|
<< " \"Batched, unprepared\" - similar, but using serial calls"
|
|
|
|
<< "\n to single query interface." << std::endl
|
|
|
|
<< " \"Random filter\" - a filter is chosen at random as target"
|
|
|
|
<< "\n of each query." << std::endl
|
|
|
|
<< " \"Skewed X% in Y%\" - like \"Random filter\" except Y% of"
|
|
|
|
<< "\n the filters are designated as \"hot\" and receive X%"
|
|
|
|
<< "\n of queries." << std::endl;
|
|
|
|
} else {
|
|
|
|
FilterBench b;
|
2020-02-21 22:29:41 +01:00
|
|
|
for (uint32_t i = 0; i < FLAGS_runs; ++i) {
|
|
|
|
b.Go();
|
|
|
|
FLAGS_seed += 100;
|
|
|
|
b.random_.Seed(FLAGS_seed);
|
|
|
|
}
|
2019-10-08 05:09:27 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2019-10-28 22:10:51 +01:00
|
|
|
#endif // !defined(GFLAGS) || defined(ROCKSDB_LITE)
|