andreacavalli/rocksdb
0
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
efe9d5c823
rocksdb/util/bloom_test.cc
Peter Dillinger a8b3b9a20c Refine Ribbon configuration, improve testing, add Homogeneous (#7879) 
Summary:
This change only affects non-schema-critical aspects of the production candidate Ribbon filter. Specifically, it refines choice of internal configuration parameters based on inputs. The changes are minor enough that the schema tests in bloom_test, some of which depend on this, are unaffected. There are also some minor optimizations and refactorings.

This would be a schema change for "smash" Ribbon, to fix some known issues with small filters, but "smash" Ribbon is not accessible in public APIs. Unit test CompactnessAndBacktrackAndFpRate updated to test small and medium-large filters. Run with --thoroughness=100 or so for much better detection power (not appropriate for continuous regression testing).

Homogenous Ribbon:
This change adds internally a Ribbon filter variant we call Homogeneous Ribbon, in collaboration with Stefan Walzer. The expected "result" value for every key is zero, instead of computed from a hash. Entropy for queries not to be false positives comes from free variables ("overhead") in the solution structure, which are populated pseudorandomly. Construction is slightly faster for not tracking result values, and never fails. Instead, FP rate can jump up whenever and whereever entries are packed too tightly. For small structures, we can choose overhead to make this FP rate jump unlikely, as seen in updated unit test CompactnessAndBacktrackAndFpRate.

Unlike standard Ribbon, Homogeneous Ribbon seems to scale to arbitrary number of keys when accepting an FP rate penalty for small pockets of high FP rate in the structure. For example, 64-bit ribbon with 8 solution columns and 10% allocated space overhead for slots seems to achieve about 10.5% space overhead vs. information-theoretic minimum based on its observed FP rate with expected pockets of degradation. (FP rate is close to 1/256.) If targeting a higher FP rate with fewer solution columns, Homogeneous Ribbon can be even more space efficient, because the penalty from degradation is relatively smaller. If targeting a lower FP rate, Homogeneous Ribbon is less space efficient, as more allocated overhead is needed to keep the FP rate impact of degradation relatively under control. The new OptimizeHomogAtScale tool in ribbon_test helps to find these optimal allocation overheads for different numbers of solution columns. And Ribbon widths, with 128-bit Ribbon apparently cutting space overheads in half vs. 64-bit.

Other misc item specifics:
* Ribbon APIs in util/ribbon_config.h now provide configuration data for not just 5% construction failure rate (95% success), but also 50% and 0.1%.
  * Note that the Ribbon structure does not exhibit "threshold" behavior as standard Xor filter does, so there is a roughly fixed space penalty to cut construction failure rate in half. Thus, there isn't really an "almost sure" setting.
  * Although we can extrapolate settings for large filters, we don't have a good formula for configuring smaller filters (< 2^17 slots or so), and efforts to summarize with a formula have failed. Thus, small data is hard-coded from updated FindOccupancy tool.
* Enhances ApproximateNumEntries for public API Ribbon using more precise data (new API GetNumToAdd), thus a more accurate but not perfect reversal of CalculateSpace. (bloom_test updated to expect the greater precision)
* Move EndianSwapValue from coding.h to coding_lean.h to keep Ribbon code easily transferable from RocksDB
* Add some missing 'const' to member functions
* Small optimization to 128-bit BitParity
* Small refactoring of BandingStorage in ribbon_alg.h to support Homogeneous Ribbon
* CompactnessAndBacktrackAndFpRate now has an "expand" test: on construction failure, a possible alternative to re-seeding hash functions is simply to increase the number of slots (allocated space overhead) and try again with essentially the same hash values. (Start locations will be different roundings of the same scaled hash values--because fastrange not mod.) This seems to be as effective or more effective than re-seeding, as long as we increase the number of slots (m) by roughly m += m/w where w is the Ribbon width. This way, there is effectively an expansion by one slot for each ribbon-width window in the banding. (This approach assumes that getting "bad data" from your hash function is as unlikely as it naturally should be, e.g. no adversary.)
* 32-bit and 16-bit Ribbon configurations are added to ribbon_test for understanding their behavior, e.g. with FindOccupancy. They are not considered useful at this time and not tested with CompactnessAndBacktrackAndFpRate.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/7879

Test Plan: unit test updates included

Reviewed By: jay-zhuang

Differential Revision: D26371245

Pulled By: pdillinger

fbshipit-source-id: da6600d90a3785b99ad17a88b2a3027710b4ea3a
2021-02-26 08:50:42 -08:00

1194 lines
39 KiB
C++
Raw Blame History

// 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).
//
// Copyright (c) 2012 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#ifndef GFLAGS
#include <cstdio>
int main() {
fprintf(stderr, "Please install gflags to run this test... Skipping...\n");
return 0;
}
#else
#include <array>
#include <cmath>
#include <vector>
#include "memory/arena.h"
#include "port/jemalloc_helper.h"
#include "rocksdb/filter_policy.h"
#include "table/block_based/filter_policy_internal.h"
#include "test_util/testharness.h"
#include "test_util/testutil.h"
#include "util/gflags_compat.h"
#include "util/hash.h"
using GFLAGS_NAMESPACE::ParseCommandLineFlags;
DEFINE_int32(bits_per_key, 10, "");
namespace ROCKSDB_NAMESPACE {
static const int kVerbose = 1;
static Slice Key(int i, char* buffer) {
std::string s;
PutFixed32(&s, static_cast<uint32_t>(i));
memcpy(buffer, s.c_str(), sizeof(i));
return Slice(buffer, sizeof(i));
}
static int NextLength(int length) {
if (length < 10) {
length += 1;
} else if (length < 100) {
length += 10;
} else if (length < 1000) {
length += 100;
} else {
length += 1000;
}
return length;
}
class BlockBasedBloomTest : public testing::Test {
private:
std::unique_ptr<const FilterPolicy> policy_;
std::string filter_;
std::vector<std::string> keys_;
public:
BlockBasedBloomTest() { ResetPolicy(); }
void Reset() {
keys_.clear();
filter_.clear();
}
void ResetPolicy(double bits_per_key) {
policy_.reset(new BloomFilterPolicy(bits_per_key,
BloomFilterPolicy::kDeprecatedBlock));
Reset();
}
void ResetPolicy() { ResetPolicy(FLAGS_bits_per_key); }
void Add(const Slice& s) {
keys_.push_back(s.ToString());
}
void Build() {
std::vector<Slice> key_slices;
for (size_t i = 0; i < keys_.size(); i++) {
key_slices.push_back(Slice(keys_[i]));
}
filter_.clear();
policy_->CreateFilter(&key_slices[0], static_cast<int>(key_slices.size()),
&filter_);
keys_.clear();
if (kVerbose >= 2) DumpFilter();
}
size_t FilterSize() const {
return filter_.size();
}
Slice FilterData() const { return Slice(filter_); }
void DumpFilter() {
fprintf(stderr, "F(");
for (size_t i = 0; i+1 < filter_.size(); i++) {
const unsigned int c = static_cast<unsigned int>(filter_[i]);
for (int j = 0; j < 8; j++) {
fprintf(stderr, "%c", (c & (1 <<j)) ? '1' : '.');
}
}
fprintf(stderr, ")\n");
}
bool Matches(const Slice& s) {
if (!keys_.empty()) {
Build();
}
return policy_->KeyMayMatch(s, filter_);
}
double FalsePositiveRate() {
char buffer[sizeof(int)];
int result = 0;
for (int i = 0; i < 10000; i++) {
if (Matches(Key(i + 1000000000, buffer))) {
result++;
}
}
return result / 10000.0;
}
};
TEST_F(BlockBasedBloomTest, EmptyFilter) {
ASSERT_TRUE(! Matches("hello"));
ASSERT_TRUE(! Matches("world"));
}
TEST_F(BlockBasedBloomTest, Small) {
Add("hello");
Add("world");
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
ASSERT_TRUE(! Matches("x"));
ASSERT_TRUE(! Matches("foo"));
}
TEST_F(BlockBasedBloomTest, VaryingLengths) {
char buffer[sizeof(int)];
// Count number of filters that significantly exceed the false positive rate
int mediocre_filters = 0;
int good_filters = 0;
for (int length = 1; length <= 10000; length = NextLength(length)) {
Reset();
for (int i = 0; i < length; i++) {
Add(Key(i, buffer));
}
Build();
ASSERT_LE(FilterSize(), (size_t)((length * 10 / 8) + 40)) << length;
// All added keys must match
for (int i = 0; i < length; i++) {
ASSERT_TRUE(Matches(Key(i, buffer)))
<< "Length " << length << "; key " << i;
}
// Check false positive rate
double rate = FalsePositiveRate();
if (kVerbose >= 1) {
fprintf(stderr, "False positives: %5.2f%% @ length = %6d ; bytes = %6d\n",
rate*100.0, length, static_cast<int>(FilterSize()));
}
ASSERT_LE(rate, 0.02); // Must not be over 2%
if (rate > 0.0125) mediocre_filters++; // Allowed, but not too often
else good_filters++;
}
if (kVerbose >= 1) {
fprintf(stderr, "Filters: %d good, %d mediocre\n",
good_filters, mediocre_filters);
}
ASSERT_LE(mediocre_filters, good_filters/5);
}
// Ensure the implementation doesn't accidentally change in an
// incompatible way
TEST_F(BlockBasedBloomTest, Schema) {
char buffer[sizeof(int)];
ResetPolicy(8); // num_probes = 5
for (int key = 0; key < 87; key++) {
Add(Key(key, buffer));
}
Build();
ASSERT_EQ(BloomHash(FilterData()), 3589896109U);
ResetPolicy(9); // num_probes = 6
for (int key = 0; key < 87; key++) {
Add(Key(key, buffer));
}
Build();
ASSERT_EQ(BloomHash(FilterData()), 969445585U);
ResetPolicy(11); // num_probes = 7
for (int key = 0; key < 87; key++) {
Add(Key(key, buffer));
}
Build();
ASSERT_EQ(BloomHash(FilterData()), 1694458207U);
ResetPolicy(10); // num_probes = 6
for (int key = 0; key < 87; key++) {
Add(Key(key, buffer));
}
Build();
ASSERT_EQ(BloomHash(FilterData()), 2373646410U);
ResetPolicy(10);
for (int key = /*CHANGED*/ 1; key < 87; key++) {
Add(Key(key, buffer));
}
Build();
ASSERT_EQ(BloomHash(FilterData()), 1908442116U);
ResetPolicy(10);
for (int key = 1; key < /*CHANGED*/ 88; key++) {
Add(Key(key, buffer));
}
Build();
ASSERT_EQ(BloomHash(FilterData()), 3057004015U);
// With new fractional bits_per_key, check that we are rounding to
// whole bits per key for old Bloom filters.
ResetPolicy(9.5); // Treated as 10
for (int key = 1; key < 88; key++) {
Add(Key(key, buffer));
}
Build();
ASSERT_EQ(BloomHash(FilterData()), /*SAME*/ 3057004015U);
ResetPolicy(10.499); // Treated as 10
for (int key = 1; key < 88; key++) {
Add(Key(key, buffer));
}
Build();
ASSERT_EQ(BloomHash(FilterData()), /*SAME*/ 3057004015U);
ResetPolicy();
}
// Different bits-per-byte
class FullBloomTest : public testing::TestWithParam<BloomFilterPolicy::Mode> {
protected:
BlockBasedTableOptions table_options_;
private:
std::shared_ptr<const FilterPolicy>& policy_;
std::unique_ptr<FilterBitsBuilder> bits_builder_;
std::unique_ptr<FilterBitsReader> bits_reader_;
std::unique_ptr<const char[]> buf_;
size_t filter_size_;
public:
FullBloomTest() : policy_(table_options_.filter_policy), filter_size_(0) {
ResetPolicy();
}
BuiltinFilterBitsBuilder* GetBuiltinFilterBitsBuilder() {
// Throws on bad cast
return &dynamic_cast<BuiltinFilterBitsBuilder&>(*bits_builder_);
}
const BloomFilterPolicy* GetBloomFilterPolicy() {
// Throws on bad cast
return &dynamic_cast<const BloomFilterPolicy&>(*policy_);
}
void Reset() {
bits_builder_.reset(BloomFilterPolicy::GetBuilderFromContext(
FilterBuildingContext(table_options_)));
bits_reader_.reset(nullptr);
buf_.reset(nullptr);
filter_size_ = 0;
}
void ResetPolicy(double bits_per_key) {
policy_.reset(new BloomFilterPolicy(bits_per_key, GetParam()));
Reset();
}
void ResetPolicy() { ResetPolicy(FLAGS_bits_per_key); }
void Add(const Slice& s) {
bits_builder_->AddKey(s);
}
void OpenRaw(const Slice& s) {
bits_reader_.reset(policy_->GetFilterBitsReader(s));
}
void Build() {
Slice filter = bits_builder_->Finish(&buf_);
bits_reader_.reset(policy_->GetFilterBitsReader(filter));
filter_size_ = filter.size();
}
size_t FilterSize() const {
return filter_size_;
}
Slice FilterData() { return Slice(buf_.get(), filter_size_); }
int GetNumProbesFromFilterData() {
assert(filter_size_ >= 5);
int8_t raw_num_probes = static_cast<int8_t>(buf_.get()[filter_size_ - 5]);
if (raw_num_probes == -1) { // New bloom filter marker
return static_cast<uint8_t>(buf_.get()[filter_size_ - 3]);
} else {
return raw_num_probes;
}
}
int GetRibbonSeedFromFilterData() {
assert(filter_size_ >= 5);
// Check for ribbon marker
assert(-2 == static_cast<int8_t>(buf_.get()[filter_size_ - 5]));
return static_cast<uint8_t>(buf_.get()[filter_size_ - 4]);
}
bool Matches(const Slice& s) {
if (bits_reader_ == nullptr) {
Build();
}
return bits_reader_->MayMatch(s);
}
// Provides a kind of fingerprint on the Bloom filter's
// behavior, for reasonbly high FP rates.
uint64_t PackedMatches() {
char buffer[sizeof(int)];
uint64_t result = 0;
for (int i = 0; i < 64; i++) {
if (Matches(Key(i + 12345, buffer))) {
result |= uint64_t{1} << i;
}
}
return result;
}
// Provides a kind of fingerprint on the Bloom filter's
// behavior, for lower FP rates.
std::string FirstFPs(int count) {
char buffer[sizeof(int)];
std::string rv;
int fp_count = 0;
for (int i = 0; i < 1000000; i++) {
// Pack four match booleans into each hexadecimal digit
if (Matches(Key(i + 1000000, buffer))) {
++fp_count;
rv += std::to_string(i);
if (fp_count == count) {
break;
}
rv += ',';
}
}
return rv;
}
double FalsePositiveRate() {
char buffer[sizeof(int)];
int result = 0;
for (int i = 0; i < 10000; i++) {
if (Matches(Key(i + 1000000000, buffer))) {
result++;
}
}
return result / 10000.0;
}
};
TEST_P(FullBloomTest, FilterSize) {
// In addition to checking the consistency of space computation, we are
// checking that denoted and computed doubles are interpreted as expected
// as bits_per_key values.
bool some_computed_less_than_denoted = false;
// Note: enforced minimum is 1 bit per key (1000 millibits), and enforced
// maximum is 100 bits per key (100000 millibits).
for (auto bpk :
std::vector<std::pair<double, int> >{{-HUGE_VAL, 1000},
{-INFINITY, 1000},
{0.0, 1000},
{1.234, 1234},
{3.456, 3456},
{9.5, 9500},
{10.0, 10000},
{10.499, 10499},
{21.345, 21345},
{99.999, 99999},
{1234.0, 100000},
{HUGE_VAL, 100000},
{INFINITY, 100000},
{NAN, 100000}}) {
ResetPolicy(bpk.first);
auto bfp = GetBloomFilterPolicy();
EXPECT_EQ(bpk.second, bfp->GetMillibitsPerKey());
EXPECT_EQ((bpk.second + 500) / 1000, bfp->GetWholeBitsPerKey());
double computed = bpk.first;
// This transforms e.g. 9.5 -> 9.499999999999998, which we still
// round to 10 for whole bits per key.
computed += 0.5;
computed /= 1234567.0;
computed *= 1234567.0;
computed -= 0.5;
some_computed_less_than_denoted |= (computed < bpk.first);
ResetPolicy(computed);
bfp = GetBloomFilterPolicy();
EXPECT_EQ(bpk.second, bfp->GetMillibitsPerKey());
EXPECT_EQ((bpk.second + 500) / 1000, bfp->GetWholeBitsPerKey());
auto bits_builder = GetBuiltinFilterBitsBuilder();
size_t n = 1;
size_t space = 0;
for (; n < 1000000; n += 1 + n / 1000) {
// Ensure consistency between CalculateSpace and ApproximateNumEntries
space = bits_builder->CalculateSpace(n);
size_t n2 = bits_builder->ApproximateNumEntries(space);
EXPECT_GE(n2, n);
size_t space2 = bits_builder->CalculateSpace(n2);
if (n > 12000 && GetParam() == BloomFilterPolicy::kStandard128Ribbon) {
// TODO(peterd): better approximation?
EXPECT_GE(space2, space);
EXPECT_LE(space2 * 0.998, space * 1.0);
} else {
EXPECT_EQ(space2, space);
}
}
// Until size_t overflow
for (; n < (n + n / 3); n += n / 3) {
// Ensure space computation is not overflowing; capped is OK
size_t space2 = bits_builder->CalculateSpace(n);
EXPECT_GE(space2, space);
space = space2;
}
}
// Check that the compiler hasn't optimized our computation into nothing
EXPECT_TRUE(some_computed_less_than_denoted);
ResetPolicy();
}
TEST_P(FullBloomTest, FullEmptyFilter) {
// Empty filter is not match, at this level
ASSERT_TRUE(!Matches("hello"));
ASSERT_TRUE(!Matches("world"));
}
TEST_P(FullBloomTest, FullSmall) {
Add("hello");
Add("world");
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
ASSERT_TRUE(!Matches("x"));
ASSERT_TRUE(!Matches("foo"));
}
TEST_P(FullBloomTest, FullVaryingLengths) {
char buffer[sizeof(int)];
// Count number of filters that significantly exceed the false positive rate
int mediocre_filters = 0;
int good_filters = 0;
for (int length = 1; length <= 10000; length = NextLength(length)) {
Reset();
for (int i = 0; i < length; i++) {
Add(Key(i, buffer));
}
Build();
EXPECT_LE(FilterSize(),
(size_t)((length * 10 / 8) + CACHE_LINE_SIZE * 2 + 5));
// All added keys must match
for (int i = 0; i < length; i++) {
ASSERT_TRUE(Matches(Key(i, buffer)))
<< "Length " << length << "; key " << i;
}
// Check false positive rate
double rate = FalsePositiveRate();
if (kVerbose >= 1) {
fprintf(stderr, "False positives: %5.2f%% @ length = %6d ; bytes = %6d\n",
rate*100.0, length, static_cast<int>(FilterSize()));
}
EXPECT_LE(rate, 0.02); // Must not be over 2%
if (rate > 0.0125)
mediocre_filters++; // Allowed, but not too often
else
good_filters++;
}
if (kVerbose >= 1) {
fprintf(stderr, "Filters: %d good, %d mediocre\n",
good_filters, mediocre_filters);
}
EXPECT_LE(mediocre_filters, good_filters / 5);
}
TEST_P(FullBloomTest, OptimizeForMemory) {
char buffer[sizeof(int)];
for (bool offm : {true, false}) {
table_options_.optimize_filters_for_memory = offm;
ResetPolicy();
Random32 rnd(12345);
uint64_t total_size = 0;
uint64_t total_mem = 0;
int64_t total_keys = 0;
double total_fp_rate = 0;
constexpr int nfilters = 100;
for (int i = 0; i < nfilters; ++i) {
int nkeys = static_cast<int>(rnd.Uniformish(10000)) + 100;
Reset();
for (int j = 0; j < nkeys; ++j) {
Add(Key(j, buffer));
}
Build();
size_t size = FilterData().size();
total_size += size;
// optimize_filters_for_memory currently depends on malloc_usable_size
// but we run the rest of the test to ensure no bad behavior without it.
#ifdef ROCKSDB_MALLOC_USABLE_SIZE
size = malloc_usable_size(const_cast<char*>(FilterData().data()));
#endif // ROCKSDB_MALLOC_USABLE_SIZE
total_mem += size;
total_keys += nkeys;
total_fp_rate += FalsePositiveRate();
}
EXPECT_LE(total_fp_rate / double{nfilters}, 0.011);
EXPECT_GE(total_fp_rate / double{nfilters}, 0.008);
int64_t ex_min_total_size = int64_t{FLAGS_bits_per_key} * total_keys / 8;
if (GetParam() == BloomFilterPolicy::kStandard128Ribbon) {
// ~ 30% savings vs. Bloom filter
ex_min_total_size = 7 * ex_min_total_size / 10;
}
EXPECT_GE(static_cast<int64_t>(total_size), ex_min_total_size);
int64_t blocked_bloom_overhead = nfilters * (CACHE_LINE_SIZE + 5);
if (GetParam() == BloomFilterPolicy::kLegacyBloom) {
// this config can add extra cache line to make odd number
blocked_bloom_overhead += nfilters * CACHE_LINE_SIZE;
}
EXPECT_GE(total_mem, total_size);
// optimize_filters_for_memory not implemented with legacy Bloom
if (offm && GetParam() != BloomFilterPolicy::kLegacyBloom) {
// This value can include a small extra penalty for kExtraPadding
fprintf(stderr, "Internal fragmentation (optimized): %g%%\n",
(total_mem - total_size) * 100.0 / total_size);
// Less than 1% internal fragmentation
EXPECT_LE(total_mem, total_size * 101 / 100);
// Up to 2% storage penalty
EXPECT_LE(static_cast<int64_t>(total_size),
ex_min_total_size * 102 / 100 + blocked_bloom_overhead);
} else {
fprintf(stderr, "Internal fragmentation (not optimized): %g%%\n",
(total_mem - total_size) * 100.0 / total_size);
// TODO: add control checks for more allocators?
#ifdef ROCKSDB_JEMALLOC
fprintf(stderr, "Jemalloc detected? %d\n", HasJemalloc());
if (HasJemalloc()) {
#ifdef ROCKSDB_MALLOC_USABLE_SIZE
// More than 5% internal fragmentation
EXPECT_GE(total_mem, total_size * 105 / 100);
#endif // ROCKSDB_MALLOC_USABLE_SIZE
}
#endif // ROCKSDB_JEMALLOC
// No storage penalty, just usual overhead
EXPECT_LE(static_cast<int64_t>(total_size),
ex_min_total_size + blocked_bloom_overhead);
}
}
}
namespace {
inline uint32_t SelectByCacheLineSize(uint32_t for64, uint32_t for128,
uint32_t for256) {
(void)for64;
(void)for128;
(void)for256;
#if CACHE_LINE_SIZE == 64
return for64;
#elif CACHE_LINE_SIZE == 128
return for128;
#elif CACHE_LINE_SIZE == 256
return for256;
#else
#error "CACHE_LINE_SIZE unknown or unrecognized"
#endif
}
} // namespace
// Ensure the implementation doesn't accidentally change in an
// incompatible way. This test doesn't check the reading side
// (FirstFPs/PackedMatches) for LegacyBloom because it requires the
// ability to read filters generated using other cache line sizes.
// See RawSchema.
TEST_P(FullBloomTest, Schema) {
#define EXPECT_EQ_Bloom(a, b) \
{ \
if (GetParam() != BloomFilterPolicy::kStandard128Ribbon) { \
EXPECT_EQ(a, b); \
} \
}
#define EXPECT_EQ_Ribbon(a, b) \
{ \
if (GetParam() == BloomFilterPolicy::kStandard128Ribbon) { \
EXPECT_EQ(a, b); \
} \
}
#define EXPECT_EQ_FastBloom(a, b) \
{ \
if (GetParam() == BloomFilterPolicy::kFastLocalBloom) { \
EXPECT_EQ(a, b); \
} \
}
#define EXPECT_EQ_LegacyBloom(a, b) \
{ \
if (GetParam() == BloomFilterPolicy::kLegacyBloom) { \
EXPECT_EQ(a, b); \
} \
}
#define EXPECT_EQ_NotLegacy(a, b) \
{ \
if (GetParam() != BloomFilterPolicy::kLegacyBloom) { \
EXPECT_EQ(a, b); \
} \
}
char buffer[sizeof(int)];
// First do a small number of keys, where Ribbon config will fall back on
// fast Bloom filter and generate the same data
ResetPolicy(5); // num_probes = 3
for (int key = 0; key < 87; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ(GetNumProbesFromFilterData(), 3);
EXPECT_EQ_NotLegacy(BloomHash(FilterData()), 4130687756U);
EXPECT_EQ_NotLegacy("31,38,40,43,61,83,86,112,125,131", FirstFPs(10));
// Now use enough keys so that changing bits / key by 1 is guaranteed to
// change number of allocated cache lines. So keys > max cache line bits.
// Note that the first attempted Ribbon seed is determined by the hash
// of the first key added (for pseudorandomness in practice, determinism in
// testing)
ResetPolicy(2); // num_probes = 1
for (int key = 0; key < 2087; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_Bloom(GetNumProbesFromFilterData(), 1);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 61);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(1567096579, 1964771444, 2659542661U));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 3817481309U);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 1705851228);
EXPECT_EQ_FastBloom("11,13,17,25,29,30,35,37,45,53", FirstFPs(10));
EXPECT_EQ_Ribbon("3,8,10,17,19,20,23,28,31,32", FirstFPs(10));
ResetPolicy(3); // num_probes = 2
for (int key = 0; key < 2087; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_Bloom(GetNumProbesFromFilterData(), 2);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 61);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(2707206547U, 2571983456U, 218344685));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 2807269961U);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 1095342358);
EXPECT_EQ_FastBloom("4,15,17,24,27,28,29,53,63,70", FirstFPs(10));
EXPECT_EQ_Ribbon("3,17,20,28,32,33,36,43,49,54", FirstFPs(10));
ResetPolicy(5); // num_probes = 3
for (int key = 0; key < 2087; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_Bloom(GetNumProbesFromFilterData(), 3);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 61);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(515748486, 94611728, 2436112214U));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 204628445);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 3971337699U);
EXPECT_EQ_FastBloom("15,24,29,39,53,87,89,100,103,104", FirstFPs(10));
EXPECT_EQ_Ribbon("3,33,36,43,67,70,76,78,84,102", FirstFPs(10));
ResetPolicy(8); // num_probes = 5
for (int key = 0; key < 2087; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_Bloom(GetNumProbesFromFilterData(), 5);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 61);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(1302145999, 2811644657U, 756553699));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 355564975);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 3651449053U);
EXPECT_EQ_FastBloom("16,60,66,126,220,238,244,256,265,287", FirstFPs(10));
EXPECT_EQ_Ribbon("33,187,203,296,300,322,411,419,547,582", FirstFPs(10));
ResetPolicy(9); // num_probes = 6
for (int key = 0; key < 2087; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_Bloom(GetNumProbesFromFilterData(), 6);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 61);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(2092755149, 661139132, 1182970461));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 2137566013U);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 1005676675);
EXPECT_EQ_FastBloom("156,367,791,872,945,1015,1139,1159,1265", FirstFPs(9));
EXPECT_EQ_Ribbon("33,187,203,296,411,419,604,612,615,619", FirstFPs(10));
ResetPolicy(11); // num_probes = 7
for (int key = 0; key < 2087; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_Bloom(GetNumProbesFromFilterData(), 7);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 61);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(3755609649U, 1812694762, 1449142939));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 2561502687U);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 3129900846U);
EXPECT_EQ_FastBloom("34,74,130,236,643,882,962,1015,1035,1110", FirstFPs(10));
EXPECT_EQ_Ribbon("411,419,623,665,727,794,955,1052,1323,1330", FirstFPs(10));
// This used to be 9 probes, but 8 is a better choice for speed,
// especially with SIMD groups of 8 probes, with essentially no
// change in FP rate.
// FP rate @ 9 probes, old Bloom: 0.4321%
// FP rate @ 9 probes, new Bloom: 0.1846%
// FP rate @ 8 probes, new Bloom: 0.1843%
ResetPolicy(14); // num_probes = 8 (new), 9 (old)
for (int key = 0; key < 2087; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_LegacyBloom(GetNumProbesFromFilterData(), 9);
EXPECT_EQ_FastBloom(GetNumProbesFromFilterData(), 8);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 61);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(178861123, 379087593, 2574136516U));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 3709876890U);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 1855638875);
EXPECT_EQ_FastBloom("130,240,522,565,989,2002,2526,3147,3543", FirstFPs(9));
EXPECT_EQ_Ribbon("665,727,1323,1755,3866,4232,4442,4492,4736", FirstFPs(9));
// This used to be 11 probes, but 9 is a better choice for speed
// AND accuracy.
// FP rate @ 11 probes, old Bloom: 0.3571%
// FP rate @ 11 probes, new Bloom: 0.0884%
// FP rate @ 9 probes, new Bloom: 0.0843%
ResetPolicy(16); // num_probes = 9 (new), 11 (old)
for (int key = 0; key < 2087; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_LegacyBloom(GetNumProbesFromFilterData(), 11);
EXPECT_EQ_FastBloom(GetNumProbesFromFilterData(), 9);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 61);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(1129406313, 3049154394U, 1727750964));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 1087138490);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 459379967);
EXPECT_EQ_FastBloom("3299,3611,3916,6620,7822,8079,8482,8942", FirstFPs(8));
EXPECT_EQ_Ribbon("727,1323,1755,4442,4736,5386,6974,7154,8222", FirstFPs(9));
ResetPolicy(10); // num_probes = 6, but different memory ratio vs. 9
for (int key = 0; key < 2087; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_Bloom(GetNumProbesFromFilterData(), 6);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 61);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(1478976371, 2910591341U, 1182970461));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 2498541272U);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 1273231667);
EXPECT_EQ_FastBloom("16,126,133,422,466,472,813,1002,1035", FirstFPs(9));
EXPECT_EQ_Ribbon("296,411,419,612,619,623,630,665,686,727", FirstFPs(10));
ResetPolicy(10);
for (int key = /*CHANGED*/ 1; key < 2087; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_Bloom(GetNumProbesFromFilterData(), 6);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), /*CHANGED*/ 184);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(4205696321U, 1132081253U, 2385981855U));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 2058382345U);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 3007790572U);
EXPECT_EQ_FastBloom("16,126,133,422,466,472,813,1002,1035", FirstFPs(9));
EXPECT_EQ_Ribbon("33,152,383,497,589,633,737,781,911,990", FirstFPs(10));
ResetPolicy(10);
for (int key = 1; key < /*CHANGED*/ 2088; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_Bloom(GetNumProbesFromFilterData(), 6);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 184);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
SelectByCacheLineSize(2885052954U, 769447944, 4175124908U));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 23699164);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 1942323379);
EXPECT_EQ_FastBloom("16,126,133,422,466,472,813,1002,1035", FirstFPs(9));
EXPECT_EQ_Ribbon("33,95,360,589,737,911,990,1048,1081,1414", FirstFPs(10));
// With new fractional bits_per_key, check that we are rounding to
// whole bits per key for old Bloom filters but fractional for
// new Bloom filter.
ResetPolicy(9.5);
for (int key = 1; key < 2088; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_Bloom(GetNumProbesFromFilterData(), 6);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 184);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
/*SAME*/ SelectByCacheLineSize(2885052954U, 769447944, 4175124908U));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 3166884174U);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 1148258663);
EXPECT_EQ_FastBloom("126,156,367,444,458,791,813,976,1015", FirstFPs(9));
EXPECT_EQ_Ribbon("33,54,95,360,589,693,737,911,990,1048", FirstFPs(10));
ResetPolicy(10.499);
for (int key = 1; key < 2088; key++) {
Add(Key(key, buffer));
}
Build();
EXPECT_EQ_LegacyBloom(GetNumProbesFromFilterData(), 6);
EXPECT_EQ_FastBloom(GetNumProbesFromFilterData(), 7);
EXPECT_EQ_Ribbon(GetRibbonSeedFromFilterData(), 184);
EXPECT_EQ_LegacyBloom(
BloomHash(FilterData()),
/*SAME*/ SelectByCacheLineSize(2885052954U, 769447944, 4175124908U));
EXPECT_EQ_FastBloom(BloomHash(FilterData()), 4098502778U);
EXPECT_EQ_Ribbon(BloomHash(FilterData()), 792138188);
EXPECT_EQ_FastBloom("16,236,240,472,1015,1045,1111,1409,1465", FirstFPs(9));
EXPECT_EQ_Ribbon("33,95,360,589,737,990,1048,1081,1414,1643", FirstFPs(10));
ResetPolicy();
}
// A helper class for testing custom or corrupt filter bits as read by
// built-in FilterBitsReaders.
struct RawFilterTester {
// Buffer, from which we always return a tail Slice, so the
// last five bytes are always the metadata bytes.
std::array<char, 3000> data_;
// Points five bytes from the end
char* metadata_ptr_;
RawFilterTester() : metadata_ptr_(&*(data_.end() - 5)) {}
Slice ResetNoFill(uint32_t len_without_metadata, uint32_t num_lines,
uint32_t num_probes) {
metadata_ptr_[0] = static_cast<char>(num_probes);
EncodeFixed32(metadata_ptr_ + 1, num_lines);
uint32_t len = len_without_metadata + /*metadata*/ 5;
assert(len <= data_.size());
return Slice(metadata_ptr_ - len_without_metadata, len);
}
Slice Reset(uint32_t len_without_metadata, uint32_t num_lines,
uint32_t num_probes, bool fill_ones) {
data_.fill(fill_ones ? 0xff : 0);
return ResetNoFill(len_without_metadata, num_lines, num_probes);
}
Slice ResetWeirdFill(uint32_t len_without_metadata, uint32_t num_lines,
uint32_t num_probes) {
for (uint32_t i = 0; i < data_.size(); ++i) {
data_[i] = static_cast<char>(0x7b7b >> (i % 7));
}
return ResetNoFill(len_without_metadata, num_lines, num_probes);
}
};
TEST_P(FullBloomTest, RawSchema) {
RawFilterTester cft;
// Legacy Bloom configurations
// Two probes, about 3/4 bits set: ~50% "FP" rate
// One 256-byte cache line.
OpenRaw(cft.ResetWeirdFill(256, 1, 2));
EXPECT_EQ(uint64_t{11384799501900898790U}, PackedMatches());
// Two 128-byte cache lines.
OpenRaw(cft.ResetWeirdFill(256, 2, 2));
EXPECT_EQ(uint64_t{10157853359773492589U}, PackedMatches());
// Four 64-byte cache lines.
OpenRaw(cft.ResetWeirdFill(256, 4, 2));
EXPECT_EQ(uint64_t{7123594913907464682U}, PackedMatches());
// Fast local Bloom configurations (marker 255 -> -1)
// Two probes, about 3/4 bits set: ~50% "FP" rate
// Four 64-byte cache lines.
OpenRaw(cft.ResetWeirdFill(256, 2U << 8, 255));
EXPECT_EQ(uint64_t{9957045189927952471U}, PackedMatches());
// Ribbon configurations (marker 254 -> -2)
// Even though the builder never builds configurations this
// small (preferring Bloom), we can test that the configuration
// can be read, for possible future-proofing.
// 256 slots, one result column = 32 bytes (2 blocks, seed 0)
// ~50% FP rate:
// 0b0101010111110101010000110000011011011111100100001110010011101010
OpenRaw(cft.ResetWeirdFill(32, 2U << 8, 254));
EXPECT_EQ(uint64_t{6193930559317665002U}, PackedMatches());
// 256 slots, three-to-four result columns = 112 bytes
// ~ 1 in 10 FP rate:
// 0b0000000000100000000000000000000001000001000000010000101000000000
OpenRaw(cft.ResetWeirdFill(112, 2U << 8, 254));
EXPECT_EQ(uint64_t{9007200345328128U}, PackedMatches());
}
TEST_P(FullBloomTest, CorruptFilters) {
RawFilterTester cft;
for (bool fill : {false, true}) {
// Legacy Bloom configurations
// Good filter bits - returns same as fill
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 1, 6, fill));
ASSERT_EQ(fill, Matches("hello"));
ASSERT_EQ(fill, Matches("world"));
// Good filter bits - returns same as fill
OpenRaw(cft.Reset(CACHE_LINE_SIZE * 3, 3, 6, fill));
ASSERT_EQ(fill, Matches("hello"));
ASSERT_EQ(fill, Matches("world"));
// Good filter bits - returns same as fill
// 256 is unusual but legal cache line size
OpenRaw(cft.Reset(256 * 3, 3, 6, fill));
ASSERT_EQ(fill, Matches("hello"));
ASSERT_EQ(fill, Matches("world"));
// Good filter bits - returns same as fill
// 30 should be max num_probes
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 1, 30, fill));
ASSERT_EQ(fill, Matches("hello"));
ASSERT_EQ(fill, Matches("world"));
// Good filter bits - returns same as fill
// 1 should be min num_probes
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 1, 1, fill));
ASSERT_EQ(fill, Matches("hello"));
ASSERT_EQ(fill, Matches("world"));
// Type 1 trivial filter bits - returns true as if FP by zero probes
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 1, 0, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
// Type 2 trivial filter bits - returns false as if built from zero keys
OpenRaw(cft.Reset(0, 0, 6, fill));
ASSERT_FALSE(Matches("hello"));
ASSERT_FALSE(Matches("world"));
// Type 2 trivial filter bits - returns false as if built from zero keys
OpenRaw(cft.Reset(0, 37, 6, fill));
ASSERT_FALSE(Matches("hello"));
ASSERT_FALSE(Matches("world"));
// Type 2 trivial filter bits - returns false as 0 size trumps 0 probes
OpenRaw(cft.Reset(0, 0, 0, fill));
ASSERT_FALSE(Matches("hello"));
ASSERT_FALSE(Matches("world"));
// Bad filter bits - returns true for safety
// No solution to 0 * x == CACHE_LINE_SIZE
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 0, 6, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
// Bad filter bits - returns true for safety
// Can't have 3 * x == 4 for integer x
OpenRaw(cft.Reset(4, 3, 6, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
// Bad filter bits - returns true for safety
// 97 bytes is not a power of two, so not a legal cache line size
OpenRaw(cft.Reset(97 * 3, 3, 6, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
// Bad filter bits - returns true for safety
// 65 bytes is not a power of two, so not a legal cache line size
OpenRaw(cft.Reset(65 * 3, 3, 6, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
// Bad filter bits - returns false as if built from zero keys
// < 5 bytes overall means missing even metadata
OpenRaw(cft.Reset(static_cast<uint32_t>(-1), 3, 6, fill));
ASSERT_FALSE(Matches("hello"));
ASSERT_FALSE(Matches("world"));
OpenRaw(cft.Reset(static_cast<uint32_t>(-5), 3, 6, fill));
ASSERT_FALSE(Matches("hello"));
ASSERT_FALSE(Matches("world"));
// Dubious filter bits - returns same as fill (for now)
// 31 is not a useful num_probes, nor generated by RocksDB unless directly
// using filter bits API without BloomFilterPolicy.
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 1, 31, fill));
ASSERT_EQ(fill, Matches("hello"));
ASSERT_EQ(fill, Matches("world"));
// Dubious filter bits - returns same as fill (for now)
// Similar, with 127, largest positive char
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 1, 127, fill));
ASSERT_EQ(fill, Matches("hello"));
ASSERT_EQ(fill, Matches("world"));
// Dubious filter bits - returns true (for now)
// num_probes set to 128 / -128, lowest negative char
// NB: Bug in implementation interprets this as negative and has same
// effect as zero probes, but effectively reserves negative char values
// for future use.
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 1, 128, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
// Dubious filter bits - returns true (for now)
// Similar, with 253 / -3
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 1, 253, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
// #########################################################
// Fast local Bloom configurations (marker 255 -> -1)
// Good config with six probes
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 6U << 8, 255, fill));
ASSERT_EQ(fill, Matches("hello"));
ASSERT_EQ(fill, Matches("world"));
// Becomes bad/reserved config (always true) if any other byte set
OpenRaw(cft.Reset(CACHE_LINE_SIZE, (6U << 8) | 1U, 255, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
OpenRaw(cft.Reset(CACHE_LINE_SIZE, (6U << 8) | (1U << 16), 255, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
OpenRaw(cft.Reset(CACHE_LINE_SIZE, (6U << 8) | (1U << 24), 255, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
// Good config, max 30 probes
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 30U << 8, 255, fill));
ASSERT_EQ(fill, Matches("hello"));
ASSERT_EQ(fill, Matches("world"));
// Bad/reserved config (always true) if more than 30
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 31U << 8, 255, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 33U << 8, 255, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 66U << 8, 255, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
OpenRaw(cft.Reset(CACHE_LINE_SIZE, 130U << 8, 255, fill));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
}
// #########################################################
// Ribbon configurations (marker 254 -> -2)
// ("fill" doesn't work to detect good configurations, we just
// have to rely on TN probability)
// Good: 2 blocks * 16 bytes / segment * 4 columns = 128 bytes
// seed = 123
OpenRaw(cft.Reset(128, (2U << 8) + 123U, 254, false));
ASSERT_FALSE(Matches("hello"));
ASSERT_FALSE(Matches("world"));
// Good: 2 blocks * 16 bytes / segment * 8 columns = 256 bytes
OpenRaw(cft.Reset(256, (2U << 8) + 123U, 254, false));
ASSERT_FALSE(Matches("hello"));
ASSERT_FALSE(Matches("world"));
// Surprisingly OK: 5000 blocks (640,000 slots) in only 1024 bits
// -> average close to 0 columns
OpenRaw(cft.Reset(128, (5000U << 8) + 123U, 254, false));
// *Almost* all FPs
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
// Need many queries to find a "true negative"
for (int i = 0; Matches(ToString(i)); ++i) {
ASSERT_LT(i, 1000);
}
// Bad: 1 block not allowed (for implementation detail reasons)
OpenRaw(cft.Reset(128, (1U << 8) + 123U, 254, false));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
// Bad: 0 blocks not allowed
OpenRaw(cft.Reset(128, (0U << 8) + 123U, 254, false));
ASSERT_TRUE(Matches("hello"));
ASSERT_TRUE(Matches("world"));
}
INSTANTIATE_TEST_CASE_P(Full, FullBloomTest,
testing::Values(BloomFilterPolicy::kLegacyBloom,
BloomFilterPolicy::kFastLocalBloom,
BloomFilterPolicy::kStandard128Ribbon));
} // namespace ROCKSDB_NAMESPACE
int main(int argc, char** argv) {
::testing::InitGoogleTest(&argc, argv);
ParseCommandLineFlags(&argc, &argv, true);
return RUN_ALL_TESTS();
}
#endif // GFLAGS
Reference in New Issue View Git Blame Copy Permalink