AuroraMiddleware/brotli
1
0
Fork 0
mirror of https://github.com/google/brotli.git synced 2024-11-22 11:40:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

update brotli encoder with latest improvements

Browse Source
This commit is contained in:
Lode Vandevenne 2015-08-28 16:09:23 +02:00
parent db71549ac5
commit 6511d6b016
10 changed files with 152 additions and 65 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
enc/backward_references.cc
View File

@ -112,7 +112,7 @@ class ZopfliCostModel {
uint64_t copynumextra = copyextra[copycode];
uint16_t dist_symbol;
uint32_t distextra;
GetDistCode(dist_code, &dist_symbol, &distextra);
PrefixEncodeCopyDistance(dist_code, 0, 0, &dist_symbol, &distextra);
uint32_t distnumextra = distextra >> 24;
double result = insnumextra + copynumextra + distnumextra;
@ -517,7 +517,7 @@ void CreateBackwardReferences(size_t num_bytes,
// Minimum score to accept a backward reference.
const int kMinScore = 4.0;
while (i + 3 < i_end) {
while (i + Hasher::kHashTypeLength - 1 < i_end) {
int max_length = i_end - i;
size_t max_distance = std::min(i + i_diff, max_backward_limit);
int best_len = 0;

12
enc/bit_cost.h
View File

@ -17,6 +17,7 @@
#ifndef BROTLI_ENC_BIT_COST_H_
#define BROTLI_ENC_BIT_COST_H_
#include <stdint.h>
#include "./entropy_encode.h"
@ -24,7 +25,8 @@
namespace brotli {
static inline double BitsEntropy(const int *population, int size) {
static inline double ShannonEntropy(const int *population, int size,
int *total) {
int sum = 0;
double retval = 0;
const int *population_end = population + size;
@ -42,6 +44,13 @@ static inline double BitsEntropy(const int *population, int size) {
retval -= p * FastLog2(p);
}
if (sum) retval += sum * FastLog2(sum);
*total = sum;
return retval;
}
static inline double BitsEntropy(const int *population, int size) {
int sum;
double retval = ShannonEntropy(population, size, &sum);
if (retval < sum) {
// At least one bit per literal is needed.
retval = sum;
@ -49,6 +58,7 @@ static inline double BitsEntropy(const int *population, int size) {
return retval;
}
template<int kSize>
double PopulationCost(const Histogram<kSize>& histogram) {
if (histogram.total_count_ == 0) {

2
enc/brotli_bit_stream.cc
View File

@ -731,7 +731,7 @@ bool StoreMetaBlock(const uint8_t* input,
int distnumextra = cmd.dist_extra_ >> 24;
int distextra = cmd.dist_extra_ & 0xffffff;
if (mb.distance_context_map.empty()) {
distance_enc.StoreSymbol(dist_code, storage_ix, storage);
distance_enc.StoreSymbol(dist_code, storage_ix, storage);
} else {
int context = cmd.DistanceContext();
distance_enc.StoreSymbolWithContext<kDistanceContextBits>(

2
enc/cluster.h
View File

@ -213,6 +213,7 @@ void HistogramRemap(const HistogramType* in, int in_size,
symbols[i] = best_out;
}
// Recompute each out based on raw and symbols.
for (std::set<int>::const_iterator k = all_symbols.begin();
k != all_symbols.end(); ++k) {
@ -283,6 +284,7 @@ void ClusterHistograms(const std::vector<HistogramType>& in,
HistogramReindex(out, histogram_symbols);
}
} // namespace brotli
#endif // BROTLI_ENC_CLUSTER_H_

20
enc/command.h
View File

@ -19,23 +19,10 @@
#include <stdint.h>
#include "./fast_log.h"
#include "./prefix.h"
namespace brotli {
static inline void GetDistCode(int distance_code,
uint16_t* code, uint32_t* extra) {
if (distance_code < 16) {
*code = distance_code;
*extra = 0;
} else {
distance_code -= 12;
int numextra = Log2FloorNonZero(distance_code) - 1;
int prefix = distance_code >> numextra;
*code = 12 + 2 * numextra + prefix;
*extra = (numextra << 24) | (distance_code - (prefix << numextra));
}
}
static int insbase[] = { 0, 1, 2, 3, 4, 5, 6, 8, 10, 14, 18, 26, 34, 50, 66,
98, 130, 194, 322, 578, 1090, 2114, 6210, 22594 };
static int insextra[] = { 0, 0, 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5,
@ -108,7 +95,10 @@ struct Command {
// distance_code is e.g. 0 for same-as-last short code, or 16 for offset 1.
Command(int insertlen, int copylen, int copylen_code, int distance_code)
: insert_len_(insertlen), copy_len_(copylen) {
GetDistCode(distance_code, &dist_prefix_, &dist_extra_);
// The distance prefix and extra bits are stored in this Command as if
// npostfix and ndirect were 0, they are only recomputed later after the
// clustering if needed.
PrefixEncodeCopyDistance(distance_code, 0, 0, &dist_prefix_, &dist_extra_);
GetLengthCode(insertlen, copylen_code, dist_prefix_,
&cmd_prefix_, &cmd_extra_);
}

109
enc/encode.cc
View File

@ -108,8 +108,7 @@ void RecomputeDistancePrefixes(Command* cmds,
size_t num_commands,
int num_direct_distance_codes,
int distance_postfix_bits) {
if (num_direct_distance_codes == 0 &&
distance_postfix_bits == 0) {
if (num_direct_distance_codes == 0 && distance_postfix_bits == 0) {
return;
}
for (int i = 0; i < num_commands; ++i) {
@ -187,9 +186,12 @@ BrotliCompressor::BrotliCompressor(BrotliParams params)
} else if (params_.lgwin == 17) {
last_byte_ = 1;
last_byte_bits_ = 7;
} else {
} else if (params_.lgwin > 17) {
last_byte_ = ((params_.lgwin - 17) << 1) | 1;
last_byte_bits_ = 4;
} else {
last_byte_ = ((params_.lgwin - 8) << 4) | 1;
last_byte_bits_ = 7;
}
// Initialize distance cache.
@ -340,6 +342,71 @@ bool BrotliCompressor::WriteBrotliData(const bool is_last,
return WriteMetaBlockInternal(is_last, utf8_mode, out_size, output);
}
// Decide about the context map based on the ability of the prediction
// ability of the previous byte UTF8-prefix on the next byte. The
// prediction ability is calculated as shannon entropy. Here we need
// shannon entropy instead of 'BitsEntropy' since the prefix will be
// encoded with the remaining 6 bits of the following byte, and
// BitsEntropy will assume that symbol to be stored alone using Huffman
// coding.
void ChooseContextMap(int quality,
int* bigram_histo,
int* num_literal_contexts,
const int** literal_context_map) {
int monogram_histo[3] = { 0 };
int two_prefix_histo[6] = { 0 };
int total = 0;
for (int i = 0; i < 9; ++i) {
total += bigram_histo[i];
monogram_histo[i % 3] += bigram_histo[i];
int j = i;
if (j >= 6) {
j -= 6;
}
two_prefix_histo[j] += bigram_histo[i];
}
int dummy;
double entropy1 = ShannonEntropy(monogram_histo, 3, &dummy);
double entropy2 = (ShannonEntropy(two_prefix_histo, 3, &dummy) +
ShannonEntropy(two_prefix_histo + 3, 3, &dummy));
double entropy3 = 0;
for (int k = 0; k < 3; ++k) {
entropy3 += ShannonEntropy(bigram_histo + 3 * k, 3, &dummy);
}
entropy1 *= (1.0 / total);
entropy2 *= (1.0 / total);
entropy3 *= (1.0 / total);
static const int kStaticContextMapContinuation[64] = {
1, 1, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
};
static const int kStaticContextMapSimpleUTF8[64] = {
0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
};
if (quality < 7) {
// 3 context models is a bit slower, don't use it at lower qualities.
entropy3 = entropy1 * 10;
}
// If expected savings by symbol are less than 0.2 bits, skip the
// context modeling -- in exchange for faster decoding speed.
if (entropy1 - entropy2 < 0.2 &&
entropy1 - entropy3 < 0.2) {
*num_literal_contexts = 1;
} else if (entropy2 - entropy3 < 0.02) {
*num_literal_contexts = 2;
*literal_context_map = kStaticContextMapSimpleUTF8;
} else {
*num_literal_contexts = 3;
*literal_context_map = kStaticContextMapContinuation;
}
}
void DecideOverLiteralContextModeling(const uint8_t* input,
size_t start_pos,
size_t length,
@ -351,40 +418,24 @@ void DecideOverLiteralContextModeling(const uint8_t* input,
if (quality < kMinQualityForContextModeling || length < 64) {
return;
}
// Simple heuristics to guess if the data is UTF8 or not. The goal is to
// recognize non-UTF8 data quickly by searching for the following obvious
// violations: a continuation byte following an ASCII byte or an ASCII or
// lead byte following a lead byte. If we find such violation we decide that
// the data is not UTF8. To make the analysis of UTF8 data faster we only
// examine 64 byte long strides at every 4kB intervals, if there are no
// violations found, we assume the whole data is UTF8.
// Gather bigram data of the UTF8 byte prefixes. To make the analysis of
// UTF8 data faster we only examine 64 byte long strides at every 4kB
// intervals.
const size_t end_pos = start_pos + length;
int bigram_prefix_histo[9] = { 0 };
for (; start_pos + 64 < end_pos; start_pos += 4096) {
static const int lut[4] = { 0, 0, 1, 2 };
const size_t stride_end_pos = start_pos + 64;
uint8_t prev = input[start_pos & mask];
int prev = lut[input[start_pos & mask] >> 6] * 3;
for (size_t pos = start_pos + 1; pos < stride_end_pos; ++pos) {
const uint8_t literal = input[pos & mask];
if ((prev < 128 && (literal & 0xc0) == 0x80) ||
(prev >= 192 && (literal & 0xc0) != 0x80)) {
return;
}
prev = literal;
++bigram_prefix_histo[prev + lut[literal >> 6]];
prev = lut[literal >> 6] * 3;
}
}
*literal_context_mode = CONTEXT_UTF8;
// If the data is UTF8, this static context map distinguishes between ASCII
// or lead bytes and continuation bytes: the UTF8 context value based on the
// last two bytes is 2 or 3 if and only if the next byte is a continuation
// byte (see table in context.h).
static const int kStaticContextMap[64] = {
0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
};
static const int kNumLiteralContexts = 2;
*num_literal_contexts = kNumLiteralContexts;
*literal_context_map = kStaticContextMap;
ChooseContextMap(quality, &bigram_prefix_histo[0], num_literal_contexts,
literal_context_map);
}
bool BrotliCompressor::WriteMetaBlockInternal(const bool is_last,

4
enc/encode.h
View File

@ -30,7 +30,7 @@
namespace brotli {
static const int kMaxWindowBits = 24;
static const int kMinWindowBits = 16;
static const int kMinWindowBits = 10;
static const int kMinInputBlockBits = 16;
static const int kMaxInputBlockBits = 24;
@ -59,7 +59,7 @@ struct BrotliParams {
// Controls the compression-speed vs compression-density tradeoffs. The higher
// the quality, the slower the compression. Range is 0 to 11.
int quality;
// Base 2 logarithm of the sliding window size. Range is 16 to 24.
// Base 2 logarithm of the sliding window size. Range is 10 to 24.
int lgwin;
// Base 2 logarithm of the maximum input block size. Range is 16 to 24.
// If set to 0, the value will be set based on the quality.

54
enc/hash.h
View File

@ -130,9 +130,7 @@ class HashLongestMatchQuickly {
// not filling will make the results of the compression stochastic
// (but correct). This is because random data would cause the
// system to find accidentally good backward references here and there.
std::fill(&buckets_[0],
&buckets_[sizeof(buckets_) / sizeof(buckets_[0])],
0);
memset(&buckets_[0], 0, sizeof(buckets_));
num_dict_lookups_ = 0;
num_dict_matches_ = 0;
}
@ -140,7 +138,7 @@ class HashLongestMatchQuickly {
// Compute a hash from these, and store the value somewhere within
// [ix .. ix+3].
inline void Store(const uint8_t *data, const int ix) {
const uint32_t key = Hash<kBucketBits>(data);
const uint32_t key = HashBytes(data);
// Wiggle the value with the bucket sweep range.
const uint32_t off = (static_cast<uint32_t>(ix) >> 3) % kBucketSweep;
buckets_[key + off] = ix;
@ -201,7 +199,7 @@ class HashLongestMatchQuickly {
}
}
}
const uint32_t key = Hash<kBucketBits>(&ring_buffer[cur_ix_masked]);
const uint32_t key = HashBytes(&ring_buffer[cur_ix_masked]);
if (kBucketSweep == 1) {
// Only one to look for, don't bother to prepare for a loop.
prev_ix = buckets_[key];
@ -291,6 +289,21 @@ class HashLongestMatchQuickly {
return match_found;
}
enum { kHashLength = 5 };
enum { kHashTypeLength = 8 };
// HashBytes is the function that chooses the bucket to place
// the address in. The HashLongestMatch and HashLongestMatchQuickly
// classes have separate, different implementations of hashing.
static uint32_t HashBytes(const uint8_t *data) {
// Computing a hash based on 5 bytes works much better for
// qualities 1 and 3, where the next hash value is likely to replace
static const uint32_t kHashMul32 = 0x1e35a7bd;
uint64_t h = (BROTLI_UNALIGNED_LOAD64(data) << 24) * kHashMul32;
// The higher bits contain more mixture from the multiplication,
// so we take our results from there.
return h >> (64 - kBucketBits);
}
private:
static const uint32_t kBucketSize = 1 << kBucketBits;
uint32_t buckets_[kBucketSize + kBucketSweep];
@ -317,7 +330,7 @@ class HashLongestMatch {
}
void Reset() {
std::fill(&num_[0], &num_[sizeof(num_) / sizeof(num_[0])], 0);
memset(&num_[0], 0, sizeof(num_));
num_dict_lookups_ = 0;
num_dict_matches_ = 0;
}
@ -325,7 +338,7 @@ class HashLongestMatch {
// Look at 3 bytes at data.
// Compute a hash from these, and store the value of ix at that position.
inline void Store(const uint8_t *data, const int ix) {
const uint32_t key = Hash<kBucketBits>(data);
const uint32_t key = HashBytes(data);
const int minor_ix = num_[key] & kBlockMask;
buckets_[key][minor_ix] = ix;
++num_[key];
@ -401,7 +414,7 @@ class HashLongestMatch {
}
}
}
const uint32_t key = Hash<kBucketBits>(&data[cur_ix_masked]);
const uint32_t key = HashBytes(&data[cur_ix_masked]);
const int * __restrict const bucket = &buckets_[key][0];
const int down = (num_[key] > kBlockSize) ? (num_[key] - kBlockSize) : 0;
for (int i = num_[key] - 1; i >= down; --i) {
@ -518,7 +531,7 @@ class HashLongestMatch {
*matches++ = BackwardMatch(backward, len);
}
}
const uint32_t key = Hash<kBucketBits>(&data[cur_ix_masked]);
const uint32_t key = HashBytes(&data[cur_ix_masked]);
const int * __restrict const bucket = &buckets_[key][0];
const int down = (num_[key] > kBlockSize) ? (num_[key] - kBlockSize) : 0;
for (int i = num_[key] - 1; i >= down; --i) {
@ -562,6 +575,27 @@ class HashLongestMatch {
*num_matches += matches - orig_matches;
}
enum { kHashLength = 4 };
enum { kHashTypeLength = 4 };
// HashBytes is the function that chooses the bucket to place
// the address in. The HashLongestMatch and HashLongestMatchQuickly
// classes have separate, different implementations of hashing.
static uint32_t HashBytes(const uint8_t *data) {
// kHashMul32 multiplier has these properties:
// * The multiplier must be odd. Otherwise we may lose the highest bit.
// * No long streaks of 1s or 0s.
// * Is not unfortunate (see the unittest) for the English language.
// * There is no effort to ensure that it is a prime, the oddity is enough
// for this use.
// * The number has been tuned heuristically against compression benchmarks.
static const uint32_t kHashMul32 = 0x1e35a7bd;
uint32_t h = BROTLI_UNALIGNED_LOAD32(data) * kHashMul32;
// The higher bits contain more mixture from the multiplication,
// so we take our results from there.
return h >> (32 - kBucketBits);
}
private:
// Number of hash buckets.
static const uint32_t kBucketSize = 1 << kBucketBits;
@ -614,7 +648,7 @@ struct Hashers {
template<typename Hasher>
void WarmupHash(const size_t size, const uint8_t* dict, Hasher* hasher) {
for (size_t i = 0; i + 3 < size; i++) {
for (size_t i = 0; i + Hasher::kHashTypeLength - 1 < size; i++) {
hasher->Store(dict, i);
}
}

6
enc/ringbuffer.h
View File

@ -36,10 +36,10 @@ class RingBuffer {
mask_((1 << window_bits) - 1),
tail_size_(1 << tail_bits),
pos_(0) {
static const int kSlackForFourByteHashingEverywhere = 3;
static const int kSlackForEightByteHashingEverywhere = 7;
const int buflen = (1 << window_bits_) + tail_size_;
buffer_ = new uint8_t[buflen + kSlackForFourByteHashingEverywhere];
for (int i = 0; i < kSlackForFourByteHashingEverywhere; ++i) {
buffer_ = new uint8_t[buflen + kSlackForEightByteHashingEverywhere];
for (int i = 0; i < kSlackForEightByteHashingEverywhere; ++i) {
buffer_[buflen + i] = 0;
}
}

4
enc/static_dict.cc
View File

@ -377,7 +377,7 @@ bool FindAllStaticDictionaryMatches(const uint8_t* data,
// Transforms with prefixes "e ", "s ", ", " and "\xc2\xa0"
if ((data[1] == ' ' &&
(data[0] == 'e' || data[0] == 's' || data[0] == ',')) ||
(data[0] == '\xc2' && data[1] == '\xa0')) {
(data[0] == 0xc2 && data[1] == 0xa0)) {
key = Hash(&data[2]);
bucket = kStaticDictionaryBuckets[key];
int num = bucket & 0xff;
@ -388,7 +388,7 @@ bool FindAllStaticDictionaryMatches(const uint8_t* data,
const int n = 1 << kBrotliDictionarySizeBitsByLength[l];
const int id = w.idx;
if (w.transform == 0 && IsMatch(w, &data[2], max_length - 2)) {
if (data[0] == '\xc2') {
if (data[0] == 0xc2) {
AddMatch(id + 102 * n, l + 2, l, matches);
found_match = true;
} else if (l + 2 < max_length && data[l + 2] == ' ') {