tool
/
starrocks
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[Enhancement] faster from_unixtime (#60174)

This commit is contained in:
Murphy 2025-07-09 17:04:12 +08:00 committed by GitHub
parent a04ba3afca
commit 46804c47bf
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 141 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
thirdparty/download-thirdparty.sh vendored
View File

@ -614,3 +614,14 @@ if [[ -d $TP_SOURCE_DIR/$AZURE_SOURCE ]] ; then
cd -
echo "Finished patching $AZURE_SOURCE"
fi
#patch cctz
if [[ -d $TP_SOURCE_DIR/$CCTZ_SOURCE ]] ; then
cd $TP_SOURCE_DIR/$CCTZ_SOURCE
if [ ! -f "$PATCHED_MARK" ] && [[ $CCTZ_SOURCE == "cctz-2.3" ]] ; then
patch -p1 < "$TP_PATCH_DIR/cctz_civil_cache.patch"
touch "$PATCHED_MARK"
fi
cd -
echo "Finished patching $CCTZ_SOURCE"
fi

130
thirdparty/patches/cctz_civil_cache.patch vendored Normal file
View File

@ -0,0 +1,130 @@
diff --git a/src/cctz_benchmark.cc b/src/cctz_benchmark.cc
index 179ae50..87a7722 100644
--- a/src/cctz_benchmark.cc
+++ b/src/cctz_benchmark.cc
@@ -797,6 +797,20 @@ void BM_Time_ToCivil_CCTZ(benchmark::State& state) {
}
BENCHMARK(BM_Time_ToCivil_CCTZ);
+void BM_Time_ToCivil_Cached(benchmark::State &state) {
+ const cctz::time_zone tz = TestTimeZone();
+ std::chrono::system_clock::time_point tp =
+ std::chrono::system_clock::from_time_t(1750656243);
+ std::chrono::system_clock::time_point tp2 =
+ std::chrono::system_clock::from_time_t(1750656245);
+ while (state.KeepRunning()) {
+ std::swap(tp, tp2);
+ tp += std::chrono::seconds(1);
+ benchmark::DoNotOptimize(cctz::convert(tp, tz));
+ }
+}
+BENCHMARK(BM_Time_ToCivil_Cached);
+
void BM_Time_ToCivil_Libc(benchmark::State& state) {
// No timezone support, so just use localtime.
time_t t = 1384569027;
diff --git a/src/time_zone_info.cc b/src/time_zone_info.cc
index eb1cd8a..6258c27 100644
--- a/src/time_zone_info.cc
+++ b/src/time_zone_info.cc
@@ -41,12 +41,14 @@
#include <cstring>
#include <functional>
#include <iostream>
+#include <mutex>
#include <memory>
#include <sstream>
#include <string>
#include "cctz/civil_time.h"
#include "time_zone_fixed.h"
+#include "time_zone_if.h"
#include "time_zone_posix.h"
namespace cctz {
@@ -79,6 +81,21 @@ const std::int_least32_t kSecsPerYear[2] = {
366 * kSecsPerDay,
};
+static constexpr int kCivilCacheDays = 200 * 365;
+static civil_day gCivilCache[kCivilCacheDays];
+static std::once_flag gCivilCacheInit;
+
+static void InitCivilDaysCache() {
+ std::call_once(gCivilCacheInit, [&]() {
+ civil_day epoch;
+ for (int i = 0; i < kCivilCacheDays; i++) {
+ gCivilCache[i] = epoch;
+ epoch++;
+ }
+ });
+}
+
+
// Single-byte, unsigned numeric values are encoded directly.
inline std::uint_fast8_t Decode8(const char* cp) {
return static_cast<std::uint_fast8_t>(*cp) & 0xff;
@@ -751,14 +768,56 @@ time_zone::absolute_lookup TimeZoneInfo::LocalTime(
tt.utc_offset, tt.is_dst, &abbreviations_[tt.abbr_index]};
}
+static inline int64_t floor_div(int64_t n, int64_t d) {
+ int64_t q = n / d;
+ int64_t r = n % d;
+ return (r < 0) ? q - 1 : q;
+}
+
// BreakTime() translation for a particular transition.
-time_zone::absolute_lookup TimeZoneInfo::LocalTime(
- std::int_fast64_t unix_time, const Transition& tr) const {
- const TransitionType& tt = transition_types_[tr.type_index];
- // Note: (unix_time - tr.unix_time) will never overflow as we
- // have ensured that there is always a "nearby" transition.
- return {tr.civil_sec + (unix_time - tr.unix_time), // TODO: Optimize.
- tt.utc_offset, tt.is_dst, &abbreviations_[tt.abbr_index]};
+time_zone::absolute_lookup TimeZoneInfo::LocalTime(std::int_fast64_t unix_time,
+ const Transition &tr) const {
+ const TransitionType &tt = transition_types_[tr.type_index];
+ // Baseline
+ // return {tr.civil_sec + (unix_time-tr.unix_time), tt.utc_offset, tt.is_dst,
+ // &abbreviations_[tt.abbr_index]};
+
+ // Optimization: Utilizing a cache for n_days() operation to mitigate its 30+ns performance impact.
+ // Benchmark Time CPU Time Old Time New CPU Old CPU New
+ // ----------------------------------------------------------------------------------------------------------------------
+ // BM_Time_ToCivil_CCTZ -0.2508 -0.2508 55 41 55 41
+ // BM_Time_ToCivil_Cached -0.5182 -0.5182 71 34 71 34
+ // BM_Time_ToCivil_Libc +0.0020 +0.0020 217 217 217 217
+ // BM_Time_ToCivilUTC_CCTZ -0.6477 -0.6477 73 26 73 26
+ // BM_Time_ToCivilUTC_Libc +0.0145 +0.0144 43 44 43 44
+ // OVERALL_GEOMEAN -0.3358 -0.3358 0 0 0 0
+ std::int_fast64_t local_unix = unix_time + tt.utc_offset;
+ if (local_unix < unix_time) {
+ // Handle overflow case
+ return {tr.civil_sec + (unix_time - tr.unix_time), tt.utc_offset, tt.is_dst,
+ &abbreviations_[tt.abbr_index]};
+ }
+
+ constexpr int64_t SECS_PER_DAY = 86400;
+ std::int_fast64_t days = floor_div(local_unix, SECS_PER_DAY);
+ std::int_fast64_t sod = local_unix - days * SECS_PER_DAY;
+ if (sod < 0) {
+ sod += SECS_PER_DAY;
+ }
+
+ if (0 <= days && days < kCivilCacheDays) {
+ InitCivilDaysCache();
+ civil_day local_civil_day = gCivilCache[days];
+ civil_second local_civil(local_civil_day.year(), local_civil_day.month(), local_civil_day.day(),
+ sod / (3600), sod % 3600 / 60, sod % 60);
+ return {local_civil, tt.utc_offset, tt.is_dst,
+ &abbreviations_[tt.abbr_index]};
+ } else {
+ // Note: (unix_time - tr.unix_time) will never overflow as we
+ // have ensured that there is always a "nearby" transition.
+ return {tr.civil_sec + (unix_time - tr.unix_time), tt.utc_offset, tt.is_dst,
+ &abbreviations_[tt.abbr_index]};
+ }
}
// MakeTime() translation with a conversion-preserving +N * 400-year shift.