[libcxx] r345525 - Bug 39129: Speeding up partition_point/lower_bound/upper_bound/ by using unsigned division by 2 when possible.
Eric Fiselier
eric at efcs.ca
Mon Oct 29 12:25:03 PDT 2018
Author: ericwf
Date: Mon Oct 29 12:25:02 2018
New Revision: 345525
URL: http://llvm.org/viewvc/llvm-project?rev=345525&view=rev
Log:
Bug 39129: Speeding up partition_point/lower_bound/upper_bound/ by using unsigned division by 2 when possible.
Patch by Denis Yaroshevskiy (denis.yaroshevskij at gmail.com)
The rational and measurements can be found in the bug description: https://bugs.llvm.org/show_bug.cgi?id=39129
Reviewed as https://reviews.llvm.org/D52697
Added:
libcxx/trunk/test/libcxx/algorithms/half_positive.pass.cpp
Modified:
libcxx/trunk/benchmarks/algorithms.bench.cpp
libcxx/trunk/include/algorithm
Modified: libcxx/trunk/benchmarks/algorithms.bench.cpp
URL: http://llvm.org/viewvc/llvm-project/libcxx/trunk/benchmarks/algorithms.bench.cpp?rev=345525&r1=345524&r2=345525&view=diff
==============================================================================
--- libcxx/trunk/benchmarks/algorithms.bench.cpp (original)
+++ libcxx/trunk/benchmarks/algorithms.bench.cpp Mon Oct 29 12:25:02 2018
@@ -58,5 +58,69 @@ BENCHMARK_CAPTURE(BM_Sort, sorted_descen
BENCHMARK_CAPTURE(BM_Sort, single_element_strings,
getDuplicateStringInputs)->Arg(TestNumInputs);
+template <typename GenInputs, typename Alg>
+void do_binary_search_benchmark(benchmark::State& st, GenInputs gen, Alg alg)
+{
+ using ValueType = typename decltype(gen(0))::value_type;
+ auto in = gen(st.range(0));
+ std::sort(in.begin(), in.end());
+
+ const auto every_10_percentile = [&]() -> std::vector<ValueType*> {
+ size_t step = in.size() / 10;
+
+ if (step == 0) {
+ st.SkipWithError("Input doesn't contain enough elements");
+ return {};
+ }
+
+ std::vector<ValueType*> res;
+ for (size_t i = 0; i < in.size(); i += step)
+ res.push_back(&in[i]);
+
+ return res;
+ }();
+
+ for (auto _ : st)
+ {
+ for (auto* test : every_10_percentile)
+ benchmark::DoNotOptimize(alg(in.begin(), in.end(), *test));
+ }
+}
+
+template <typename GenInputs>
+void BM_LowerBound(benchmark::State& st, GenInputs gen)
+{
+ do_binary_search_benchmark(st, gen, [](auto f, auto l, const auto& v) {
+ return std::lower_bound(f, l, v);
+ });
+}
+
+BENCHMARK_CAPTURE(BM_LowerBound, random_int32, getRandomIntegerInputs<int32_t>)
+ ->Arg(TestNumInputs) // Small int32_t vector
+ ->Arg(TestNumInputs * TestNumInputs); // Big int32_t vector
+
+BENCHMARK_CAPTURE(BM_LowerBound, random_int64, getRandomIntegerInputs<int64_t>)
+ ->Arg(TestNumInputs); // Small int64_t vector. Should also represent pointers.
+
+BENCHMARK_CAPTURE(BM_LowerBound, random_strings, getRandomStringInputs)
+ ->Arg(TestNumInputs); // Small string vector. What happens if the comparison is not very cheap.
+
+template <typename GenInputs>
+void BM_EqualRange(benchmark::State& st, GenInputs gen)
+{
+ do_binary_search_benchmark(st, gen, [](auto f, auto l, const auto& v) {
+ return std::equal_range(f, l, v);
+ });
+}
+
+BENCHMARK_CAPTURE(BM_EqualRange, random_int32, getRandomIntegerInputs<int32_t>)
+ ->Arg(TestNumInputs) // Small int32_t vector
+ ->Arg(TestNumInputs * TestNumInputs); // Big int32_t vector
+
+BENCHMARK_CAPTURE(BM_EqualRange, random_int64, getRandomIntegerInputs<int64_t>)
+ ->Arg(TestNumInputs); // Small int64_t vector. Should also represent pointers.
+
+BENCHMARK_CAPTURE(BM_EqualRange, random_strings, getRandomStringInputs)
+ ->Arg(TestNumInputs); // Small string vector. What happens if the comparison is not very cheap.
BENCHMARK_MAIN();
Modified: libcxx/trunk/include/algorithm
URL: http://llvm.org/viewvc/llvm-project/libcxx/trunk/include/algorithm?rev=345525&r1=345524&r2=345525&view=diff
==============================================================================
--- libcxx/trunk/include/algorithm (original)
+++ libcxx/trunk/include/algorithm Mon Oct 29 12:25:02 2018
@@ -750,6 +750,32 @@ public:
bool operator()(const _T1& __x, const _T2& __y) {return __p_(__y, __x);}
};
+// Perform division by two quickly for positive integers (llvm.org/PR39129)
+
+template <typename _Integral>
+_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR
+typename enable_if
+<
+ is_integral<_Integral>::value,
+ _Integral
+>::type
+__half_positive(_Integral __value)
+{
+ return static_cast<_Integral>(static_cast<typename make_unsigned<_Integral>::type>(__value) / 2);
+}
+
+template <typename _Tp>
+_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR
+typename enable_if
+<
+ !is_integral<_Tp>::value,
+ _Tp
+>::type
+__half_positive(_Tp __value)
+{
+ return __value / 2;
+}
+
#ifdef _LIBCPP_DEBUG
template <class _Compare>
@@ -3202,7 +3228,7 @@ partition_point(_ForwardIterator __first
difference_type __len = _VSTD::distance(__first, __last);
while (__len != 0)
{
- difference_type __l2 = __len / 2;
+ difference_type __l2 = _VSTD::__half_positive(__len);
_ForwardIterator __m = __first;
_VSTD::advance(__m, __l2);
if (__pred(*__m))
@@ -4069,7 +4095,7 @@ __lower_bound(_ForwardIterator __first,
difference_type __len = _VSTD::distance(__first, __last);
while (__len != 0)
{
- difference_type __l2 = __len / 2;
+ difference_type __l2 = _VSTD::__half_positive(__len);
_ForwardIterator __m = __first;
_VSTD::advance(__m, __l2);
if (__comp(*__m, __value_))
@@ -4111,7 +4137,7 @@ __upper_bound(_ForwardIterator __first,
difference_type __len = _VSTD::distance(__first, __last);
while (__len != 0)
{
- difference_type __l2 = __len / 2;
+ difference_type __l2 = _VSTD::__half_positive(__len);
_ForwardIterator __m = __first;
_VSTD::advance(__m, __l2);
if (__comp(__value_, *__m))
@@ -4153,7 +4179,7 @@ __equal_range(_ForwardIterator __first,
difference_type __len = _VSTD::distance(__first, __last);
while (__len != 0)
{
- difference_type __l2 = __len / 2;
+ difference_type __l2 = _VSTD::__half_positive(__len);
_ForwardIterator __m = __first;
_VSTD::advance(__m, __l2);
if (__comp(*__m, __value_))
Added: libcxx/trunk/test/libcxx/algorithms/half_positive.pass.cpp
URL: http://llvm.org/viewvc/llvm-project/libcxx/trunk/test/libcxx/algorithms/half_positive.pass.cpp?rev=345525&view=auto
==============================================================================
--- libcxx/trunk/test/libcxx/algorithms/half_positive.pass.cpp (added)
+++ libcxx/trunk/test/libcxx/algorithms/half_positive.pass.cpp Mon Oct 29 12:25:02 2018
@@ -0,0 +1,59 @@
+//===----------------------------------------------------------------------===//
+//
+// The LLVM Compiler Infrastructure
+//
+// This file is dual licensed under the MIT and the University of Illinois Open
+// Source Licenses. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+// <algorithm>
+
+// template <typename _Tp> _Tp __half_positive(const _Tp&);
+
+// __half_positive divide integer number by 2 as unsigned number
+// if it's safe to do so. It can be an important optimization for lower bound,
+// for example.
+
+#include <algorithm>
+#include <cassert>
+#include <limits>
+#include <type_traits>
+
+#include "test_macros.h"
+#include "user_defined_integral.hpp"
+
+namespace {
+
+template <class IntType, class UnderlyingType = IntType>
+TEST_CONSTEXPR bool test(IntType max_v = IntType(std::numeric_limits<UnderlyingType>::max())) {
+ return std::__half_positive(max_v) == max_v / 2;
+}
+
+} // namespace
+
+int main()
+{
+ {
+ assert(test<char>());
+ assert(test<int>());
+ assert(test<long>());
+ assert((test<UserDefinedIntegral<int>, int>()));
+ assert(test<size_t>());
+#if !defined(_LIBCPP_HAS_NO_INT128)
+ assert(test<__int128_t>());
+#endif // !defined(_LIBCPP_HAS_NO_INT128)
+ }
+
+#if TEST_STD_VER >= 11
+ {
+ static_assert(test<char>(), "");
+ static_assert(test<int>(), "");
+ static_assert(test<long>(), "");
+ static_assert(test<size_t>(), "");
+#if !defined(_LIBCPP_HAS_NO_INT128)
+ static_assert(test<__int128_t>(), "");
+#endif // !defined(_LIBCPP_HAS_NO_INT128)
+ }
+#endif // TEST_STD_VER >= 11
+}
More information about the libcxx-commits
mailing list