[libc++] Optimize ranges::equal for vector<bool>::iterator #121084
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
winner245
force-pushed
the
optimize-ranges-equal
branch
3 times, most recently
from
8dbf8de to
40b8a42
Compare
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
winner245
force-pushed
the
optimize-ranges-equal
branch
from
40b8a42 to
098b68d
Compare
winner245
force-pushed
the
optimize-ranges-equal
branch
2 times, most recently
from
d51edbb to
878808f
Compare
llvmbot
added
the
libc++
|
@llvm/pr-subscribers-libcxx Author: Peng Liu (winner245) ChangesThis PR optimizes the performance of
Patch is 36.80 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/121084.diff 6 Files Affected:
diff --git a/libcxx/include/__algorithm/equal.h b/libcxx/include/__algorithm/equal.h
index a276bb9954c9bb..0f8f1147b193c3 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -11,19 +11,27 @@
#define _LIBCPP___ALGORITHM_EQUAL_H
#include <__algorithm/comp.h>
+#include <__algorithm/min.h>
#include <__algorithm/unwrap_iter.h>
#include <__config>
#include <__functional/identity.h>
+#include <__fwd/bit_reference.h>
#include <__iterator/distance.h>
#include <__iterator/iterator_traits.h>
+#include <__memory/pointer_traits.h>
#include <__string/constexpr_c_functions.h>
#include <__type_traits/desugars_to.h>
#include <__type_traits/enable_if.h>
#include <__type_traits/invoke.h>
#include <__type_traits/is_equality_comparable.h>
+#include <__type_traits/is_same.h>
#include <__type_traits/is_volatile.h>
#include <__utility/move.h>
+#if _LIBCPP_STD_VER >= 20
+# include <__functional/ranges_operations.h>
+#endif
+
#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
# pragma GCC system_header
#endif
@@ -33,6 +41,132 @@ _LIBCPP_PUSH_MACROS
_LIBCPP_BEGIN_NAMESPACE_STD
+template <class _Cp, bool _IC1, bool _IC2>
+[[__nodiscard__]] _LIBCPP_CONSTEXPR_SINCE_CXX20 _LIBCPP_HIDE_FROM_ABI bool __equal_unaligned(
+ __bit_iterator<_Cp, _IC1> __first1, __bit_iterator<_Cp, _IC1> __last1, __bit_iterator<_Cp, _IC2> __first2) {
+ using _It = __bit_iterator<_Cp, _IC1>;
+ using difference_type = typename _It::difference_type;
+ using __storage_type = typename _It::__storage_type;
+
+ const int __bits_per_word = _It::__bits_per_word;
+ difference_type __n = __last1 - __first1;
+ if (__n > 0) {
+ // do first word
+ if (__first1.__ctz_ != 0) {
+ unsigned __clz_f = __bits_per_word - __first1.__ctz_;
+ difference_type __dn = std::min(static_cast<difference_type>(__clz_f), __n);
+ __n -= __dn;
+ __storage_type __m = (~__storage_type(0) << __first1.__ctz_) & (~__storage_type(0) >> (__clz_f - __dn));
+ __storage_type __b = *__first1.__seg_ & __m;
+ unsigned __clz_r = __bits_per_word - __first2.__ctz_;
+ __storage_type __ddn = std::min<__storage_type>(__dn, __clz_r);
+ __m = (~__storage_type(0) << __first2.__ctz_) & (~__storage_type(0) >> (__clz_r - __ddn));
+ if (__first2.__ctz_ > __first1.__ctz_) {
+ if ((*__first2.__seg_ & __m) != (__b << (__first2.__ctz_ - __first1.__ctz_)))
+ return false;
+ } else {
+ if ((*__first2.__seg_ & __m) != (__b >> (__first1.__ctz_ - __first2.__ctz_)))
+ return false;
+ }
+ __first2.__seg_ += (__ddn + __first2.__ctz_) / __bits_per_word;
+ __first2.__ctz_ = static_cast<unsigned>((__ddn + __first2.__ctz_) % __bits_per_word);
+ __dn -= __ddn;
+ if (__dn > 0) {
+ __m = ~__storage_type(0) >> (__bits_per_word - __dn);
+ if ((*__first2.__seg_ & __m) != (__b >> (__first1.__ctz_ + __ddn)))
+ return false;
+ __first2.__ctz_ = static_cast<unsigned>(__dn);
+ }
+ ++__first1.__seg_;
+ // __first1.__ctz_ = 0;
+ }
+ // __first1.__ctz_ == 0;
+ // do middle words
+ unsigned __clz_r = __bits_per_word - __first2.__ctz_;
+ __storage_type __m = ~__storage_type(0) << __first2.__ctz_;
+ for (; __n >= __bits_per_word; __n -= __bits_per_word, ++__first1.__seg_) {
+ __storage_type __b = *__first1.__seg_;
+ if ((*__first2.__seg_ & __m) != (__b << __first2.__ctz_))
+ return false;
+ ++__first2.__seg_;
+ if ((*__first2.__seg_ & ~__m) != (__b >> __clz_r))
+ return false;
+ }
+ // do last word
+ if (__n > 0) {
+ __m = ~__storage_type(0) >> (__bits_per_word - __n);
+ __storage_type __b = *__first1.__seg_ & __m;
+ __storage_type __dn = std::min(__n, static_cast<difference_type>(__clz_r));
+ __m = (~__storage_type(0) << __first2.__ctz_) & (~__storage_type(0) >> (__clz_r - __dn));
+ if ((*__first2.__seg_ & __m) != (__b << __first2.__ctz_))
+ return false;
+ __first2.__seg_ += (__dn + __first2.__ctz_) / __bits_per_word;
+ __first2.__ctz_ = static_cast<unsigned>((__dn + __first2.__ctz_) % __bits_per_word);
+ __n -= __dn;
+ if (__n > 0) {
+ __m = ~__storage_type(0) >> (__bits_per_word - __n);
+ if ((*__first2.__seg_ & __m) != (__b >> __dn))
+ return false;
+ }
+ }
+ }
+ return true;
+}
+
+template <class _Cp, bool _IC1, bool _IC2>
+[[__nodiscard__]] _LIBCPP_CONSTEXPR_SINCE_CXX20 _LIBCPP_HIDE_FROM_ABI bool __equal_aligned(
+ __bit_iterator<_Cp, _IC1> __first1, __bit_iterator<_Cp, _IC1> __last1, __bit_iterator<_Cp, _IC2> __first2) {
+ using _It = __bit_iterator<_Cp, _IC1>;
+ using difference_type = typename _It::difference_type;
+ using __storage_type = typename _It::__storage_type;
+
+ const int __bits_per_word = _It::__bits_per_word;
+ difference_type __n = __last1 - __first1;
+ if (__n > 0) {
+ // do first word
+ if (__first1.__ctz_ != 0) {
+ unsigned __clz = __bits_per_word - __first1.__ctz_;
+ difference_type __dn = std::min(static_cast<difference_type>(__clz), __n);
+ __n -= __dn;
+ __storage_type __m = (~__storage_type(0) << __first1.__ctz_) & (~__storage_type(0) >> (__clz - __dn));
+ if ((*__first2.__seg_ & __m) != (*__first1.__seg_ & __m))
+ return false;
+ ++__first2.__seg_;
+ ++__first1.__seg_;
+ // __first1.__ctz_ = 0;
+ // __first2.__ctz_ = 0;
+ }
+ // __first1.__ctz_ == 0;
+ // __first2.__ctz_ == 0;
+ // do middle words
+ for (; __n >= __bits_per_word; __n -= __bits_per_word, ++__first1.__seg_, ++__first2.__seg_)
+ if (*__first2.__seg_ != *__first1.__seg_)
+ return false;
+ // do last word
+ if (__n > 0) {
+ __storage_type __m = ~__storage_type(0) >> (__bits_per_word - __n);
+ if ((*__first2.__seg_ & __m) != (*__first1.__seg_ & __m))
+ return false;
+ }
+ }
+ return true;
+}
+
+template <class _Cp,
+ bool _IC1,
+ bool _IC2,
+ class _BinaryPredicate,
+ __enable_if_t<std::is_same<_BinaryPredicate, __equal_to>::value, int> = 0>
+[[__nodiscard__]] inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 bool __equal_iter_impl(
+ __bit_iterator<_Cp, _IC1> __first1,
+ __bit_iterator<_Cp, _IC1> __last1,
+ __bit_iterator<_Cp, _IC2> __first2,
+ _BinaryPredicate) {
+ if (__first1.__ctz_ == __first2.__ctz_)
+ return std::__equal_aligned(__first1, __last1, __first2);
+ return std::__equal_unaligned(__first1, __last1, __first2);
+}
+
template <class _InputIterator1, class _InputIterator2, class _BinaryPredicate>
[[__nodiscard__]] inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 bool __equal_iter_impl(
_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _BinaryPredicate& __pred) {
@@ -94,6 +228,31 @@ __equal_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&,
return std::__constexpr_memcmp_equal(__first1, __first2, __element_count(__last1 - __first1));
}
+template <class _Cp,
+ bool _IC1,
+ bool _IC2,
+ class _Pred,
+ class _Proj1,
+ class _Proj2,
+ __enable_if_t<(is_same<_Pred, __equal_to>::value
+# if _LIBCPP_STD_VER >= 20
+ || is_same<_Pred, ranges::equal_to>::value
+# endif
+ ) &&
+ __desugars_to_v<__equal_tag, _Pred, bool, bool> && __is_identity<_Proj1>::value &&
+ __is_identity<_Proj2>::value,
+ int> = 0>
+[[__nodiscard__]] inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 bool __equal_impl(
+ __bit_iterator<_Cp, _IC1> __first1,
+ __bit_iterator<_Cp, _IC1> __last1,
+ __bit_iterator<_Cp, _IC2> __first2,
+ __bit_iterator<_Cp, _IC2>,
+ _Pred&,
+ _Proj1&,
+ _Proj2&) {
+ return std::__equal_iter_impl(__first1, __last1, __first2, __equal_to());
+}
+
template <class _InputIterator1, class _InputIterator2, class _BinaryPredicate>
[[__nodiscard__]] inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 bool
equal(_InputIterator1 __first1,
diff --git a/libcxx/include/__bit_reference b/libcxx/include/__bit_reference
index 67abb023122edf..8a0c4c93bcbaf7 100644
--- a/libcxx/include/__bit_reference
+++ b/libcxx/include/__bit_reference
@@ -10,7 +10,9 @@
#ifndef _LIBCPP___BIT_REFERENCE
#define _LIBCPP___BIT_REFERENCE
+#include <__algorithm/comp.h>
#include <__algorithm/copy_n.h>
+#include <__algorithm/equal.h>
#include <__algorithm/min.h>
#include <__bit/countr.h>
#include <__compare/ordering.h>
@@ -22,7 +24,9 @@
#include <__memory/construct_at.h>
#include <__memory/pointer_traits.h>
#include <__type_traits/conditional.h>
+#include <__type_traits/enable_if.h>
#include <__type_traits/is_constant_evaluated.h>
+#include <__type_traits/is_same.h>
#include <__type_traits/void_t.h>
#include <__utility/swap.h>
@@ -669,127 +673,6 @@ rotate(__bit_iterator<_Cp, false> __first, __bit_iterator<_Cp, false> __middle,
return __r;
}
-// equal
-
-template <class _Cp, bool _IC1, bool _IC2>
-_LIBCPP_CONSTEXPR_SINCE_CXX20 _LIBCPP_HIDE_FROM_ABI bool __equal_unaligned(
- __bit_iterator<_Cp, _IC1> __first1, __bit_iterator<_Cp, _IC1> __last1, __bit_iterator<_Cp, _IC2> __first2) {
- using _It = __bit_iterator<_Cp, _IC1>;
- using difference_type = typename _It::difference_type;
- using __storage_type = typename _It::__storage_type;
-
- const int __bits_per_word = _It::__bits_per_word;
- difference_type __n = __last1 - __first1;
- if (__n > 0) {
- // do first word
- if (__first1.__ctz_ != 0) {
- unsigned __clz_f = __bits_per_word - __first1.__ctz_;
- difference_type __dn = std::min(static_cast<difference_type>(__clz_f), __n);
- __n -= __dn;
- __storage_type __m = (~__storage_type(0) << __first1.__ctz_) & (~__storage_type(0) >> (__clz_f - __dn));
- __storage_type __b = *__first1.__seg_ & __m;
- unsigned __clz_r = __bits_per_word - __first2.__ctz_;
- __storage_type __ddn = std::min<__storage_type>(__dn, __clz_r);
- __m = (~__storage_type(0) << __first2.__ctz_) & (~__storage_type(0) >> (__clz_r - __ddn));
- if (__first2.__ctz_ > __first1.__ctz_) {
- if ((*__first2.__seg_ & __m) != (__b << (__first2.__ctz_ - __first1.__ctz_)))
- return false;
- } else {
- if ((*__first2.__seg_ & __m) != (__b >> (__first1.__ctz_ - __first2.__ctz_)))
- return false;
- }
- __first2.__seg_ += (__ddn + __first2.__ctz_) / __bits_per_word;
- __first2.__ctz_ = static_cast<unsigned>((__ddn + __first2.__ctz_) % __bits_per_word);
- __dn -= __ddn;
- if (__dn > 0) {
- __m = ~__storage_type(0) >> (__bits_per_word - __dn);
- if ((*__first2.__seg_ & __m) != (__b >> (__first1.__ctz_ + __ddn)))
- return false;
- __first2.__ctz_ = static_cast<unsigned>(__dn);
- }
- ++__first1.__seg_;
- // __first1.__ctz_ = 0;
- }
- // __first1.__ctz_ == 0;
- // do middle words
- unsigned __clz_r = __bits_per_word - __first2.__ctz_;
- __storage_type __m = ~__storage_type(0) << __first2.__ctz_;
- for (; __n >= __bits_per_word; __n -= __bits_per_word, ++__first1.__seg_) {
- __storage_type __b = *__first1.__seg_;
- if ((*__first2.__seg_ & __m) != (__b << __first2.__ctz_))
- return false;
- ++__first2.__seg_;
- if ((*__first2.__seg_ & ~__m) != (__b >> __clz_r))
- return false;
- }
- // do last word
- if (__n > 0) {
- __m = ~__storage_type(0) >> (__bits_per_word - __n);
- __storage_type __b = *__first1.__seg_ & __m;
- __storage_type __dn = std::min(__n, static_cast<difference_type>(__clz_r));
- __m = (~__storage_type(0) << __first2.__ctz_) & (~__storage_type(0) >> (__clz_r - __dn));
- if ((*__first2.__seg_ & __m) != (__b << __first2.__ctz_))
- return false;
- __first2.__seg_ += (__dn + __first2.__ctz_) / __bits_per_word;
- __first2.__ctz_ = static_cast<unsigned>((__dn + __first2.__ctz_) % __bits_per_word);
- __n -= __dn;
- if (__n > 0) {
- __m = ~__storage_type(0) >> (__bits_per_word - __n);
- if ((*__first2.__seg_ & __m) != (__b >> __dn))
- return false;
- }
- }
- }
- return true;
-}
-
-template <class _Cp, bool _IC1, bool _IC2>
-_LIBCPP_CONSTEXPR_SINCE_CXX20 _LIBCPP_HIDE_FROM_ABI bool __equal_aligned(
- __bit_iterator<_Cp, _IC1> __first1, __bit_iterator<_Cp, _IC1> __last1, __bit_iterator<_Cp, _IC2> __first2) {
- using _It = __bit_iterator<_Cp, _IC1>;
- using difference_type = typename _It::difference_type;
- using __storage_type = typename _It::__storage_type;
-
- const int __bits_per_word = _It::__bits_per_word;
- difference_type __n = __last1 - __first1;
- if (__n > 0) {
- // do first word
- if (__first1.__ctz_ != 0) {
- unsigned __clz = __bits_per_word - __first1.__ctz_;
- difference_type __dn = std::min(static_cast<difference_type>(__clz), __n);
- __n -= __dn;
- __storage_type __m = (~__storage_type(0) << __first1.__ctz_) & (~__storage_type(0) >> (__clz - __dn));
- if ((*__first2.__seg_ & __m) != (*__first1.__seg_ & __m))
- return false;
- ++__first2.__seg_;
- ++__first1.__seg_;
- // __first1.__ctz_ = 0;
- // __first2.__ctz_ = 0;
- }
- // __first1.__ctz_ == 0;
- // __first2.__ctz_ == 0;
- // do middle words
- for (; __n >= __bits_per_word; __n -= __bits_per_word, ++__first1.__seg_, ++__first2.__seg_)
- if (*__first2.__seg_ != *__first1.__seg_)
- return false;
- // do last word
- if (__n > 0) {
- __storage_type __m = ~__storage_type(0) >> (__bits_per_word - __n);
- if ((*__first2.__seg_ & __m) != (*__first1.__seg_ & __m))
- return false;
- }
- }
- return true;
-}
-
-template <class _Cp, bool _IC1, bool _IC2>
-inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 bool
-equal(__bit_iterator<_Cp, _IC1> __first1, __bit_iterator<_Cp, _IC1> __last1, __bit_iterator<_Cp, _IC2> __first2) {
- if (__first1.__ctz_ == __first2.__ctz_)
- return std::__equal_aligned(__first1, __last1, __first2);
- return std::__equal_unaligned(__first1, __last1, __first2);
-}
-
template <class _Cp, bool _IsConst, typename _Cp::__storage_type>
class __bit_iterator {
public:
@@ -1018,9 +901,13 @@ private:
template <class _Dp, bool _IC1, bool _IC2>
_LIBCPP_CONSTEXPR_SINCE_CXX20 friend bool
__equal_unaligned(__bit_iterator<_Dp, _IC1>, __bit_iterator<_Dp, _IC1>, __bit_iterator<_Dp, _IC2>);
- template <class _Dp, bool _IC1, bool _IC2>
- _LIBCPP_CONSTEXPR_SINCE_CXX20 friend bool
- equal(__bit_iterator<_Dp, _IC1>, __bit_iterator<_Dp, _IC1>, __bit_iterator<_Dp, _IC2>);
+ template <class _Dp,
+ bool _IC1,
+ bool _IC2,
+ class _BinaryPredicate,
+ __enable_if_t<std::is_same<_BinaryPredicate, __equal_to>::value, int> >
+ _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 friend bool __equal_iter_impl(
+ __bit_iterator<_Dp, _IC1>, __bit_iterator<_Dp, _IC1>, __bit_iterator<_Dp, _IC2>, _BinaryPredicate);
template <bool _ToFind, class _Dp, bool _IC>
_LIBCPP_CONSTEXPR_SINCE_CXX20 friend __bit_iterator<_Dp, _IC>
__find_bool(__bit_iterator<_Dp, _IC>, typename __size_difference_type_traits<_Dp>::size_type);
diff --git a/libcxx/include/bitset b/libcxx/include/bitset
index 10576eb80bf2ee..a8c499df04232f 100644
--- a/libcxx/include/bitset
+++ b/libcxx/include/bitset
@@ -130,6 +130,7 @@ template <size_t N> struct hash<std::bitset<N>>;
# include <__cxx03/bitset>
#else
# include <__algorithm/count.h>
+# include <__algorithm/equal.h>
# include <__algorithm/fill.h>
# include <__algorithm/fill_n.h>
# include <__algorithm/find.h>
diff --git a/libcxx/test/benchmarks/algorithms/equal.bench.cpp b/libcxx/test/benchmarks/algorithms/equal.bench.cpp
index 2dc11585c15c7f..ac3aa28bb28b39 100644
--- a/libcxx/test/benchmarks/algorithms/equal.bench.cpp
+++ b/libcxx/test/benchmarks/algorithms/equal.bench.cpp
@@ -45,4 +45,55 @@ static void bm_ranges_equal(benchmark::State& state) {
}
BENCHMARK(bm_ranges_equal)->DenseRange(1, 8)->Range(16, 1 << 20);
+static void bm_ranges_equal_vb_aligned(benchmark::State& state) {
+ auto n = state.range();
+ std::vector<bool> vec1(n, true);
+ std::vector<bool> vec2(n, true);
+ for (auto _ : state) {
+ benchmark::DoNotOptimize(std::ranges::equal(vec1, vec2));
+ benchmark::DoNotOptimize(&vec1);
+ benchmark::DoNotOptimize(&vec2);
+ }
+}
+
+static void bm_ranges_equal_vb_unaligned(benchmark::State& state) {
+ auto n = state.range();
+ std::vector<bool> vec1(n, true);
+ std::vector<bool> vec2(n + 8, true);
+ auto beg1 = std::ranges::begin(vec1);
+ auto end1 = std::ranges::end(vec1);
+ auto beg2 = std::ranges::begin(vec2) + 4;
+ auto end2 = std::ranges::end(vec2) - 4;
+ for (auto _ : state) {
+ benchmark::DoNotOptimize(std::ranges::equal(beg1, end1, beg2, end2));
+ benchmark::DoNotOptimize(&vec1);
+ benchmark::DoNotOptimize(&vec2);
+ }
+}
+
+// Test std::ranges::equal for vector<bool>::iterator
+BENCHMARK(bm_ranges_equal_vb_aligned)->Range(8, 1 << 20);
+BENCHMARK(bm_ranges_equal_vb_unaligned)->Range(8, 1 << 20);
+
+static void bm_equal_vb(benchmark::State& state, bool aligned) {
+ auto n = state.range();
+ std::vector<bool> vec1(n, true);
+ std::vector<bool> vec2(aligned ? n : n + 8, true);
+ auto beg1 = vec1.begin();
+ auto end1 = vec1.end();
+ auto beg2 = aligned ? vec2.begin() : vec2.begin() + 4;
+ for (auto _ : state) {
+ benchmark::DoNotOptimize(std::equal(beg1, end1, beg2));
+ benchmark::DoNotOptimize(&vec1);
+ benchmark::DoNotOptimize(&vec2);
+ }
+}
+
+static void bm_equal_vb_aligned(benchmark::State& state) { bm_equal_vb(state, true); }
+static void bm_equal_vb_unaligned(benchmark::State& state) { bm_equal_vb(state, false); }
+
+// Test std::equal for vector<bool>::iterator
+BENCHMARK(bm_equal_vb_aligned)->Range(8, 1 << 20);
+BENCHMARK(bm_equal_vb_unaligned)->Range(8, 1 << 20);
+
BENCHMARK_MAIN();
diff --git a/libcxx/test/std/algorithms/alg.nonmodifying/alg.equal/equal.pass.cpp b/libcxx/test/std/algorithms/alg.nonmodifying/alg.equal/equal.pass.cpp
index c3ba3f89b4de3c..a88f041013da6c 100644
--- a/libcxx/test/std/algorithms/alg.nonmodifying/alg.equal/equal.pass.cpp
+++ b/libcxx/test/std/algorithms/alg.nonmodifying/alg.equal/equal.pass.cpp
@@ -28,6 +28,7 @@
#include <algorithm>
#include <cassert>
#include <functional>
+#include <vector>
#include "test_iterators.h"
#include "test_macros.h"
@@ -123,6 +124,30 @@ class trivially_equality_comparable {
#endif
+template <std::size_t N>
+TEST_CONSTEXPR_CXX20 void test_vector_bool() {
+ std::vector<bool> in(N, false);
+ for (std::size_t i = 0; i < N; i += 2)
+ in[i] = true;
+
+ { // Test equal() with aligned bytes
+ std::vector<bool> out = in;
+ assert(std::equal(in.begin(), in.end(), out.begin()));
+#if TEST_STD_VER >= 14
+ assert(std::equal(in.begin(), in.end(), out.begin(), out.end()));
+#endif
+ }
+
+ { // Test equal() with unaligned bytes
+ std::vector<bool> out(N + 8);
+ std::copy(in.begin(), in.end(), out.begin() + 4);
+ assert(std::equal(in.begin(), in.end(), out.begin() + 4));
+#if TEST_STD_VER >= 14
+ assert(std::equal(in.begin(), in.end(), out.begin() + 4, out.end() - 4));
+#endif
+ }
+}
+
TEST_CONSTEXPR_CXX20 bool test() {
types::for_each(types::cpp17_input_iterator_list<int*>(), TestIter2<int, types::cpp17_input_iterator_list<int*> >());
types::for_each(
@@ -138,6 +163,14 @@ TEST_CONSTEXPR_CXX20 bool test() {
TestIter2<trivially_equality_comparable, types::cpp17_input_iterator_list<trivially_equality_comparable*>>{});
#endif
+ { // Test vector<bool>::iterator optimization
+ test_vector_bool<8>();
+ test_vector_bool<16>();
+ test_vector_bool<32>();
+ test_vector_bool<64>();
+ test_vector_bool<256>();
+ }
+
return true;
}
diff --git a/libcxx/test/std/algorithms/alg.nonmodifying/alg.equal/ranges.equal.pass.cpp b/libcxx/test/std/algorithms/alg.nonmodifying/alg.equal/ranges.equal.pass.cpp
index f36cd2e0896552..37c3677b445c00 100644
--- a/libcxx/test/std/algorithms/alg.nonmodifying/alg.equal/ra...
[truncated]
|
winner245
force-pushed
the
optimize-ranges-equal
branch
2 times, most recently
from
2b9442a to
c47c0d8
Compare
ldionne
approved these changes
Feb 5, 2025
winner245
force-pushed
the
optimize-ranges-equal
branch
5 times, most recently
from
2019761 to
a6ee161
Compare
|
I've updated the tests for the The purpose of this change is to ensure that this PR can be merged independently of the issue resolution. |
ldionne
requested changes
Feb 19, 2025
libcxx/test/std/algorithms/alg.modifying.operations/alg.fill/fill.pass.cpp
Outdated
Show resolved
Hide resolved
winner245
force-pushed
the
optimize-ranges-equal
branch
from
a6ee161 to
cb3dc20
Compare
winner245
force-pushed
the
optimize-ranges-equal
branch
from
cb3dc20 to
915e176
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR optimizes the performance of
std::ranges::equalforvector<bool>::iterator, addressing a subtask outlined in issue #64038. The optimizations yield performance improvements of up to 188x for aligned equality comparison and 82x for unaligned equality comparison. Moreover, comprehensive tests covering up to 4 storage words (256 bytes) with odd and even bit sizes are provided, which validate the proposed optimizations in this patch.