[libc++] Optimize ranges::swap_ranges for vector<bool>::iterator #121150
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-swap_ranges
branch
2 times, most recently
from
55c9ee7 to
4267694
Compare
winner245
force-pushed
the
optimize-swap_ranges
branch
3 times, most recently
from
b82d52e to
0bfcf34
Compare
llvmbot
added
the
libc++
|
@llvm/pr-subscribers-libcxx Author: Peng Liu (winner245) ChangesThis PR optimizes the performance of
Patch is 35.62 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/121150.diff 6 Files Affected:
diff --git a/libcxx/include/__algorithm/swap_ranges.h b/libcxx/include/__algorithm/swap_ranges.h
index 54b453b72360e0..2731d4feaa63d4 100644
--- a/libcxx/include/__algorithm/swap_ranges.h
+++ b/libcxx/include/__algorithm/swap_ranges.h
@@ -10,9 +10,12 @@
#define _LIBCPP___ALGORITHM_SWAP_RANGES_H
#include <__algorithm/iterator_operations.h>
+#include <__algorithm/min.h>
#include <__config>
+#include <__fwd/bit_reference.h>
#include <__utility/move.h>
#include <__utility/pair.h>
+#include <__utility/swap.h>
#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
# pragma GCC system_header
@@ -23,6 +26,165 @@ _LIBCPP_PUSH_MACROS
_LIBCPP_BEGIN_NAMESPACE_STD
+template <class _Cl, class _Cr>
+_LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 __bit_iterator<_Cr, false> __swap_ranges_aligned(
+ __bit_iterator<_Cl, false> __first, __bit_iterator<_Cl, false> __last, __bit_iterator<_Cr, false> __result) {
+ using _I1 = __bit_iterator<_Cl, false>;
+ using difference_type = typename _I1::difference_type;
+ using __storage_type = typename _I1::__storage_type;
+
+ const int __bits_per_word = _I1::__bits_per_word;
+ difference_type __n = __last - __first;
+ if (__n > 0) {
+ // do first word
+ if (__first.__ctz_ != 0) {
+ unsigned __clz = __bits_per_word - __first.__ctz_;
+ difference_type __dn = std::min(static_cast<difference_type>(__clz), __n);
+ __n -= __dn;
+ __storage_type __m = (~__storage_type(0) << __first.__ctz_) & (~__storage_type(0) >> (__clz - __dn));
+ __storage_type __b1 = *__first.__seg_ & __m;
+ *__first.__seg_ &= ~__m;
+ __storage_type __b2 = *__result.__seg_ & __m;
+ *__result.__seg_ &= ~__m;
+ *__result.__seg_ |= __b1;
+ *__first.__seg_ |= __b2;
+ __result.__seg_ += (__dn + __result.__ctz_) / __bits_per_word;
+ __result.__ctz_ = static_cast<unsigned>((__dn + __result.__ctz_) % __bits_per_word);
+ ++__first.__seg_;
+ // __first.__ctz_ = 0;
+ }
+ // __first.__ctz_ == 0;
+ // do middle words
+ for (; __n >= __bits_per_word; __n -= __bits_per_word, ++__first.__seg_, ++__result.__seg_)
+ swap(*__first.__seg_, *__result.__seg_);
+ // do last word
+ if (__n > 0) {
+ __storage_type __m = ~__storage_type(0) >> (__bits_per_word - __n);
+ __storage_type __b1 = *__first.__seg_ & __m;
+ *__first.__seg_ &= ~__m;
+ __storage_type __b2 = *__result.__seg_ & __m;
+ *__result.__seg_ &= ~__m;
+ *__result.__seg_ |= __b1;
+ *__first.__seg_ |= __b2;
+ __result.__ctz_ = static_cast<unsigned>(__n);
+ }
+ }
+ return __result;
+}
+
+template <class _Cl, class _Cr>
+_LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 __bit_iterator<_Cr, false> __swap_ranges_unaligned(
+ __bit_iterator<_Cl, false> __first, __bit_iterator<_Cl, false> __last, __bit_iterator<_Cr, false> __result) {
+ using _I1 = __bit_iterator<_Cl, false>;
+ using difference_type = typename _I1::difference_type;
+ using __storage_type = typename _I1::__storage_type;
+
+ const int __bits_per_word = _I1::__bits_per_word;
+ difference_type __n = __last - __first;
+ if (__n > 0) {
+ // do first word
+ if (__first.__ctz_ != 0) {
+ unsigned __clz_f = __bits_per_word - __first.__ctz_;
+ difference_type __dn = std::min(static_cast<difference_type>(__clz_f), __n);
+ __n -= __dn;
+ __storage_type __m = (~__storage_type(0) << __first.__ctz_) & (~__storage_type(0) >> (__clz_f - __dn));
+ __storage_type __b1 = *__first.__seg_ & __m;
+ *__first.__seg_ &= ~__m;
+ unsigned __clz_r = __bits_per_word - __result.__ctz_;
+ __storage_type __ddn = std::min<__storage_type>(__dn, __clz_r);
+ __m = (~__storage_type(0) << __result.__ctz_) & (~__storage_type(0) >> (__clz_r - __ddn));
+ __storage_type __b2 = *__result.__seg_ & __m;
+ *__result.__seg_ &= ~__m;
+ if (__result.__ctz_ > __first.__ctz_) {
+ unsigned __s = __result.__ctz_ - __first.__ctz_;
+ *__result.__seg_ |= __b1 << __s;
+ *__first.__seg_ |= __b2 >> __s;
+ } else {
+ unsigned __s = __first.__ctz_ - __result.__ctz_;
+ *__result.__seg_ |= __b1 >> __s;
+ *__first.__seg_ |= __b2 << __s;
+ }
+ __result.__seg_ += (__ddn + __result.__ctz_) / __bits_per_word;
+ __result.__ctz_ = static_cast<unsigned>((__ddn + __result.__ctz_) % __bits_per_word);
+ __dn -= __ddn;
+ if (__dn > 0) {
+ __m = ~__storage_type(0) >> (__bits_per_word - __dn);
+ __b2 = *__result.__seg_ & __m;
+ *__result.__seg_ &= ~__m;
+ unsigned __s = __first.__ctz_ + __ddn;
+ *__result.__seg_ |= __b1 >> __s;
+ *__first.__seg_ |= __b2 << __s;
+ __result.__ctz_ = static_cast<unsigned>(__dn);
+ }
+ ++__first.__seg_;
+ // __first.__ctz_ = 0;
+ }
+ // __first.__ctz_ == 0;
+ // do middle words
+ __storage_type __m = ~__storage_type(0) << __result.__ctz_;
+ unsigned __clz_r = __bits_per_word - __result.__ctz_;
+ for (; __n >= __bits_per_word; __n -= __bits_per_word, ++__first.__seg_) {
+ __storage_type __b1 = *__first.__seg_;
+ __storage_type __b2 = *__result.__seg_ & __m;
+ *__result.__seg_ &= ~__m;
+ *__result.__seg_ |= __b1 << __result.__ctz_;
+ *__first.__seg_ = __b2 >> __result.__ctz_;
+ ++__result.__seg_;
+ __b2 = *__result.__seg_ & ~__m;
+ *__result.__seg_ &= __m;
+ *__result.__seg_ |= __b1 >> __clz_r;
+ *__first.__seg_ |= __b2 << __clz_r;
+ }
+ // do last word
+ if (__n > 0) {
+ __m = ~__storage_type(0) >> (__bits_per_word - __n);
+ __storage_type __b1 = *__first.__seg_ & __m;
+ *__first.__seg_ &= ~__m;
+ __storage_type __dn = std::min<__storage_type>(__n, __clz_r);
+ __m = (~__storage_type(0) << __result.__ctz_) & (~__storage_type(0) >> (__clz_r - __dn));
+ __storage_type __b2 = *__result.__seg_ & __m;
+ *__result.__seg_ &= ~__m;
+ *__result.__seg_ |= __b1 << __result.__ctz_;
+ *__first.__seg_ |= __b2 >> __result.__ctz_;
+ __result.__seg_ += (__dn + __result.__ctz_) / __bits_per_word;
+ __result.__ctz_ = static_cast<unsigned>((__dn + __result.__ctz_) % __bits_per_word);
+ __n -= __dn;
+ if (__n > 0) {
+ __m = ~__storage_type(0) >> (__bits_per_word - __n);
+ __b2 = *__result.__seg_ & __m;
+ *__result.__seg_ &= ~__m;
+ *__result.__seg_ |= __b1 >> __dn;
+ *__first.__seg_ |= __b2 << __dn;
+ __result.__ctz_ = static_cast<unsigned>(__n);
+ }
+ }
+ }
+ return __result;
+}
+
+// 2+1 iterators: size2 >= size1; used by std::swap_ranges.
+template <class, class _Cl, class _Cr>
+_LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 pair<__bit_iterator<_Cl, false>, __bit_iterator<_Cr, false> >
+__swap_ranges(__bit_iterator<_Cl, false> __first1,
+ __bit_iterator<_Cl, false> __last1,
+ __bit_iterator<_Cr, false> __first2) {
+ if (__first1.__ctz_ == __first2.__ctz_)
+ return std::make_pair(__last1, std::__swap_ranges_aligned(__first1, __last1, __first2));
+ return std::make_pair(__last1, std::__swap_ranges_unaligned(__first1, __last1, __first2));
+}
+
+// 2+2 iterators: used by std::ranges::swap_ranges.
+template <class _AlgPolicy, class _Cl, class _Cr>
+_LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 pair<__bit_iterator<_Cl, false>, __bit_iterator<_Cr, false> >
+__swap_ranges(__bit_iterator<_Cl, false> __first1,
+ __bit_iterator<_Cl, false> __last1,
+ __bit_iterator<_Cr, false> __first2,
+ __bit_iterator<_Cr, false> __last2) {
+ if (__last1 - __first1 < __last2 - __first2)
+ return std::make_pair(__last1, std::__swap_ranges<_AlgPolicy>(__first1, __last1, __first2).second);
+ return std::make_pair(std::__swap_ranges<_AlgPolicy>(__first2, __last2, __first1).second, __last2);
+}
+
// 2+2 iterators: the shorter size will be used.
template <class _AlgPolicy, class _ForwardIterator1, class _Sentinel1, class _ForwardIterator2, class _Sentinel2>
_LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 pair<_ForwardIterator1, _ForwardIterator2>
diff --git a/libcxx/include/__bit_reference b/libcxx/include/__bit_reference
index 67abb023122edf..26dfb330384a36 100644
--- a/libcxx/include/__bit_reference
+++ b/libcxx/include/__bit_reference
@@ -12,6 +12,7 @@
#include <__algorithm/copy_n.h>
#include <__algorithm/min.h>
+#include <__algorithm/swap_ranges.h>
#include <__bit/countr.h>
#include <__compare/ordering.h>
#include <__config>
@@ -24,6 +25,7 @@
#include <__type_traits/conditional.h>
#include <__type_traits/is_constant_evaluated.h>
#include <__type_traits/void_t.h>
+#include <__utility/pair.h>
#include <__utility/swap.h>
#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -451,152 +453,6 @@ inline _LIBCPP_HIDE_FROM_ABI __bit_iterator<_Cp, false> move_backward(
return std::copy_backward(__first, __last, __result);
}
-// swap_ranges
-
-template <class _Cl, class _Cr>
-_LIBCPP_HIDE_FROM_ABI __bit_iterator<_Cr, false> __swap_ranges_aligned(
- __bit_iterator<_Cl, false> __first, __bit_iterator<_Cl, false> __last, __bit_iterator<_Cr, false> __result) {
- using _I1 = __bit_iterator<_Cl, false>;
- using difference_type = typename _I1::difference_type;
- using __storage_type = typename _I1::__storage_type;
-
- const int __bits_per_word = _I1::__bits_per_word;
- difference_type __n = __last - __first;
- if (__n > 0) {
- // do first word
- if (__first.__ctz_ != 0) {
- unsigned __clz = __bits_per_word - __first.__ctz_;
- difference_type __dn = std::min(static_cast<difference_type>(__clz), __n);
- __n -= __dn;
- __storage_type __m = (~__storage_type(0) << __first.__ctz_) & (~__storage_type(0) >> (__clz - __dn));
- __storage_type __b1 = *__first.__seg_ & __m;
- *__first.__seg_ &= ~__m;
- __storage_type __b2 = *__result.__seg_ & __m;
- *__result.__seg_ &= ~__m;
- *__result.__seg_ |= __b1;
- *__first.__seg_ |= __b2;
- __result.__seg_ += (__dn + __result.__ctz_) / __bits_per_word;
- __result.__ctz_ = static_cast<unsigned>((__dn + __result.__ctz_) % __bits_per_word);
- ++__first.__seg_;
- // __first.__ctz_ = 0;
- }
- // __first.__ctz_ == 0;
- // do middle words
- for (; __n >= __bits_per_word; __n -= __bits_per_word, ++__first.__seg_, ++__result.__seg_)
- swap(*__first.__seg_, *__result.__seg_);
- // do last word
- if (__n > 0) {
- __storage_type __m = ~__storage_type(0) >> (__bits_per_word - __n);
- __storage_type __b1 = *__first.__seg_ & __m;
- *__first.__seg_ &= ~__m;
- __storage_type __b2 = *__result.__seg_ & __m;
- *__result.__seg_ &= ~__m;
- *__result.__seg_ |= __b1;
- *__first.__seg_ |= __b2;
- __result.__ctz_ = static_cast<unsigned>(__n);
- }
- }
- return __result;
-}
-
-template <class _Cl, class _Cr>
-_LIBCPP_HIDE_FROM_ABI __bit_iterator<_Cr, false> __swap_ranges_unaligned(
- __bit_iterator<_Cl, false> __first, __bit_iterator<_Cl, false> __last, __bit_iterator<_Cr, false> __result) {
- using _I1 = __bit_iterator<_Cl, false>;
- using difference_type = typename _I1::difference_type;
- using __storage_type = typename _I1::__storage_type;
-
- const int __bits_per_word = _I1::__bits_per_word;
- difference_type __n = __last - __first;
- if (__n > 0) {
- // do first word
- if (__first.__ctz_ != 0) {
- unsigned __clz_f = __bits_per_word - __first.__ctz_;
- difference_type __dn = std::min(static_cast<difference_type>(__clz_f), __n);
- __n -= __dn;
- __storage_type __m = (~__storage_type(0) << __first.__ctz_) & (~__storage_type(0) >> (__clz_f - __dn));
- __storage_type __b1 = *__first.__seg_ & __m;
- *__first.__seg_ &= ~__m;
- unsigned __clz_r = __bits_per_word - __result.__ctz_;
- __storage_type __ddn = std::min<__storage_type>(__dn, __clz_r);
- __m = (~__storage_type(0) << __result.__ctz_) & (~__storage_type(0) >> (__clz_r - __ddn));
- __storage_type __b2 = *__result.__seg_ & __m;
- *__result.__seg_ &= ~__m;
- if (__result.__ctz_ > __first.__ctz_) {
- unsigned __s = __result.__ctz_ - __first.__ctz_;
- *__result.__seg_ |= __b1 << __s;
- *__first.__seg_ |= __b2 >> __s;
- } else {
- unsigned __s = __first.__ctz_ - __result.__ctz_;
- *__result.__seg_ |= __b1 >> __s;
- *__first.__seg_ |= __b2 << __s;
- }
- __result.__seg_ += (__ddn + __result.__ctz_) / __bits_per_word;
- __result.__ctz_ = static_cast<unsigned>((__ddn + __result.__ctz_) % __bits_per_word);
- __dn -= __ddn;
- if (__dn > 0) {
- __m = ~__storage_type(0) >> (__bits_per_word - __dn);
- __b2 = *__result.__seg_ & __m;
- *__result.__seg_ &= ~__m;
- unsigned __s = __first.__ctz_ + __ddn;
- *__result.__seg_ |= __b1 >> __s;
- *__first.__seg_ |= __b2 << __s;
- __result.__ctz_ = static_cast<unsigned>(__dn);
- }
- ++__first.__seg_;
- // __first.__ctz_ = 0;
- }
- // __first.__ctz_ == 0;
- // do middle words
- __storage_type __m = ~__storage_type(0) << __result.__ctz_;
- unsigned __clz_r = __bits_per_word - __result.__ctz_;
- for (; __n >= __bits_per_word; __n -= __bits_per_word, ++__first.__seg_) {
- __storage_type __b1 = *__first.__seg_;
- __storage_type __b2 = *__result.__seg_ & __m;
- *__result.__seg_ &= ~__m;
- *__result.__seg_ |= __b1 << __result.__ctz_;
- *__first.__seg_ = __b2 >> __result.__ctz_;
- ++__result.__seg_;
- __b2 = *__result.__seg_ & ~__m;
- *__result.__seg_ &= __m;
- *__result.__seg_ |= __b1 >> __clz_r;
- *__first.__seg_ |= __b2 << __clz_r;
- }
- // do last word
- if (__n > 0) {
- __m = ~__storage_type(0) >> (__bits_per_word - __n);
- __storage_type __b1 = *__first.__seg_ & __m;
- *__first.__seg_ &= ~__m;
- __storage_type __dn = std::min<__storage_type>(__n, __clz_r);
- __m = (~__storage_type(0) << __result.__ctz_) & (~__storage_type(0) >> (__clz_r - __dn));
- __storage_type __b2 = *__result.__seg_ & __m;
- *__result.__seg_ &= ~__m;
- *__result.__seg_ |= __b1 << __result.__ctz_;
- *__first.__seg_ |= __b2 >> __result.__ctz_;
- __result.__seg_ += (__dn + __result.__ctz_) / __bits_per_word;
- __result.__ctz_ = static_cast<unsigned>((__dn + __result.__ctz_) % __bits_per_word);
- __n -= __dn;
- if (__n > 0) {
- __m = ~__storage_type(0) >> (__bits_per_word - __n);
- __b2 = *__result.__seg_ & __m;
- *__result.__seg_ &= ~__m;
- *__result.__seg_ |= __b1 >> __dn;
- *__first.__seg_ |= __b2 << __dn;
- __result.__ctz_ = static_cast<unsigned>(__n);
- }
- }
- }
- return __result;
-}
-
-template <class _Cl, class _Cr>
-inline _LIBCPP_HIDE_FROM_ABI __bit_iterator<_Cr, false> swap_ranges(
- __bit_iterator<_Cl, false> __first1, __bit_iterator<_Cl, false> __last1, __bit_iterator<_Cr, false> __first2) {
- if (__first1.__ctz_ == __first2.__ctz_)
- return std::__swap_ranges_aligned(__first1, __last1, __first2);
- return std::__swap_ranges_unaligned(__first1, __last1, __first2);
-}
-
// rotate
template <class _Cp>
@@ -1001,14 +857,14 @@ private:
_LIBCPP_CONSTEXPR_SINCE_CXX20 friend __bit_iterator<_Dp, false>
copy_backward(__bit_iterator<_Dp, _IC> __first, __bit_iterator<_Dp, _IC> __last, __bit_iterator<_Dp, false> __result);
template <class _Cl, class _Cr>
- friend __bit_iterator<_Cr, false>
+ _LIBCPP_CONSTEXPR_SINCE_CXX20 friend __bit_iterator<_Cr, false>
__swap_ranges_aligned(__bit_iterator<_Cl, false>, __bit_iterator<_Cl, false>, __bit_iterator<_Cr, false>);
template <class _Cl, class _Cr>
- friend __bit_iterator<_Cr, false>
+ _LIBCPP_CONSTEXPR_SINCE_CXX20 friend __bit_iterator<_Cr, false>
__swap_ranges_unaligned(__bit_iterator<_Cl, false>, __bit_iterator<_Cl, false>, __bit_iterator<_Cr, false>);
- template <class _Cl, class _Cr>
- friend __bit_iterator<_Cr, false>
- swap_ranges(__bit_iterator<_Cl, false>, __bit_iterator<_Cl, false>, __bit_iterator<_Cr, false>);
+ template <class, class _Cl, class _Cr>
+ _LIBCPP_CONSTEXPR_SINCE_CXX20 friend pair<__bit_iterator<_Cl, false>, __bit_iterator<_Cr, false> >
+ __swap_ranges(__bit_iterator<_Cl, false>, __bit_iterator<_Cl, false>, __bit_iterator<_Cr, false>);
template <class _Dp>
_LIBCPP_CONSTEXPR_SINCE_CXX20 friend __bit_iterator<_Dp, false>
rotate(__bit_iterator<_Dp, false>, __bit_iterator<_Dp, false>, __bit_iterator<_Dp, false>);
diff --git a/libcxx/test/benchmarks/algorithms/swap_ranges.bench.cpp b/libcxx/test/benchmarks/algorithms/swap_ranges.bench.cpp
new file mode 100644
index 00000000000000..66629799ac5ee0
--- /dev/null
+++ b/libcxx/test/benchmarks/algorithms/swap_ranges.bench.cpp
@@ -0,0 +1,66 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+// UNSUPPORTED: c++03, c++11, c++14, c++17
+
+#include <algorithm>
+#include <benchmark/benchmark.h>
+#include <vector>
+
+static void bm_ranges_swap_ranges_vb_aligned(benchmark::State& state) {
+ auto n = state.range();
+ std::vector<bool> vec1(n, true);
+ std::vector<bool> vec2(n, false);
+ for (auto _ : state) {
+ benchmark::DoNotOptimize(std::ranges::swap_ranges(vec1, vec2));
+ benchmark::DoNotOptimize(&vec1);
+ benchmark::DoNotOptimize(&vec2);
+ }
+}
+
+static void bm_ranges_swap_ranges_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::swap_ranges(beg1, end1, beg2, end2));
+ benchmark::DoNotOptimize(&vec1);
+ benchmark::DoNotOptimize(&vec2);
+ }
+}
+
+// Test std::ranges::swap_ranges for vector<bool>::iterator
+BENCHMARK(bm_ranges_swap_ranges_vb_aligned)->Range(8, 1 << 20);
+BENCHMARK(bm_ranges_swap_ranges_vb_unaligned)->Range(8, 1 << 20);
+
+static void bm_swap_ranges_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::swap_ranges(beg1, end1, beg2));
+ benchmark::DoNotOptimize(&vec1);
+ benchmark::DoNotOptimize(&vec2);
+ }
+}
+
+static void bm_swap_ranges_vb_aligned(benchmark::State& state) { bm_swap_ranges_vb(state, true); }
+static void bm_swap_ranges_vb_unaligned(benchmark::State& state) { bm_swap_ranges_vb(state, false); }
+
+// Test std::swap_ranges for vector<bool>::iterator
+BENCHMARK(bm_swap_ranges_vb_aligned)->Range(8, 1 << 20);
+BENCHMARK(bm_swap_ranges_vb_unaligned)->Range(8, 1 << 20);
+
+BENCHMARK_MAIN();
diff --git a/libcxx/test/std/algorithms/alg.modifying.operations/alg.swap/iter_swap.pass.cpp b/libcxx/test/std/algorithms/alg.modifying.operations/alg.swap/iter_swap.pass.cpp
index b3a9f5fc259ef7..0394a48a0bb9a2 100644
--- a/libcxx/test/std/algorithms/alg.modifying.operations/alg.swap/iter_swap.pass.cpp
+++ b/libcxx/test/std/algorithms/alg.modifying.operations/alg.swap/iter_swap.pass.cpp
@@ -19,25 +19,23 @@
#include "test_macros.h"
#if TEST_STD_VER > 17
-constexpr bool test_swap_constexpr()
-{
- int i = 1;
- int j = 2;
- std::iter_swap(&i, &j);
- return i == 2 && j == 1;
+constexpr bool test_swap_constexpr() {
+ int i = 1;
+ int j = 2;
+ std::iter_swap(&i, &j);
+ return i == 2 && j == 1;
}
#endif // TEST_STD_VER > 17
-int main(int, char**)
-{
- int i = 1;
- int j = 2;
- std::iter_swap(&i, &j);
- assert(i == 2);
- assert(j == 1);
+int main(int, char**) {
+ int i = 1;
+...
[truncated]
|
winner245
force-pushed
the
optimize-swap_ranges
branch
from
423af4f to
61aaca6
Compare
ldionne
approved these changes
Feb 5, 2025
libcxx/test/std/algorithms/alg.modifying.operations/alg.swap/iter_swap.pass.cpp
Outdated
Show resolved
Hide resolved
winner245
force-pushed
the
optimize-swap_ranges
branch
from
61aaca6 to
facac24
Compare
winner245
force-pushed
the
optimize-swap_ranges
branch
from
facac24 to
ca03497
Compare
ldionne
approved these changes
Feb 26, 2025
winner245
force-pushed
the
optimize-swap_ranges
branch
from
ca03497 to
da8cb0a
Compare
jph-13
pushed a commit
to jph-13/llvm-project
that referenced
this pull request
Mar 21, 2025
…m#121150) This PR optimizes the performance of `std::ranges::swap_ranges` for `vector<bool>::iterator`, addressing a subtask outlined in issue llvm#64038. The optimizations yield performance improvements of up to **611x** for aligned range swap and **78x** for unaligned range swap comparison. Additionally, 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.
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::swap_rangesforvector<bool>::iterator, addressing a subtask outlined in issue #64038. The optimizations yield performance improvements of up to 611x for aligned range swap and 78x for unaligned range swap comparison. Additionally, 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.