[libc++] Optimize ranges::copy_backward for vector<bool>::iterator #121026
+321
−199
Conversation
This file contains 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-copy_backward
branch
2 times, most recently
from
ea1e6cc to
0093747
Compare
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
winner245
force-pushed
the
optimize-ranges-copy_backward
branch
5 times, most recently
from
3e3fd98 to
4173a6e
Compare
llvmbot
added
the
libc++
|
@llvm/pr-subscribers-libcxx Author: Peng Liu (winner245) ChangesAs a follow-up to #121013, this PR optimizes the performance of
Patch is 33.99 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/121026.diff 8 Files Affected:
diff --git a/libcxx/docs/ReleaseNotes/20.rst b/libcxx/docs/ReleaseNotes/20.rst
index c8a07fb8b73348..9c4e43899226f6 100644
--- a/libcxx/docs/ReleaseNotes/20.rst
+++ b/libcxx/docs/ReleaseNotes/20.rst
@@ -73,6 +73,9 @@ Improvements and New Features
optimized, resulting in a performance improvement of up to 2x for trivial element types (e.g., `std::vector<int>`),
and up to 3.4x for non-trivial element types (e.g., `std::vector<std::vector<int>>`).
+- The ``std::ranges::copy_backward`` algorithm has been optimized for ``std::vector<bool>::iterator``\s, resulting in
+ a performance improvement of up to 2000x.
+
Deprecations and Removals
-------------------------
diff --git a/libcxx/include/__algorithm/copy_backward.h b/libcxx/include/__algorithm/copy_backward.h
index 48a768f577f553..02ffc14361e6a9 100644
--- a/libcxx/include/__algorithm/copy_backward.h
+++ b/libcxx/include/__algorithm/copy_backward.h
@@ -10,11 +10,14 @@
#define _LIBCPP___ALGORITHM_COPY_BACKWARD_H
#include <__algorithm/copy_move_common.h>
+#include <__algorithm/copy_n.h>
#include <__algorithm/iterator_operations.h>
#include <__algorithm/min.h>
#include <__config>
+#include <__fwd/bit_reference.h>
#include <__iterator/iterator_traits.h>
#include <__iterator/segmented_iterator.h>
+#include <__memory/pointer_traits.h>
#include <__type_traits/common_type.h>
#include <__type_traits/enable_if.h>
#include <__type_traits/is_constructible.h>
@@ -34,6 +37,124 @@ template <class _AlgPolicy, class _InIter, class _Sent, class _OutIter>
_LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 pair<_InIter, _OutIter>
__copy_backward(_InIter __first, _Sent __last, _OutIter __result);
+template <class _Cp, bool _IsConst>
+_LIBCPP_CONSTEXPR_SINCE_CXX20 _LIBCPP_HIDE_FROM_ABI __bit_iterator<_Cp, false> __copy_backward_aligned(
+ __bit_iterator<_Cp, _IsConst> __first, __bit_iterator<_Cp, _IsConst> __last, __bit_iterator<_Cp, false> __result) {
+ using _In = __bit_iterator<_Cp, _IsConst>;
+ using difference_type = typename _In::difference_type;
+ using __storage_type = typename _In::__storage_type;
+
+ const int __bits_per_word = _In::__bits_per_word;
+ difference_type __n = __last - __first;
+ if (__n > 0) {
+ // do first word
+ if (__last.__ctz_ != 0) {
+ difference_type __dn = std::min(static_cast<difference_type>(__last.__ctz_), __n);
+ __n -= __dn;
+ unsigned __clz = __bits_per_word - __last.__ctz_;
+ __storage_type __m = (~__storage_type(0) << (__last.__ctz_ - __dn)) & (~__storage_type(0) >> __clz);
+ __storage_type __b = *__last.__seg_ & __m;
+ *__result.__seg_ &= ~__m;
+ *__result.__seg_ |= __b;
+ __result.__ctz_ = static_cast<unsigned>(((-__dn & (__bits_per_word - 1)) + __result.__ctz_) % __bits_per_word);
+ // __last.__ctz_ = 0
+ }
+ // __last.__ctz_ == 0 || __n == 0
+ // __result.__ctz_ == 0 || __n == 0
+ // do middle words
+ __storage_type __nw = __n / __bits_per_word;
+ __result.__seg_ -= __nw;
+ __last.__seg_ -= __nw;
+ std::copy_n(std::__to_address(__last.__seg_), __nw, std::__to_address(__result.__seg_));
+ __n -= __nw * __bits_per_word;
+ // do last word
+ if (__n > 0) {
+ __storage_type __m = ~__storage_type(0) << (__bits_per_word - __n);
+ __storage_type __b = *--__last.__seg_ & __m;
+ *--__result.__seg_ &= ~__m;
+ *__result.__seg_ |= __b;
+ __result.__ctz_ = static_cast<unsigned>(-__n & (__bits_per_word - 1));
+ }
+ }
+ return __result;
+}
+
+template <class _Cp, bool _IsConst>
+_LIBCPP_CONSTEXPR_SINCE_CXX20 _LIBCPP_HIDE_FROM_ABI __bit_iterator<_Cp, false> __copy_backward_unaligned(
+ __bit_iterator<_Cp, _IsConst> __first, __bit_iterator<_Cp, _IsConst> __last, __bit_iterator<_Cp, false> __result) {
+ using _In = __bit_iterator<_Cp, _IsConst>;
+ using difference_type = typename _In::difference_type;
+ using __storage_type = typename _In::__storage_type;
+
+ const int __bits_per_word = _In::__bits_per_word;
+ difference_type __n = __last - __first;
+ if (__n > 0) {
+ // do first word
+ if (__last.__ctz_ != 0) {
+ difference_type __dn = std::min(static_cast<difference_type>(__last.__ctz_), __n);
+ __n -= __dn;
+ unsigned __clz_l = __bits_per_word - __last.__ctz_;
+ __storage_type __m = (~__storage_type(0) << (__last.__ctz_ - __dn)) & (~__storage_type(0) >> __clz_l);
+ __storage_type __b = *__last.__seg_ & __m;
+ unsigned __clz_r = __bits_per_word - __result.__ctz_;
+ __storage_type __ddn = std::min(__dn, static_cast<difference_type>(__result.__ctz_));
+ if (__ddn > 0) {
+ __m = (~__storage_type(0) << (__result.__ctz_ - __ddn)) & (~__storage_type(0) >> __clz_r);
+ *__result.__seg_ &= ~__m;
+ if (__result.__ctz_ > __last.__ctz_)
+ *__result.__seg_ |= __b << (__result.__ctz_ - __last.__ctz_);
+ else
+ *__result.__seg_ |= __b >> (__last.__ctz_ - __result.__ctz_);
+ __result.__ctz_ = static_cast<unsigned>(((-__ddn & (__bits_per_word - 1)) + __result.__ctz_) % __bits_per_word);
+ __dn -= __ddn;
+ }
+ if (__dn > 0) {
+ // __result.__ctz_ == 0
+ --__result.__seg_;
+ __result.__ctz_ = static_cast<unsigned>(-__dn & (__bits_per_word - 1));
+ __m = ~__storage_type(0) << __result.__ctz_;
+ *__result.__seg_ &= ~__m;
+ __last.__ctz_ -= __dn + __ddn;
+ *__result.__seg_ |= __b << (__result.__ctz_ - __last.__ctz_);
+ }
+ // __last.__ctz_ = 0
+ }
+ // __last.__ctz_ == 0 || __n == 0
+ // __result.__ctz_ != 0 || __n == 0
+ // do middle words
+ unsigned __clz_r = __bits_per_word - __result.__ctz_;
+ __storage_type __m = ~__storage_type(0) >> __clz_r;
+ for (; __n >= __bits_per_word; __n -= __bits_per_word) {
+ __storage_type __b = *--__last.__seg_;
+ *__result.__seg_ &= ~__m;
+ *__result.__seg_ |= __b >> __clz_r;
+ *--__result.__seg_ &= __m;
+ *__result.__seg_ |= __b << __result.__ctz_;
+ }
+ // do last word
+ if (__n > 0) {
+ __m = ~__storage_type(0) << (__bits_per_word - __n);
+ __storage_type __b = *--__last.__seg_ & __m;
+ __clz_r = __bits_per_word - __result.__ctz_;
+ __storage_type __dn = std::min(__n, static_cast<difference_type>(__result.__ctz_));
+ __m = (~__storage_type(0) << (__result.__ctz_ - __dn)) & (~__storage_type(0) >> __clz_r);
+ *__result.__seg_ &= ~__m;
+ *__result.__seg_ |= __b >> (__bits_per_word - __result.__ctz_);
+ __result.__ctz_ = static_cast<unsigned>(((-__dn & (__bits_per_word - 1)) + __result.__ctz_) % __bits_per_word);
+ __n -= __dn;
+ if (__n > 0) {
+ // __result.__ctz_ == 0
+ --__result.__seg_;
+ __result.__ctz_ = static_cast<unsigned>(-__n & (__bits_per_word - 1));
+ __m = ~__storage_type(0) << __result.__ctz_;
+ *__result.__seg_ &= ~__m;
+ *__result.__seg_ |= __b << (__result.__ctz_ - (__bits_per_word - __n - __dn));
+ }
+ }
+ }
+ return __result;
+}
+
template <class _AlgPolicy>
struct __copy_backward_impl {
template <class _InIter, class _Sent, class _OutIter>
@@ -107,6 +228,16 @@ struct __copy_backward_impl {
}
}
+ template <class _Cp, bool _IsConst>
+ _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 pair<__bit_iterator<_Cp, _IsConst>, __bit_iterator<_Cp, false> >
+ operator()(__bit_iterator<_Cp, _IsConst> __first,
+ __bit_iterator<_Cp, _IsConst> __last,
+ __bit_iterator<_Cp, false> __result) {
+ if (__last.__ctz_ == __result.__ctz_)
+ return std::make_pair(__last, std::__copy_backward_aligned(__first, __last, __result));
+ return std::make_pair(__last, std::__copy_backward_unaligned(__first, __last, __result));
+ }
+
// At this point, the iterators have been unwrapped so any `contiguous_iterator` has been unwrapped to a pointer.
template <class _In, class _Out, __enable_if_t<__can_lower_copy_assignment_to_memmove<_In, _Out>::value, int> = 0>
_LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX14 pair<_In*, _Out*>
diff --git a/libcxx/include/__bit_reference b/libcxx/include/__bit_reference
index 9fa24c98d493fd..1d9294993201fd 100644
--- a/libcxx/include/__bit_reference
+++ b/libcxx/include/__bit_reference
@@ -10,6 +10,7 @@
#ifndef _LIBCPP___BIT_REFERENCE
#define _LIBCPP___BIT_REFERENCE
+#include <__algorithm/copy_backward.h>
#include <__algorithm/copy_n.h>
#include <__algorithm/min.h>
#include <__bit/countr.h>
@@ -293,134 +294,6 @@ copy(__bit_iterator<_Cp, _IsConst> __first, __bit_iterator<_Cp, _IsConst> __last
return std::__copy_unaligned(__first, __last, __result);
}
-// copy_backward
-
-template <class _Cp, bool _IsConst>
-_LIBCPP_CONSTEXPR_SINCE_CXX20 _LIBCPP_HIDE_FROM_ABI __bit_iterator<_Cp, false> __copy_backward_aligned(
- __bit_iterator<_Cp, _IsConst> __first, __bit_iterator<_Cp, _IsConst> __last, __bit_iterator<_Cp, false> __result) {
- using _In = __bit_iterator<_Cp, _IsConst>;
- using difference_type = typename _In::difference_type;
- using __storage_type = typename _In::__storage_type;
-
- const int __bits_per_word = _In::__bits_per_word;
- difference_type __n = __last - __first;
- if (__n > 0) {
- // do first word
- if (__last.__ctz_ != 0) {
- difference_type __dn = std::min(static_cast<difference_type>(__last.__ctz_), __n);
- __n -= __dn;
- unsigned __clz = __bits_per_word - __last.__ctz_;
- __storage_type __m = (~__storage_type(0) << (__last.__ctz_ - __dn)) & (~__storage_type(0) >> __clz);
- __storage_type __b = *__last.__seg_ & __m;
- *__result.__seg_ &= ~__m;
- *__result.__seg_ |= __b;
- __result.__ctz_ = static_cast<unsigned>(((-__dn & (__bits_per_word - 1)) + __result.__ctz_) % __bits_per_word);
- // __last.__ctz_ = 0
- }
- // __last.__ctz_ == 0 || __n == 0
- // __result.__ctz_ == 0 || __n == 0
- // do middle words
- __storage_type __nw = __n / __bits_per_word;
- __result.__seg_ -= __nw;
- __last.__seg_ -= __nw;
- std::copy_n(std::__to_address(__last.__seg_), __nw, std::__to_address(__result.__seg_));
- __n -= __nw * __bits_per_word;
- // do last word
- if (__n > 0) {
- __storage_type __m = ~__storage_type(0) << (__bits_per_word - __n);
- __storage_type __b = *--__last.__seg_ & __m;
- *--__result.__seg_ &= ~__m;
- *__result.__seg_ |= __b;
- __result.__ctz_ = static_cast<unsigned>(-__n & (__bits_per_word - 1));
- }
- }
- return __result;
-}
-
-template <class _Cp, bool _IsConst>
-_LIBCPP_CONSTEXPR_SINCE_CXX20 _LIBCPP_HIDE_FROM_ABI __bit_iterator<_Cp, false> __copy_backward_unaligned(
- __bit_iterator<_Cp, _IsConst> __first, __bit_iterator<_Cp, _IsConst> __last, __bit_iterator<_Cp, false> __result) {
- using _In = __bit_iterator<_Cp, _IsConst>;
- using difference_type = typename _In::difference_type;
- using __storage_type = typename _In::__storage_type;
-
- const int __bits_per_word = _In::__bits_per_word;
- difference_type __n = __last - __first;
- if (__n > 0) {
- // do first word
- if (__last.__ctz_ != 0) {
- difference_type __dn = std::min(static_cast<difference_type>(__last.__ctz_), __n);
- __n -= __dn;
- unsigned __clz_l = __bits_per_word - __last.__ctz_;
- __storage_type __m = (~__storage_type(0) << (__last.__ctz_ - __dn)) & (~__storage_type(0) >> __clz_l);
- __storage_type __b = *__last.__seg_ & __m;
- unsigned __clz_r = __bits_per_word - __result.__ctz_;
- __storage_type __ddn = std::min(__dn, static_cast<difference_type>(__result.__ctz_));
- if (__ddn > 0) {
- __m = (~__storage_type(0) << (__result.__ctz_ - __ddn)) & (~__storage_type(0) >> __clz_r);
- *__result.__seg_ &= ~__m;
- if (__result.__ctz_ > __last.__ctz_)
- *__result.__seg_ |= __b << (__result.__ctz_ - __last.__ctz_);
- else
- *__result.__seg_ |= __b >> (__last.__ctz_ - __result.__ctz_);
- __result.__ctz_ = static_cast<unsigned>(((-__ddn & (__bits_per_word - 1)) + __result.__ctz_) % __bits_per_word);
- __dn -= __ddn;
- }
- if (__dn > 0) {
- // __result.__ctz_ == 0
- --__result.__seg_;
- __result.__ctz_ = static_cast<unsigned>(-__dn & (__bits_per_word - 1));
- __m = ~__storage_type(0) << __result.__ctz_;
- *__result.__seg_ &= ~__m;
- __last.__ctz_ -= __dn + __ddn;
- *__result.__seg_ |= __b << (__result.__ctz_ - __last.__ctz_);
- }
- // __last.__ctz_ = 0
- }
- // __last.__ctz_ == 0 || __n == 0
- // __result.__ctz_ != 0 || __n == 0
- // do middle words
- unsigned __clz_r = __bits_per_word - __result.__ctz_;
- __storage_type __m = ~__storage_type(0) >> __clz_r;
- for (; __n >= __bits_per_word; __n -= __bits_per_word) {
- __storage_type __b = *--__last.__seg_;
- *__result.__seg_ &= ~__m;
- *__result.__seg_ |= __b >> __clz_r;
- *--__result.__seg_ &= __m;
- *__result.__seg_ |= __b << __result.__ctz_;
- }
- // do last word
- if (__n > 0) {
- __m = ~__storage_type(0) << (__bits_per_word - __n);
- __storage_type __b = *--__last.__seg_ & __m;
- __clz_r = __bits_per_word - __result.__ctz_;
- __storage_type __dn = std::min(__n, static_cast<difference_type>(__result.__ctz_));
- __m = (~__storage_type(0) << (__result.__ctz_ - __dn)) & (~__storage_type(0) >> __clz_r);
- *__result.__seg_ &= ~__m;
- *__result.__seg_ |= __b >> (__bits_per_word - __result.__ctz_);
- __result.__ctz_ = static_cast<unsigned>(((-__dn & (__bits_per_word - 1)) + __result.__ctz_) % __bits_per_word);
- __n -= __dn;
- if (__n > 0) {
- // __result.__ctz_ == 0
- --__result.__seg_;
- __result.__ctz_ = static_cast<unsigned>(-__n & (__bits_per_word - 1));
- __m = ~__storage_type(0) << __result.__ctz_;
- *__result.__seg_ &= ~__m;
- *__result.__seg_ |= __b << (__result.__ctz_ - (__bits_per_word - __n - __dn));
- }
- }
- }
- return __result;
-}
-
-template <class _Cp, bool _IsConst>
-inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 __bit_iterator<_Cp, false> copy_backward(
- __bit_iterator<_Cp, _IsConst> __first, __bit_iterator<_Cp, _IsConst> __last, __bit_iterator<_Cp, false> __result) {
- if (__last.__ctz_ == __result.__ctz_)
- return std::__copy_backward_aligned(__first, __last, __result);
- return std::__copy_backward_unaligned(__first, __last, __result);
-}
-
// move
template <class _Cp, bool _IsConst>
@@ -983,9 +856,8 @@ private:
template <class _Dp, bool _IC>
_LIBCPP_CONSTEXPR_SINCE_CXX20 friend __bit_iterator<_Dp, false> __copy_backward_unaligned(
__bit_iterator<_Dp, _IC> __first, __bit_iterator<_Dp, _IC> __last, __bit_iterator<_Dp, false> __result);
- template <class _Dp, bool _IC>
- _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 _AlgPolicy>
+ friend struct __copy_backward_impl;
template <class _Cl, class _Cr>
friend __bit_iterator<_Cr, false>
__swap_ranges_aligned(__bit_iterator<_Cl, false>, __bit_iterator<_Cl, false>, __bit_iterator<_Cr, false>);
diff --git a/libcxx/include/__vector/vector_bool.h b/libcxx/include/__vector/vector_bool.h
index 525fc35b26cc9e..b683ccf21e8479 100644
--- a/libcxx/include/__vector/vector_bool.h
+++ b/libcxx/include/__vector/vector_bool.h
@@ -10,6 +10,7 @@
#define _LIBCPP___VECTOR_VECTOR_BOOL_H
#include <__algorithm/copy.h>
+#include <__algorithm/copy_backward.h>
#include <__algorithm/fill_n.h>
#include <__algorithm/iterator_operations.h>
#include <__algorithm/max.h>
diff --git a/libcxx/include/bitset b/libcxx/include/bitset
index 8b361824805571..bb09bc5415fb03 100644
--- a/libcxx/include/bitset
+++ b/libcxx/include/bitset
@@ -129,6 +129,8 @@ template <size_t N> struct hash<std::bitset<N>>;
#if __cplusplus < 201103L && defined(_LIBCPP_USE_FROZEN_CXX03_HEADERS)
# include <__cxx03/bitset>
#else
+# include <__algorithm/copy.h>
+# include <__algorithm/copy_backward.h>
# include <__algorithm/count.h>
# include <__algorithm/fill.h>
# include <__algorithm/fill_n.h>
diff --git a/libcxx/test/benchmarks/algorithms/copy_backward.bench.cpp b/libcxx/test/benchmarks/algorithms/copy_backward.bench.cpp
new file mode 100644
index 00000000000000..3f797b5555a138
--- /dev/null
+++ b/libcxx/test/benchmarks/algorithms/copy_backward.bench.cpp
@@ -0,0 +1,55 @@
+//===----------------------------------------------------------------------===//
+//
+// 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, c++20
+
+#include <algorithm>
+#include <benchmark/benchmark.h>
+#include <vector>
+
+static void bm_ranges_copy_backward(benchmark::State& state, bool aligned) {
+ auto n = state.range();
+ std::vector<bool> in(n, true);
+ std::vector<bool> out(aligned ? n : n + 8);
+ benchmark::DoNotOptimize(&in);
+ auto dst = aligned ? out.end() : out.end() - 4;
+ for (auto _ : state) {
+ benchmark::DoNotOptimize(std::ranges::copy_backward(in, dst));
+ benchmark::DoNotOptimize(&out);
+ }
+}
+
+static void bm_copy_backward(benchmark::State& state, bool aligned) {
+ auto n = state.range();
+ std::vector<bool> in(n, true);
+ std::vector<bool> out(aligned ? n : n + 8);
+ benchmark::DoNotOptimize(&in);
+ auto beg = in.begin();
+ auto end = in.end();
+ auto dst = aligned ? out.end() : out.end() - 4;
+ for (auto _ : state) {
+ benchmark::DoNotOptimize(std::copy_backward(beg, end, dst));
+ benchmark::DoNotOptimize(&out);
+ }
+}
+
+static void bm_ranges_copy_backward_aligned(benchmark::State& state) { bm_ranges_copy_backward(state, true); }
+static void bm_ranges_copy_backward_unaligned(benchmark::State& state) { bm_ranges_copy_backward(state, false); }
+
+static void bm_copy_backward_aligned(benchmark::State& state) { bm_copy_backward(state, true); }
+static void bm_copy_backward_unaligned(benchmark::State& state) { bm_copy_backward(state, false); }
+
+// Test the range version of std::copy for vector<bool>::iterator
+BENCHMARK(bm_ranges_copy_backward_aligned)->Range(8, 1 << 16)->DenseRange(102400, 204800, 4096);
+BENCHMARK(bm_ranges_copy_backward_unaligned)->Range(8, 1 << 20);
+
+// Test the iterator-pair version of std::copy for vector<bool>::iterator
+BENCHMARK(bm_copy_backward_aligned)->Range(8, 1 << 20);
+BENCHMARK(bm_copy_backward_unaligned)->Range(8, 1 << 20);
+
+BENCHMARK_MAIN();
diff --git a/libcxx/test/std/algorithms/alg.modifying.operations/alg.copy/copy_backward.pass.cpp b/libcxx/test/std/algorithms/alg.modifying.operations/alg.copy/copy_backward.pass.cpp
index 928903de1ade2d..c4d064897e9e0b 100644
--- a/libcxx/test/std/algorithms/alg.modifying.operations/alg.copy/copy_backward.pass.cpp
+++ b/libcxx/test/std/algorithms/alg.modifying.operations/alg.copy/copy_backward.pass.cpp
@@ -15,6 +15,7 @@
#include <algorithm>
#include <cassert>
+#include <vector>
#include "test_macros.h"
#include "test_iterators.h"
@@ -36,47 +37,66 @@ class Derived : public PaddedBase {
};
template <class InIter, class OutIter>
-TEST_CONSTEXPR_CXX20 void
-test_copy_backward()
-{
+TEST_CONSTEXPR_CXX20 void test_copy_backward() {
{
const unsigned N = 1000;
- int ia[N] = {};
+ int ia[N] = {};
for (unsigned i = 0; i < N; ++i)
- ia[i] = i;
+ ia[i] = i;
int ib[N] = {0};
- OutIter r = std::copy_backward(InIter(ia), InIter(ia+N), OutIter(ib+N));
+ OutIter r = std::copy_backward(InIter(ia), InIter(ia + N), OutIter(ib + N));
assert(base(r) == ib);
for (unsigned i = 0; i < N; ++i)
- assert(ia[i] == ib[i]);
+ assert(ia[i] == ib[i]);
}
}
-TEST_CONSTEXPR_CXX20 bool
-test()
-{
- test_copy_backward<bidirectional_iterator<const int*>, bidirectional_iterator<int*> >();
- test_copy_backward<bidirectional_iterator<const int*>, random_access_iterator<int...
[truncated]
|
winner245
force-pushed
the
optimize-ranges-copy_backward
branch
from
4173a6e to
9c3a750
Compare
winner245
force-pushed
the
optimize-ranges-copy_backward
branch
4 times, most recently
from
d3bd7df to
8b316b2
Compare
ldionne
approved these changes
Jan 29, 2025
libcxx/docs/ReleaseNotes/20.rst
Outdated
| @@ -111,6 +111,9 @@ Improvements and New Features | |||
| std::errc::not_a_directory``, or use ``err.default_error_condition()`` to map to an ``error_condition``, and then test | |||
| its ``value()`` and ``category()``. | |||
|
|
|||
| - The ``std::ranges::copy_backward`` algorithm has been optimized for ``std::vector<bool>::iterator``\s, resulting in | |||
libcxx/include/__bit_reference
Outdated
Comment on lines
873
to
874
| // Note: dependent nested name specifier __copy_backward_impl<_AlgPolicy>::operator() for friend declaration | ||
| // is not supported in clang. Thus, we use a friend declaration for the entire class. |
There was a problem hiding this comment.
Suggested change
| // Note: dependent nested name specifier __copy_backward_impl<_AlgPolicy>::operator() for friend declaration | |
| // is not supported in clang. Thus, we use a friend declaration for the entire class. |
I don't think this note adds enough value to pull its weight, I would leave it out.
| @@ -96,15 +113,22 @@ test() | |||
| assert(std::equal(a, a + 10, expected)); | |||
| } | |||
|
|
|||
| return true; | |||
| { // Test vector<bool>::iterator optimization | |||
| assert(test_vector_bool(8)); | |||
| @@ -321,6 +345,16 @@ constexpr bool test() { | |||
| } | |||
| } | |||
|
|
|||
| #if TEST_STD_VER >= 23 | |||
| { // Test vector<bool>::iterator optimization | |||
| assert(test_vector_bool(8)); | |||
winner245
force-pushed
the
optimize-ranges-copy_backward
branch
from
8b316b2 to
70b1d47
Compare
winner245
force-pushed
the
optimize-ranges-copy_backward
branch
from
70b1d47 to
5132d25
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.
As a follow-up to #121013 (which focused on
std::ranges::copy), this PR optimizes the performance ofstd::ranges::copy_backwardforvector<bool>::iterator, addressing a subtask outlined in issue #64038. The optimizations yield performance improvements of up to 2000x for aligned copies and 60x for unaligned copies.