Skip to content

Search for first duplicate in vectorized unique #5363

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Stephan T. Lavavej (StephanTLavavej) merged 3 commits into from
Apr 10, 2025
Merged

Search for first duplicate in vectorized unique #5363

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
8cf50ce
benchmark
AlexGuteniev Mar 25, 2025
92bb791
fixup
AlexGuteniev Mar 25, 2025
89baac9
no longer requiring that
AlexGuteniev Mar 25, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 4 additions & 1 deletion benchmarks/src/unique.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,7 @@
#include <algorithm>
#include <benchmark/benchmark.h>
#include <cstdint>
#include <numeric>
#include <random>
#include <type_traits>
#include <vector>
Expand All @@ -19,7 +20,9 @@ void u(benchmark::State& state) {
std::binomial_distribution<TD> dis(5);

std::vector<T, not_highly_aligned_allocator<T>> src(2552);
std::generate(src.begin(), src.end(), [&] { return static_cast<T>(dis(gen)); });
std::iota(src.begin(), src.begin() + 390, T{0});
std::iota(src.end() - 390, src.end(), T{0});
std::generate(src.begin() + 390, src.end() - 390, [&] { return static_cast<T>(dis(gen)); });

std::vector<T, not_highly_aligned_allocator<T>> v;
v.reserve(src.size());
Expand Down
10 changes: 8 additions & 2 deletions stl/src/vector_algorithms.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5187,8 +5187,6 @@ namespace {
// It is not possible to leave them untouched while keeping this optimization efficient.
// This should not be a problem though, as they should be either overwritten by the next step,
// or left in the removed range.
// 'remove' does not require any specific values,
// 'unique' needs the last element value to be preserved, as it will be loaded again.
for (; _Nx != _Size_h / _Ew; ++_Nx) {
// Inner loop needed for cases where the shuffle mask operates on element parts rather than whole
// elements; for whole elements there would be one iteration.
Expand Down Expand Up @@ -5476,6 +5474,8 @@ void* __stdcall __std_remove_8(void* _First, void* const _Last, const uint64_t _
}

void* __stdcall __std_unique_1(void* _First, void* _Last) noexcept {
_First = const_cast<void*>(__std_adjacent_find_1(_First, _Last));

if (_First == _Last) {
return _First;
}
Expand All @@ -5496,6 +5496,8 @@ void* __stdcall __std_unique_1(void* _First, void* _Last) noexcept {
}

void* __stdcall __std_unique_2(void* _First, void* _Last) noexcept {
_First = const_cast<void*>(__std_adjacent_find_2(_First, _Last));

if (_First == _Last) {
return _First;
}
Expand All @@ -5516,6 +5518,8 @@ void* __stdcall __std_unique_2(void* _First, void* _Last) noexcept {
}

void* __stdcall __std_unique_4(void* _First, void* _Last) noexcept {
_First = const_cast<void*>(__std_adjacent_find_4(_First, _Last));

if (_First == _Last) {
return _First;
}
Expand Down Expand Up @@ -5543,6 +5547,8 @@ void* __stdcall __std_unique_4(void* _First, void* _Last) noexcept {
}

void* __stdcall __std_unique_8(void* _First, void* _Last) noexcept {
_First = const_cast<void*>(__std_adjacent_find_8(_First, _Last));

if (_First == _Last) {
return _First;
}
Expand Down