Abseil Common Libraries (C++) (grcp 依赖) https://abseil.io/
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

968 lines
33 KiB

Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
// Copyright 2019 The Abseil Authors.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// https://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#ifndef ABSL_CONTAINER_INTERNAL_INLINED_VECTOR_INTERNAL_H_
#define ABSL_CONTAINER_INTERNAL_INLINED_VECTOR_INTERNAL_H_
#include <algorithm>
#include <cstddef>
#include <cstring>
#include <iterator>
#include <limits>
#include <memory>
#include <utility>
#include "absl/base/macros.h"
#include "absl/container/internal/compressed_tuple.h"
#include "absl/memory/memory.h"
#include "absl/meta/type_traits.h"
#include "absl/types/span.h"
namespace absl {
ABSL_NAMESPACE_BEGIN
namespace inlined_vector_internal {
Export of internal Abseil changes -- 0bfa836596a9c787a2f0bdc283011dd1f6810c6e by Benjamin Barenblat <bbaren@google.com>: Ignore missing CPU frequency on more architectures Linux on MIPS, PA-RISC, RISC-V, and SystemZ doesn’t expose the nominal CPU frequency via /sys, so don’t worry if `NominalCPUFrequency` returns 1.0 on those platforms. Some POWER machines expose the CPU frequency; others do not. Since we can’t predict which type of machine the tests will run on, simply disable testing for `NominalCPUFrequency` on POWER. PiperOrigin-RevId: 347079873 -- 492b6834ed4a07cbc3abccd846f7e37d8c556ee5 by Benjamin Barenblat <bbaren@google.com>: Use ABSL_HAVE_THREAD_LOCAL macro instead of copying code Reduce code duplication by checking the ABSL_HAVE_THREAD_LOCAL macro instead of copying code from base/config.h. PiperOrigin-RevId: 347079561 -- 8d656efce4da9cb032094377e58493d98427a536 by Abseil Team <absl-team@google.com>: Rollback PiperOrigin-RevId: 347078779 -- 221bc69ec6dd7e2777ffcff6942584f979ef6382 by Abseil Team <absl-team@google.com>: Add flag for 'shallow subcord' feature for experimental ring buffer rollout There is a potential trade-off of CPU cost vs over-sharing cord data for subcord of large cords. This flag allows making subcords shallow for ringbuffers (with a potential larger waste of referenced source cords), which allows us to make subcord fast for this apps that do no persist (unmodified / plain copied) sub cords. This change also introduces constants for the default settings, intended to keep the internal cord settings concistent with external flags. PiperOrigin-RevId: 347053271 -- 00a56c24293566734009f6bf2169a83fb37a35ba by Abseil Team <absl-team@google.com>: Revert the usage of variant<> in Cord iterator and reader. The introduction of the variant may lead to some missed compiler optimizations. PiperOrigin-RevId: 347053041 -- c7b7b5ed7e3ab46b1e75b80f1a7de0bda26c8f70 by Chris Kennelly <ckennelly@google.com>: Release library for integer power-of-2 functions and bit counting. PiperOrigin-RevId: 347035065 -- 5a035c0d9840b251967f9e7039fc6a4e01dd52f3 by Abseil Team <absl-team@google.com>: Restructure Cord::ChunkIterator for future ring buffer support. PiperOrigin-RevId: 346890054 GitOrigin-RevId: 0bfa836596a9c787a2f0bdc283011dd1f6810c6e Change-Id: I3a58e2a44cb4c6f2116c43e2a4ccbc319d3ccecf
4 years ago
// GCC does not deal very well with the below code
#if !defined(__clang__) && defined(__GNUC__)
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wmaybe-uninitialized"
#endif
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
template <typename Iterator>
using IsAtLeastForwardIterator = std::is_convertible<
typename std::iterator_traits<Iterator>::iterator_category,
std::forward_iterator_tag>;
template <typename AllocatorType,
typename ValueType =
typename absl::allocator_traits<AllocatorType>::value_type>
using IsMemcpyOk =
absl::conjunction<std::is_same<AllocatorType, std::allocator<ValueType>>,
absl::is_trivially_copy_constructible<ValueType>,
absl::is_trivially_copy_assignable<ValueType>,
absl::is_trivially_destructible<ValueType>>;
template <typename AllocatorType, typename Pointer, typename SizeType>
void DestroyElements(AllocatorType* alloc_ptr, Pointer destroy_first,
SizeType destroy_size) {
using AllocatorTraits = absl::allocator_traits<AllocatorType>;
if (destroy_first != nullptr) {
for (auto i = destroy_size; i != 0;) {
--i;
AllocatorTraits::destroy(*alloc_ptr, destroy_first + i);
}
#if !defined(NDEBUG)
{
using ValueType = typename AllocatorTraits::value_type;
// Overwrite unused memory with `0xab` so we can catch uninitialized
// usage.
//
// Cast to `void*` to tell the compiler that we don't care that we might
// be scribbling on a vtable pointer.
void* memory_ptr = destroy_first;
auto memory_size = destroy_size * sizeof(ValueType);
std::memset(memory_ptr, 0xab, memory_size);
}
#endif // !defined(NDEBUG)
}
}
Export of internal Abseil changes -- c68f1886f5e8fd90eb0c2d2e68feaf00a7cdacda by CJ Johnson <johnsoncj@google.com>: Introduce absl::Cleanup to the OSS repo PiperOrigin-RevId: 354583156 -- 17030cf388e10f7eb959e3e566326d1072ce392e by Abseil Team <absl-team@google.com>: Internal change only PiperOrigin-RevId: 354574953 -- e979d7236d4f3252e79ddda6739b67a9a326bf6d by CJ Johnson <johnsoncj@google.com>: Internal change PiperOrigin-RevId: 354545297 -- 7ea02b3783f7f49ef97d86a8f6580a19cc57df14 by Abseil Team <absl-team@google.com>: Pre-allocate memory for vectors where the size is known. PiperOrigin-RevId: 354344576 -- 9246c7cb11f1d6444f79ebe25acc69a8a9b870e0 by Matt Kulukundis <kfm@google.com>: Add support for Elbrus 2000 (e2k) Import of https://github.com/abseil/abseil-cpp/pull/889 PiperOrigin-RevId: 354344013 -- 0fc93d359cc1fb307552e917b37b7b2e7eed822f by Abseil Team <absl-team@google.com>: Integrate CordRepRing logic into cord (but do not enable it) PiperOrigin-RevId: 354312238 -- eda05622f7da71466723acb33403f783529df24b by Abseil Team <absl-team@google.com>: Protect ignore diagnostic with "__has_warning". PiperOrigin-RevId: 354112334 -- 47716c5d8fb10efa4fdd801d28bac414c6f8ec32 by Abseil Team <absl-team@google.com>: Rearrange InlinedVector copy constructor and destructor to treat a few special cases inline and then tail-call a non-inlined routine for the rest. In particular, we optimize for empty vectors in both cases. Added a couple of benchmarks that copy either an InlVec<int64> or an InlVec<InlVec<int64>>. Speed difference: ``` BM_CopyTrivial/0 0.92ns +- 0% 0.47ns +- 0% -48.91% (p=0.000 n=11+12) BM_CopyTrivial/1 0.92ns +- 0% 1.15ns +- 0% +25.00% (p=0.000 n=10+9) BM_CopyTrivial/8 8.57ns +- 0% 10.72ns +- 1% +25.16% (p=0.000 n=10+12) BM_CopyNonTrivial/0 3.21ns +- 0% 0.70ns +- 0% -78.23% (p=0.000 n=12+10) BM_CopyNonTrivial/1 5.88ns +- 1% 5.51ns +- 0% -6.28% (p=0.000 n=10+8) BM_CopyNonTrivial/8 21.5ns +- 1% 15.2ns +- 2% -29.23% (p=0.000 n=12+12) ``` Note: the slowdowns are a few cycles which is expected given the procedure call added in that case. We decided this is a good tradeoff given the code size reductions and the more significant speedups for empty vectors. Size difference (as measured by nm): ``` BM_CopyTrivial from 1048 bytes to 326 bytes. BM_CopyNonTrivial from 749 bytes to 470 bytes. ``` Code size for a large binary drops by ~500KB (from 349415719 to 348906015 348906191). All of the benchmarks that showed a significant difference: Ones that improve with this CL: ``` BM_CopyNonTrivial/0 3.21ns +- 0% 0.70ns +- 0% -78.23% (p=0.000 n=12+10) BM_InlinedVectorFillString/0 0.93ns +- 0% 0.24ns +- 0% -74.19% (p=0.000 n=12+10) BM_InlinedVectorAssignments/1 10.5ns +- 0% 4.1ns +- 0% -60.64% (p=0.000 n=11+10) BM_InlinedVectorAssignments/2 10.7ns +- 0% 4.4ns +- 0% -59.08% (p=0.000 n=11+11) BM_CopyTrivial/0 0.92ns +- 0% 0.47ns +- 0% -48.91% (p=0.000 n=11+12) BM_CopyNonTrivial/8 21.5ns +- 1% 15.2ns +- 2% -29.23% (p=0.000 n=12+12) BM_StdVectorEmpty 0.47ns +- 1% 0.35ns +- 0% -24.73% (p=0.000 n=12+12) BM_StdVectorSize 0.46ns +- 2% 0.35ns +- 0% -24.32% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableOnly>/0 3.44ns +- 0% 2.76ns +- 1% -19.83% (p=0.000 n=11+11) BM_InlinedVectorFillRange/256 20.7ns +- 1% 17.8ns +- 0% -14.08% (p=0.000 n=12+9) BM_CopyNonTrivial/1 5.88ns +- 1% 5.51ns +- 0% -6.28% (p=0.000 n=10+8) BM_SwapElements<LargeCopyableMovable>/1 4.19ns +- 0% 3.95ns +- 1% -5.63% (p=0.000 n=11+12) BM_SwapElements<LargeCopyableMovableSwappable>/1 4.18ns +- 0% 3.99ns +- 0% -4.70% (p=0.000 n=9+11) BM_SwapElements<LargeCopyableMovable>/0 2.41ns +- 0% 2.31ns +- 0% -4.45% (p=0.000 n=12+12) BM_InlinedVectorFillRange/64 8.25ns +- 0% 8.04ns +- 0% -2.51% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableOnly>/1 82.4ns +- 0% 81.5ns +- 0% -1.06% (p=0.000 n=12+12) ``` Ones that get worse with this CL: ``` BM_CopyTrivial/1 0.92ns +- 0% 1.15ns +- 0% +25.00% (p=0.000 n=10+9) BM_CopyTrivial/8 8.57ns +- 0% 10.72ns +- 1% +25.16% (p=0.000 n=10+12) BM_SwapElements<LargeCopyableMovableSwappable>/512 1.48ns +- 1% 1.66ns +- 1% +11.88% (p=0.000 n=12+12) BM_InlinedVectorFillString/1 11.5ns +- 0% 12.8ns +- 1% +11.62% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableMovableSwappable>/64 1.48ns +- 2% 1.66ns +- 1% +11.66% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableMovableSwappable>/1k 1.48ns +- 1% 1.65ns +- 2% +11.32% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableMovable>/512 1.48ns +- 2% 1.58ns +- 4% +6.62% (p=0.000 n=11+12) BM_SwapElements<LargeCopyableMovable>/1k 1.49ns +- 2% 1.58ns +- 3% +6.05% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableMovable>/64 1.48ns +- 2% 1.57ns +- 4% +6.04% (p=0.000 n=11+12) BM_InlinedVectorFillRange/1 4.81ns +- 0% 5.05ns +- 0% +4.83% (p=0.000 n=11+11) BM_InlinedVectorFillString/8 79.4ns +- 1% 83.1ns +- 1% +4.64% (p=0.000 n=10+12) BM_StdVectorFillString/1 16.3ns +- 0% 16.6ns +- 0% +2.13% (p=0.000 n=11+8) ``` PiperOrigin-RevId: 353906786 -- 8e26518b3cec9c598e5e9573c46c3bd1b03a67ef by Abseil Team <absl-team@google.com>: Internal change PiperOrigin-RevId: 353737330 -- f206ae0983e58c9904ed8b8f05f9caf564a446be by Matt Kulukundis <kfm@google.com>: Import of CCTZ from GitHub. PiperOrigin-RevId: 353682256 GitOrigin-RevId: c68f1886f5e8fd90eb0c2d2e68feaf00a7cdacda Change-Id: I5790c1036c4f543c701d1039848fabf7ae881ad8
4 years ago
// If kUseMemcpy is true, memcpy(dst, src, n); else do nothing.
// Useful to avoid compiler warnings when memcpy() is used for T values
// that are not trivially copyable in non-reachable code.
template <bool kUseMemcpy>
inline void MemcpyIfAllowed(void* dst, const void* src, size_t n);
// memcpy when allowed.
template <>
inline void MemcpyIfAllowed<true>(void* dst, const void* src, size_t n) {
memcpy(dst, src, n);
}
// Do nothing for types that are not memcpy-able. This function is only
// called from non-reachable branches.
template <>
inline void MemcpyIfAllowed<false>(void*, const void*, size_t) {}
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
template <typename AllocatorType, typename Pointer, typename ValueAdapter,
typename SizeType>
void ConstructElements(AllocatorType* alloc_ptr, Pointer construct_first,
ValueAdapter* values_ptr, SizeType construct_size) {
for (SizeType i = 0; i < construct_size; ++i) {
ABSL_INTERNAL_TRY {
values_ptr->ConstructNext(alloc_ptr, construct_first + i);
}
ABSL_INTERNAL_CATCH_ANY {
inlined_vector_internal::DestroyElements(alloc_ptr, construct_first, i);
ABSL_INTERNAL_RETHROW;
}
}
}
template <typename Pointer, typename ValueAdapter, typename SizeType>
void AssignElements(Pointer assign_first, ValueAdapter* values_ptr,
SizeType assign_size) {
for (SizeType i = 0; i < assign_size; ++i) {
values_ptr->AssignNext(assign_first + i);
}
}
template <typename AllocatorType>
struct StorageView {
using AllocatorTraits = absl::allocator_traits<AllocatorType>;
using Pointer = typename AllocatorTraits::pointer;
using SizeType = typename AllocatorTraits::size_type;
Pointer data;
SizeType size;
SizeType capacity;
};
template <typename AllocatorType, typename Iterator>
class IteratorValueAdapter {
using AllocatorTraits = absl::allocator_traits<AllocatorType>;
using Pointer = typename AllocatorTraits::pointer;
public:
explicit IteratorValueAdapter(const Iterator& it) : it_(it) {}
void ConstructNext(AllocatorType* alloc_ptr, Pointer construct_at) {
AllocatorTraits::construct(*alloc_ptr, construct_at, *it_);
++it_;
}
void AssignNext(Pointer assign_at) {
*assign_at = *it_;
++it_;
}
private:
Iterator it_;
};
template <typename AllocatorType>
class CopyValueAdapter {
using AllocatorTraits = absl::allocator_traits<AllocatorType>;
using ValueType = typename AllocatorTraits::value_type;
using Pointer = typename AllocatorTraits::pointer;
using ConstPointer = typename AllocatorTraits::const_pointer;
public:
explicit CopyValueAdapter(const ValueType& v) : ptr_(std::addressof(v)) {}
void ConstructNext(AllocatorType* alloc_ptr, Pointer construct_at) {
AllocatorTraits::construct(*alloc_ptr, construct_at, *ptr_);
}
void AssignNext(Pointer assign_at) { *assign_at = *ptr_; }
private:
ConstPointer ptr_;
};
template <typename AllocatorType>
class DefaultValueAdapter {
using AllocatorTraits = absl::allocator_traits<AllocatorType>;
using ValueType = typename AllocatorTraits::value_type;
using Pointer = typename AllocatorTraits::pointer;
public:
explicit DefaultValueAdapter() {}
void ConstructNext(AllocatorType* alloc_ptr, Pointer construct_at) {
AllocatorTraits::construct(*alloc_ptr, construct_at);
}
void AssignNext(Pointer assign_at) { *assign_at = ValueType(); }
};
template <typename AllocatorType>
class AllocationTransaction {
using AllocatorTraits = absl::allocator_traits<AllocatorType>;
using Pointer = typename AllocatorTraits::pointer;
using SizeType = typename AllocatorTraits::size_type;
public:
explicit AllocationTransaction(AllocatorType* alloc_ptr)
: alloc_data_(*alloc_ptr, nullptr) {}
~AllocationTransaction() {
if (DidAllocate()) {
AllocatorTraits::deallocate(GetAllocator(), GetData(), GetCapacity());
}
}
AllocationTransaction(const AllocationTransaction&) = delete;
void operator=(const AllocationTransaction&) = delete;
AllocatorType& GetAllocator() { return alloc_data_.template get<0>(); }
Pointer& GetData() { return alloc_data_.template get<1>(); }
SizeType& GetCapacity() { return capacity_; }
bool DidAllocate() { return GetData() != nullptr; }
Pointer Allocate(SizeType capacity) {
GetData() = AllocatorTraits::allocate(GetAllocator(), capacity);
GetCapacity() = capacity;
return GetData();
}
void Reset() {
GetData() = nullptr;
GetCapacity() = 0;
}
private:
container_internal::CompressedTuple<AllocatorType, Pointer> alloc_data_;
SizeType capacity_ = 0;
};
template <typename AllocatorType>
class ConstructionTransaction {
using AllocatorTraits = absl::allocator_traits<AllocatorType>;
using Pointer = typename AllocatorTraits::pointer;
using SizeType = typename AllocatorTraits::size_type;
public:
explicit ConstructionTransaction(AllocatorType* alloc_ptr)
: alloc_data_(*alloc_ptr, nullptr) {}
~ConstructionTransaction() {
if (DidConstruct()) {
inlined_vector_internal::DestroyElements(std::addressof(GetAllocator()),
GetData(), GetSize());
}
}
ConstructionTransaction(const ConstructionTransaction&) = delete;
void operator=(const ConstructionTransaction&) = delete;
AllocatorType& GetAllocator() { return alloc_data_.template get<0>(); }
Pointer& GetData() { return alloc_data_.template get<1>(); }
SizeType& GetSize() { return size_; }
bool DidConstruct() { return GetData() != nullptr; }
template <typename ValueAdapter>
void Construct(Pointer data, ValueAdapter* values_ptr, SizeType size) {
inlined_vector_internal::ConstructElements(std::addressof(GetAllocator()),
data, values_ptr, size);
GetData() = data;
GetSize() = size;
}
void Commit() {
GetData() = nullptr;
GetSize() = 0;
}
private:
container_internal::CompressedTuple<AllocatorType, Pointer> alloc_data_;
SizeType size_ = 0;
};
template <typename T, size_t N, typename A>
class Storage {
public:
using AllocatorTraits = absl::allocator_traits<A>;
using allocator_type = typename AllocatorTraits::allocator_type;
using value_type = typename AllocatorTraits::value_type;
using pointer = typename AllocatorTraits::pointer;
using const_pointer = typename AllocatorTraits::const_pointer;
using size_type = typename AllocatorTraits::size_type;
using difference_type = typename AllocatorTraits::difference_type;
using reference = value_type&;
using const_reference = const value_type&;
using RValueReference = value_type&&;
using iterator = pointer;
using const_iterator = const_pointer;
using reverse_iterator = std::reverse_iterator<iterator>;
using const_reverse_iterator = std::reverse_iterator<const_iterator>;
using MoveIterator = std::move_iterator<iterator>;
using IsMemcpyOk = inlined_vector_internal::IsMemcpyOk<allocator_type>;
using StorageView = inlined_vector_internal::StorageView<allocator_type>;
template <typename Iterator>
using IteratorValueAdapter =
inlined_vector_internal::IteratorValueAdapter<allocator_type, Iterator>;
using CopyValueAdapter =
inlined_vector_internal::CopyValueAdapter<allocator_type>;
using DefaultValueAdapter =
inlined_vector_internal::DefaultValueAdapter<allocator_type>;
using AllocationTransaction =
inlined_vector_internal::AllocationTransaction<allocator_type>;
using ConstructionTransaction =
inlined_vector_internal::ConstructionTransaction<allocator_type>;
static size_type NextCapacity(size_type current_capacity) {
return current_capacity * 2;
}
static size_type ComputeCapacity(size_type current_capacity,
size_type requested_capacity) {
return (std::max)(NextCapacity(current_capacity), requested_capacity);
}
// ---------------------------------------------------------------------------
// Storage Constructors and Destructor
// ---------------------------------------------------------------------------
Export of internal Abseil changes -- ff793052bd01e1e4fcf639f94d7c30c4855a9372 by Evan Brown <ezb@google.com>: Roll forward of btree_iterator refactoring. PiperOrigin-RevId: 346116047 -- 17984679f16e3e2139b0f14fa76f4a6ca16a3ef9 by Chris Kennelly <ckennelly@google.com>: Extend absl::StrContains to accept single character needles. Single characters are more efficient to search for. Extending this API allows the abseil-string-find-str-contains Clang Tidy to include this pattern. The C++ committee has adopted http://wg21.link/P1679 for inclusion in C++23. PiperOrigin-RevId: 346095060 -- ef20b31c501b1dcaa25e244fd8f8aa43dec09bd6 by Jorg Brown <jorg@google.com>: Internal change for cord ring PiperOrigin-RevId: 346087545 -- b70f2c1cb77fc9e733a126e790967d45c5fd1dc7 by Derek Mauro <dmauro@google.com>: Release layout_benchmark PiperOrigin-RevId: 345968909 -- 3a0eda337ee43622f92cfe14c2aa06f72dc71ee5 by Derek Mauro <dmauro@google.com>: Release raw_hash_set_probe_benchmark PiperOrigin-RevId: 345965969 -- abffdb4bb241a2264cb4e73a6262b660bb10447d by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 345733599 -- 7c9e24a71188df945be17fe98f700bdb51f81b16 by Derek Mauro <dmauro@google.com>: Release hash_benchmark PiperOrigin-RevId: 345721635 -- d68f33f17f9a8cd3f6da8eee3870bdb46402cdc8 by Derek Mauro <dmauro@google.com>: Release raw_hash_set_benchmark PiperOrigin-RevId: 345708384 -- 6e6c547d4d1327b226c0ffe8ff34d0aa103ce24b by Abseil Team <absl-team@google.com>: Updates the implementation of InlinedVector to accurately express the value-initialization semantics of the default constructor PiperOrigin-RevId: 345548260 -- 1532424deda97d468444c217cc0fa4614099c7c1 by Evan Brown <ezb@google.com>: Rollback btree_iterator refactoring. PiperOrigin-RevId: 345543900 GitOrigin-RevId: ff793052bd01e1e4fcf639f94d7c30c4855a9372 Change-Id: I719831981fd056de41939f9addfee3d85e3b49b2
4 years ago
Storage() : metadata_(allocator_type(), /* size and is_allocated */ 0) {}
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
Export of internal Abseil changes -- ff793052bd01e1e4fcf639f94d7c30c4855a9372 by Evan Brown <ezb@google.com>: Roll forward of btree_iterator refactoring. PiperOrigin-RevId: 346116047 -- 17984679f16e3e2139b0f14fa76f4a6ca16a3ef9 by Chris Kennelly <ckennelly@google.com>: Extend absl::StrContains to accept single character needles. Single characters are more efficient to search for. Extending this API allows the abseil-string-find-str-contains Clang Tidy to include this pattern. The C++ committee has adopted http://wg21.link/P1679 for inclusion in C++23. PiperOrigin-RevId: 346095060 -- ef20b31c501b1dcaa25e244fd8f8aa43dec09bd6 by Jorg Brown <jorg@google.com>: Internal change for cord ring PiperOrigin-RevId: 346087545 -- b70f2c1cb77fc9e733a126e790967d45c5fd1dc7 by Derek Mauro <dmauro@google.com>: Release layout_benchmark PiperOrigin-RevId: 345968909 -- 3a0eda337ee43622f92cfe14c2aa06f72dc71ee5 by Derek Mauro <dmauro@google.com>: Release raw_hash_set_probe_benchmark PiperOrigin-RevId: 345965969 -- abffdb4bb241a2264cb4e73a6262b660bb10447d by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 345733599 -- 7c9e24a71188df945be17fe98f700bdb51f81b16 by Derek Mauro <dmauro@google.com>: Release hash_benchmark PiperOrigin-RevId: 345721635 -- d68f33f17f9a8cd3f6da8eee3870bdb46402cdc8 by Derek Mauro <dmauro@google.com>: Release raw_hash_set_benchmark PiperOrigin-RevId: 345708384 -- 6e6c547d4d1327b226c0ffe8ff34d0aa103ce24b by Abseil Team <absl-team@google.com>: Updates the implementation of InlinedVector to accurately express the value-initialization semantics of the default constructor PiperOrigin-RevId: 345548260 -- 1532424deda97d468444c217cc0fa4614099c7c1 by Evan Brown <ezb@google.com>: Rollback btree_iterator refactoring. PiperOrigin-RevId: 345543900 GitOrigin-RevId: ff793052bd01e1e4fcf639f94d7c30c4855a9372 Change-Id: I719831981fd056de41939f9addfee3d85e3b49b2
4 years ago
explicit Storage(const allocator_type& alloc)
: metadata_(alloc, /* size and is_allocated */ 0) {}
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
~Storage() {
Export of internal Abseil changes -- c68f1886f5e8fd90eb0c2d2e68feaf00a7cdacda by CJ Johnson <johnsoncj@google.com>: Introduce absl::Cleanup to the OSS repo PiperOrigin-RevId: 354583156 -- 17030cf388e10f7eb959e3e566326d1072ce392e by Abseil Team <absl-team@google.com>: Internal change only PiperOrigin-RevId: 354574953 -- e979d7236d4f3252e79ddda6739b67a9a326bf6d by CJ Johnson <johnsoncj@google.com>: Internal change PiperOrigin-RevId: 354545297 -- 7ea02b3783f7f49ef97d86a8f6580a19cc57df14 by Abseil Team <absl-team@google.com>: Pre-allocate memory for vectors where the size is known. PiperOrigin-RevId: 354344576 -- 9246c7cb11f1d6444f79ebe25acc69a8a9b870e0 by Matt Kulukundis <kfm@google.com>: Add support for Elbrus 2000 (e2k) Import of https://github.com/abseil/abseil-cpp/pull/889 PiperOrigin-RevId: 354344013 -- 0fc93d359cc1fb307552e917b37b7b2e7eed822f by Abseil Team <absl-team@google.com>: Integrate CordRepRing logic into cord (but do not enable it) PiperOrigin-RevId: 354312238 -- eda05622f7da71466723acb33403f783529df24b by Abseil Team <absl-team@google.com>: Protect ignore diagnostic with "__has_warning". PiperOrigin-RevId: 354112334 -- 47716c5d8fb10efa4fdd801d28bac414c6f8ec32 by Abseil Team <absl-team@google.com>: Rearrange InlinedVector copy constructor and destructor to treat a few special cases inline and then tail-call a non-inlined routine for the rest. In particular, we optimize for empty vectors in both cases. Added a couple of benchmarks that copy either an InlVec<int64> or an InlVec<InlVec<int64>>. Speed difference: ``` BM_CopyTrivial/0 0.92ns +- 0% 0.47ns +- 0% -48.91% (p=0.000 n=11+12) BM_CopyTrivial/1 0.92ns +- 0% 1.15ns +- 0% +25.00% (p=0.000 n=10+9) BM_CopyTrivial/8 8.57ns +- 0% 10.72ns +- 1% +25.16% (p=0.000 n=10+12) BM_CopyNonTrivial/0 3.21ns +- 0% 0.70ns +- 0% -78.23% (p=0.000 n=12+10) BM_CopyNonTrivial/1 5.88ns +- 1% 5.51ns +- 0% -6.28% (p=0.000 n=10+8) BM_CopyNonTrivial/8 21.5ns +- 1% 15.2ns +- 2% -29.23% (p=0.000 n=12+12) ``` Note: the slowdowns are a few cycles which is expected given the procedure call added in that case. We decided this is a good tradeoff given the code size reductions and the more significant speedups for empty vectors. Size difference (as measured by nm): ``` BM_CopyTrivial from 1048 bytes to 326 bytes. BM_CopyNonTrivial from 749 bytes to 470 bytes. ``` Code size for a large binary drops by ~500KB (from 349415719 to 348906015 348906191). All of the benchmarks that showed a significant difference: Ones that improve with this CL: ``` BM_CopyNonTrivial/0 3.21ns +- 0% 0.70ns +- 0% -78.23% (p=0.000 n=12+10) BM_InlinedVectorFillString/0 0.93ns +- 0% 0.24ns +- 0% -74.19% (p=0.000 n=12+10) BM_InlinedVectorAssignments/1 10.5ns +- 0% 4.1ns +- 0% -60.64% (p=0.000 n=11+10) BM_InlinedVectorAssignments/2 10.7ns +- 0% 4.4ns +- 0% -59.08% (p=0.000 n=11+11) BM_CopyTrivial/0 0.92ns +- 0% 0.47ns +- 0% -48.91% (p=0.000 n=11+12) BM_CopyNonTrivial/8 21.5ns +- 1% 15.2ns +- 2% -29.23% (p=0.000 n=12+12) BM_StdVectorEmpty 0.47ns +- 1% 0.35ns +- 0% -24.73% (p=0.000 n=12+12) BM_StdVectorSize 0.46ns +- 2% 0.35ns +- 0% -24.32% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableOnly>/0 3.44ns +- 0% 2.76ns +- 1% -19.83% (p=0.000 n=11+11) BM_InlinedVectorFillRange/256 20.7ns +- 1% 17.8ns +- 0% -14.08% (p=0.000 n=12+9) BM_CopyNonTrivial/1 5.88ns +- 1% 5.51ns +- 0% -6.28% (p=0.000 n=10+8) BM_SwapElements<LargeCopyableMovable>/1 4.19ns +- 0% 3.95ns +- 1% -5.63% (p=0.000 n=11+12) BM_SwapElements<LargeCopyableMovableSwappable>/1 4.18ns +- 0% 3.99ns +- 0% -4.70% (p=0.000 n=9+11) BM_SwapElements<LargeCopyableMovable>/0 2.41ns +- 0% 2.31ns +- 0% -4.45% (p=0.000 n=12+12) BM_InlinedVectorFillRange/64 8.25ns +- 0% 8.04ns +- 0% -2.51% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableOnly>/1 82.4ns +- 0% 81.5ns +- 0% -1.06% (p=0.000 n=12+12) ``` Ones that get worse with this CL: ``` BM_CopyTrivial/1 0.92ns +- 0% 1.15ns +- 0% +25.00% (p=0.000 n=10+9) BM_CopyTrivial/8 8.57ns +- 0% 10.72ns +- 1% +25.16% (p=0.000 n=10+12) BM_SwapElements<LargeCopyableMovableSwappable>/512 1.48ns +- 1% 1.66ns +- 1% +11.88% (p=0.000 n=12+12) BM_InlinedVectorFillString/1 11.5ns +- 0% 12.8ns +- 1% +11.62% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableMovableSwappable>/64 1.48ns +- 2% 1.66ns +- 1% +11.66% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableMovableSwappable>/1k 1.48ns +- 1% 1.65ns +- 2% +11.32% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableMovable>/512 1.48ns +- 2% 1.58ns +- 4% +6.62% (p=0.000 n=11+12) BM_SwapElements<LargeCopyableMovable>/1k 1.49ns +- 2% 1.58ns +- 3% +6.05% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableMovable>/64 1.48ns +- 2% 1.57ns +- 4% +6.04% (p=0.000 n=11+12) BM_InlinedVectorFillRange/1 4.81ns +- 0% 5.05ns +- 0% +4.83% (p=0.000 n=11+11) BM_InlinedVectorFillString/8 79.4ns +- 1% 83.1ns +- 1% +4.64% (p=0.000 n=10+12) BM_StdVectorFillString/1 16.3ns +- 0% 16.6ns +- 0% +2.13% (p=0.000 n=11+8) ``` PiperOrigin-RevId: 353906786 -- 8e26518b3cec9c598e5e9573c46c3bd1b03a67ef by Abseil Team <absl-team@google.com>: Internal change PiperOrigin-RevId: 353737330 -- f206ae0983e58c9904ed8b8f05f9caf564a446be by Matt Kulukundis <kfm@google.com>: Import of CCTZ from GitHub. PiperOrigin-RevId: 353682256 GitOrigin-RevId: c68f1886f5e8fd90eb0c2d2e68feaf00a7cdacda Change-Id: I5790c1036c4f543c701d1039848fabf7ae881ad8
4 years ago
if (GetSizeAndIsAllocated() == 0) {
// Empty and not allocated; nothing to do.
} else if (IsMemcpyOk::value) {
// No destructors need to be run; just deallocate if necessary.
DeallocateIfAllocated();
} else {
DestroyContents();
}
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
}
// ---------------------------------------------------------------------------
// Storage Member Accessors
// ---------------------------------------------------------------------------
size_type& GetSizeAndIsAllocated() { return metadata_.template get<1>(); }
const size_type& GetSizeAndIsAllocated() const {
return metadata_.template get<1>();
}
size_type GetSize() const { return GetSizeAndIsAllocated() >> 1; }
bool GetIsAllocated() const { return GetSizeAndIsAllocated() & 1; }
pointer GetAllocatedData() { return data_.allocated.allocated_data; }
const_pointer GetAllocatedData() const {
return data_.allocated.allocated_data;
}
pointer GetInlinedData() {
return reinterpret_cast<pointer>(
std::addressof(data_.inlined.inlined_data[0]));
}
const_pointer GetInlinedData() const {
return reinterpret_cast<const_pointer>(
std::addressof(data_.inlined.inlined_data[0]));
}
size_type GetAllocatedCapacity() const {
return data_.allocated.allocated_capacity;
}
size_type GetInlinedCapacity() const { return static_cast<size_type>(N); }
StorageView MakeStorageView() {
return GetIsAllocated()
? StorageView{GetAllocatedData(), GetSize(),
GetAllocatedCapacity()}
: StorageView{GetInlinedData(), GetSize(), GetInlinedCapacity()};
}
allocator_type* GetAllocPtr() {
return std::addressof(metadata_.template get<0>());
}
const allocator_type* GetAllocPtr() const {
return std::addressof(metadata_.template get<0>());
}
// ---------------------------------------------------------------------------
// Storage Member Mutators
// ---------------------------------------------------------------------------
Export of internal Abseil changes -- c68f1886f5e8fd90eb0c2d2e68feaf00a7cdacda by CJ Johnson <johnsoncj@google.com>: Introduce absl::Cleanup to the OSS repo PiperOrigin-RevId: 354583156 -- 17030cf388e10f7eb959e3e566326d1072ce392e by Abseil Team <absl-team@google.com>: Internal change only PiperOrigin-RevId: 354574953 -- e979d7236d4f3252e79ddda6739b67a9a326bf6d by CJ Johnson <johnsoncj@google.com>: Internal change PiperOrigin-RevId: 354545297 -- 7ea02b3783f7f49ef97d86a8f6580a19cc57df14 by Abseil Team <absl-team@google.com>: Pre-allocate memory for vectors where the size is known. PiperOrigin-RevId: 354344576 -- 9246c7cb11f1d6444f79ebe25acc69a8a9b870e0 by Matt Kulukundis <kfm@google.com>: Add support for Elbrus 2000 (e2k) Import of https://github.com/abseil/abseil-cpp/pull/889 PiperOrigin-RevId: 354344013 -- 0fc93d359cc1fb307552e917b37b7b2e7eed822f by Abseil Team <absl-team@google.com>: Integrate CordRepRing logic into cord (but do not enable it) PiperOrigin-RevId: 354312238 -- eda05622f7da71466723acb33403f783529df24b by Abseil Team <absl-team@google.com>: Protect ignore diagnostic with "__has_warning". PiperOrigin-RevId: 354112334 -- 47716c5d8fb10efa4fdd801d28bac414c6f8ec32 by Abseil Team <absl-team@google.com>: Rearrange InlinedVector copy constructor and destructor to treat a few special cases inline and then tail-call a non-inlined routine for the rest. In particular, we optimize for empty vectors in both cases. Added a couple of benchmarks that copy either an InlVec<int64> or an InlVec<InlVec<int64>>. Speed difference: ``` BM_CopyTrivial/0 0.92ns +- 0% 0.47ns +- 0% -48.91% (p=0.000 n=11+12) BM_CopyTrivial/1 0.92ns +- 0% 1.15ns +- 0% +25.00% (p=0.000 n=10+9) BM_CopyTrivial/8 8.57ns +- 0% 10.72ns +- 1% +25.16% (p=0.000 n=10+12) BM_CopyNonTrivial/0 3.21ns +- 0% 0.70ns +- 0% -78.23% (p=0.000 n=12+10) BM_CopyNonTrivial/1 5.88ns +- 1% 5.51ns +- 0% -6.28% (p=0.000 n=10+8) BM_CopyNonTrivial/8 21.5ns +- 1% 15.2ns +- 2% -29.23% (p=0.000 n=12+12) ``` Note: the slowdowns are a few cycles which is expected given the procedure call added in that case. We decided this is a good tradeoff given the code size reductions and the more significant speedups for empty vectors. Size difference (as measured by nm): ``` BM_CopyTrivial from 1048 bytes to 326 bytes. BM_CopyNonTrivial from 749 bytes to 470 bytes. ``` Code size for a large binary drops by ~500KB (from 349415719 to 348906015 348906191). All of the benchmarks that showed a significant difference: Ones that improve with this CL: ``` BM_CopyNonTrivial/0 3.21ns +- 0% 0.70ns +- 0% -78.23% (p=0.000 n=12+10) BM_InlinedVectorFillString/0 0.93ns +- 0% 0.24ns +- 0% -74.19% (p=0.000 n=12+10) BM_InlinedVectorAssignments/1 10.5ns +- 0% 4.1ns +- 0% -60.64% (p=0.000 n=11+10) BM_InlinedVectorAssignments/2 10.7ns +- 0% 4.4ns +- 0% -59.08% (p=0.000 n=11+11) BM_CopyTrivial/0 0.92ns +- 0% 0.47ns +- 0% -48.91% (p=0.000 n=11+12) BM_CopyNonTrivial/8 21.5ns +- 1% 15.2ns +- 2% -29.23% (p=0.000 n=12+12) BM_StdVectorEmpty 0.47ns +- 1% 0.35ns +- 0% -24.73% (p=0.000 n=12+12) BM_StdVectorSize 0.46ns +- 2% 0.35ns +- 0% -24.32% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableOnly>/0 3.44ns +- 0% 2.76ns +- 1% -19.83% (p=0.000 n=11+11) BM_InlinedVectorFillRange/256 20.7ns +- 1% 17.8ns +- 0% -14.08% (p=0.000 n=12+9) BM_CopyNonTrivial/1 5.88ns +- 1% 5.51ns +- 0% -6.28% (p=0.000 n=10+8) BM_SwapElements<LargeCopyableMovable>/1 4.19ns +- 0% 3.95ns +- 1% -5.63% (p=0.000 n=11+12) BM_SwapElements<LargeCopyableMovableSwappable>/1 4.18ns +- 0% 3.99ns +- 0% -4.70% (p=0.000 n=9+11) BM_SwapElements<LargeCopyableMovable>/0 2.41ns +- 0% 2.31ns +- 0% -4.45% (p=0.000 n=12+12) BM_InlinedVectorFillRange/64 8.25ns +- 0% 8.04ns +- 0% -2.51% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableOnly>/1 82.4ns +- 0% 81.5ns +- 0% -1.06% (p=0.000 n=12+12) ``` Ones that get worse with this CL: ``` BM_CopyTrivial/1 0.92ns +- 0% 1.15ns +- 0% +25.00% (p=0.000 n=10+9) BM_CopyTrivial/8 8.57ns +- 0% 10.72ns +- 1% +25.16% (p=0.000 n=10+12) BM_SwapElements<LargeCopyableMovableSwappable>/512 1.48ns +- 1% 1.66ns +- 1% +11.88% (p=0.000 n=12+12) BM_InlinedVectorFillString/1 11.5ns +- 0% 12.8ns +- 1% +11.62% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableMovableSwappable>/64 1.48ns +- 2% 1.66ns +- 1% +11.66% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableMovableSwappable>/1k 1.48ns +- 1% 1.65ns +- 2% +11.32% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableMovable>/512 1.48ns +- 2% 1.58ns +- 4% +6.62% (p=0.000 n=11+12) BM_SwapElements<LargeCopyableMovable>/1k 1.49ns +- 2% 1.58ns +- 3% +6.05% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableMovable>/64 1.48ns +- 2% 1.57ns +- 4% +6.04% (p=0.000 n=11+12) BM_InlinedVectorFillRange/1 4.81ns +- 0% 5.05ns +- 0% +4.83% (p=0.000 n=11+11) BM_InlinedVectorFillString/8 79.4ns +- 1% 83.1ns +- 1% +4.64% (p=0.000 n=10+12) BM_StdVectorFillString/1 16.3ns +- 0% 16.6ns +- 0% +2.13% (p=0.000 n=11+8) ``` PiperOrigin-RevId: 353906786 -- 8e26518b3cec9c598e5e9573c46c3bd1b03a67ef by Abseil Team <absl-team@google.com>: Internal change PiperOrigin-RevId: 353737330 -- f206ae0983e58c9904ed8b8f05f9caf564a446be by Matt Kulukundis <kfm@google.com>: Import of CCTZ from GitHub. PiperOrigin-RevId: 353682256 GitOrigin-RevId: c68f1886f5e8fd90eb0c2d2e68feaf00a7cdacda Change-Id: I5790c1036c4f543c701d1039848fabf7ae881ad8
4 years ago
ABSL_ATTRIBUTE_NOINLINE void InitFrom(const Storage& other);
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
template <typename ValueAdapter>
void Initialize(ValueAdapter values, size_type new_size);
template <typename ValueAdapter>
void Assign(ValueAdapter values, size_type new_size);
template <typename ValueAdapter>
void Resize(ValueAdapter values, size_type new_size);
template <typename ValueAdapter>
iterator Insert(const_iterator pos, ValueAdapter values,
size_type insert_count);
template <typename... Args>
reference EmplaceBack(Args&&... args);
iterator Erase(const_iterator from, const_iterator to);
void Reserve(size_type requested_capacity);
void ShrinkToFit();
void Swap(Storage* other_storage_ptr);
void SetIsAllocated() {
GetSizeAndIsAllocated() |= static_cast<size_type>(1);
}
void UnsetIsAllocated() {
GetSizeAndIsAllocated() &= ((std::numeric_limits<size_type>::max)() - 1);
}
void SetSize(size_type size) {
GetSizeAndIsAllocated() =
(size << 1) | static_cast<size_type>(GetIsAllocated());
}
void SetAllocatedSize(size_type size) {
GetSizeAndIsAllocated() = (size << 1) | static_cast<size_type>(1);
}
void SetInlinedSize(size_type size) {
GetSizeAndIsAllocated() = size << static_cast<size_type>(1);
}
void AddSize(size_type count) {
GetSizeAndIsAllocated() += count << static_cast<size_type>(1);
}
void SubtractSize(size_type count) {
assert(count <= GetSize());
GetSizeAndIsAllocated() -= count << static_cast<size_type>(1);
}
void SetAllocatedData(pointer data, size_type capacity) {
data_.allocated.allocated_data = data;
data_.allocated.allocated_capacity = capacity;
}
void AcquireAllocatedData(AllocationTransaction* allocation_tx_ptr) {
SetAllocatedData(allocation_tx_ptr->GetData(),
allocation_tx_ptr->GetCapacity());
allocation_tx_ptr->Reset();
}
void MemcpyFrom(const Storage& other_storage) {
assert(IsMemcpyOk::value || other_storage.GetIsAllocated());
GetSizeAndIsAllocated() = other_storage.GetSizeAndIsAllocated();
data_ = other_storage.data_;
}
void DeallocateIfAllocated() {
if (GetIsAllocated()) {
AllocatorTraits::deallocate(*GetAllocPtr(), GetAllocatedData(),
GetAllocatedCapacity());
}
}
private:
Export of internal Abseil changes -- c68f1886f5e8fd90eb0c2d2e68feaf00a7cdacda by CJ Johnson <johnsoncj@google.com>: Introduce absl::Cleanup to the OSS repo PiperOrigin-RevId: 354583156 -- 17030cf388e10f7eb959e3e566326d1072ce392e by Abseil Team <absl-team@google.com>: Internal change only PiperOrigin-RevId: 354574953 -- e979d7236d4f3252e79ddda6739b67a9a326bf6d by CJ Johnson <johnsoncj@google.com>: Internal change PiperOrigin-RevId: 354545297 -- 7ea02b3783f7f49ef97d86a8f6580a19cc57df14 by Abseil Team <absl-team@google.com>: Pre-allocate memory for vectors where the size is known. PiperOrigin-RevId: 354344576 -- 9246c7cb11f1d6444f79ebe25acc69a8a9b870e0 by Matt Kulukundis <kfm@google.com>: Add support for Elbrus 2000 (e2k) Import of https://github.com/abseil/abseil-cpp/pull/889 PiperOrigin-RevId: 354344013 -- 0fc93d359cc1fb307552e917b37b7b2e7eed822f by Abseil Team <absl-team@google.com>: Integrate CordRepRing logic into cord (but do not enable it) PiperOrigin-RevId: 354312238 -- eda05622f7da71466723acb33403f783529df24b by Abseil Team <absl-team@google.com>: Protect ignore diagnostic with "__has_warning". PiperOrigin-RevId: 354112334 -- 47716c5d8fb10efa4fdd801d28bac414c6f8ec32 by Abseil Team <absl-team@google.com>: Rearrange InlinedVector copy constructor and destructor to treat a few special cases inline and then tail-call a non-inlined routine for the rest. In particular, we optimize for empty vectors in both cases. Added a couple of benchmarks that copy either an InlVec<int64> or an InlVec<InlVec<int64>>. Speed difference: ``` BM_CopyTrivial/0 0.92ns +- 0% 0.47ns +- 0% -48.91% (p=0.000 n=11+12) BM_CopyTrivial/1 0.92ns +- 0% 1.15ns +- 0% +25.00% (p=0.000 n=10+9) BM_CopyTrivial/8 8.57ns +- 0% 10.72ns +- 1% +25.16% (p=0.000 n=10+12) BM_CopyNonTrivial/0 3.21ns +- 0% 0.70ns +- 0% -78.23% (p=0.000 n=12+10) BM_CopyNonTrivial/1 5.88ns +- 1% 5.51ns +- 0% -6.28% (p=0.000 n=10+8) BM_CopyNonTrivial/8 21.5ns +- 1% 15.2ns +- 2% -29.23% (p=0.000 n=12+12) ``` Note: the slowdowns are a few cycles which is expected given the procedure call added in that case. We decided this is a good tradeoff given the code size reductions and the more significant speedups for empty vectors. Size difference (as measured by nm): ``` BM_CopyTrivial from 1048 bytes to 326 bytes. BM_CopyNonTrivial from 749 bytes to 470 bytes. ``` Code size for a large binary drops by ~500KB (from 349415719 to 348906015 348906191). All of the benchmarks that showed a significant difference: Ones that improve with this CL: ``` BM_CopyNonTrivial/0 3.21ns +- 0% 0.70ns +- 0% -78.23% (p=0.000 n=12+10) BM_InlinedVectorFillString/0 0.93ns +- 0% 0.24ns +- 0% -74.19% (p=0.000 n=12+10) BM_InlinedVectorAssignments/1 10.5ns +- 0% 4.1ns +- 0% -60.64% (p=0.000 n=11+10) BM_InlinedVectorAssignments/2 10.7ns +- 0% 4.4ns +- 0% -59.08% (p=0.000 n=11+11) BM_CopyTrivial/0 0.92ns +- 0% 0.47ns +- 0% -48.91% (p=0.000 n=11+12) BM_CopyNonTrivial/8 21.5ns +- 1% 15.2ns +- 2% -29.23% (p=0.000 n=12+12) BM_StdVectorEmpty 0.47ns +- 1% 0.35ns +- 0% -24.73% (p=0.000 n=12+12) BM_StdVectorSize 0.46ns +- 2% 0.35ns +- 0% -24.32% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableOnly>/0 3.44ns +- 0% 2.76ns +- 1% -19.83% (p=0.000 n=11+11) BM_InlinedVectorFillRange/256 20.7ns +- 1% 17.8ns +- 0% -14.08% (p=0.000 n=12+9) BM_CopyNonTrivial/1 5.88ns +- 1% 5.51ns +- 0% -6.28% (p=0.000 n=10+8) BM_SwapElements<LargeCopyableMovable>/1 4.19ns +- 0% 3.95ns +- 1% -5.63% (p=0.000 n=11+12) BM_SwapElements<LargeCopyableMovableSwappable>/1 4.18ns +- 0% 3.99ns +- 0% -4.70% (p=0.000 n=9+11) BM_SwapElements<LargeCopyableMovable>/0 2.41ns +- 0% 2.31ns +- 0% -4.45% (p=0.000 n=12+12) BM_InlinedVectorFillRange/64 8.25ns +- 0% 8.04ns +- 0% -2.51% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableOnly>/1 82.4ns +- 0% 81.5ns +- 0% -1.06% (p=0.000 n=12+12) ``` Ones that get worse with this CL: ``` BM_CopyTrivial/1 0.92ns +- 0% 1.15ns +- 0% +25.00% (p=0.000 n=10+9) BM_CopyTrivial/8 8.57ns +- 0% 10.72ns +- 1% +25.16% (p=0.000 n=10+12) BM_SwapElements<LargeCopyableMovableSwappable>/512 1.48ns +- 1% 1.66ns +- 1% +11.88% (p=0.000 n=12+12) BM_InlinedVectorFillString/1 11.5ns +- 0% 12.8ns +- 1% +11.62% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableMovableSwappable>/64 1.48ns +- 2% 1.66ns +- 1% +11.66% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableMovableSwappable>/1k 1.48ns +- 1% 1.65ns +- 2% +11.32% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableMovable>/512 1.48ns +- 2% 1.58ns +- 4% +6.62% (p=0.000 n=11+12) BM_SwapElements<LargeCopyableMovable>/1k 1.49ns +- 2% 1.58ns +- 3% +6.05% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableMovable>/64 1.48ns +- 2% 1.57ns +- 4% +6.04% (p=0.000 n=11+12) BM_InlinedVectorFillRange/1 4.81ns +- 0% 5.05ns +- 0% +4.83% (p=0.000 n=11+11) BM_InlinedVectorFillString/8 79.4ns +- 1% 83.1ns +- 1% +4.64% (p=0.000 n=10+12) BM_StdVectorFillString/1 16.3ns +- 0% 16.6ns +- 0% +2.13% (p=0.000 n=11+8) ``` PiperOrigin-RevId: 353906786 -- 8e26518b3cec9c598e5e9573c46c3bd1b03a67ef by Abseil Team <absl-team@google.com>: Internal change PiperOrigin-RevId: 353737330 -- f206ae0983e58c9904ed8b8f05f9caf564a446be by Matt Kulukundis <kfm@google.com>: Import of CCTZ from GitHub. PiperOrigin-RevId: 353682256 GitOrigin-RevId: c68f1886f5e8fd90eb0c2d2e68feaf00a7cdacda Change-Id: I5790c1036c4f543c701d1039848fabf7ae881ad8
4 years ago
ABSL_ATTRIBUTE_NOINLINE void DestroyContents();
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
using Metadata =
container_internal::CompressedTuple<allocator_type, size_type>;
struct Allocated {
pointer allocated_data;
size_type allocated_capacity;
};
struct Inlined {
alignas(value_type) char inlined_data[sizeof(value_type[N])];
};
union Data {
Allocated allocated;
Inlined inlined;
};
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
template <typename... Args>
ABSL_ATTRIBUTE_NOINLINE reference EmplaceBackSlow(Args&&... args);
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
Metadata metadata_;
Data data_;
};
Export of internal Abseil changes -- c68f1886f5e8fd90eb0c2d2e68feaf00a7cdacda by CJ Johnson <johnsoncj@google.com>: Introduce absl::Cleanup to the OSS repo PiperOrigin-RevId: 354583156 -- 17030cf388e10f7eb959e3e566326d1072ce392e by Abseil Team <absl-team@google.com>: Internal change only PiperOrigin-RevId: 354574953 -- e979d7236d4f3252e79ddda6739b67a9a326bf6d by CJ Johnson <johnsoncj@google.com>: Internal change PiperOrigin-RevId: 354545297 -- 7ea02b3783f7f49ef97d86a8f6580a19cc57df14 by Abseil Team <absl-team@google.com>: Pre-allocate memory for vectors where the size is known. PiperOrigin-RevId: 354344576 -- 9246c7cb11f1d6444f79ebe25acc69a8a9b870e0 by Matt Kulukundis <kfm@google.com>: Add support for Elbrus 2000 (e2k) Import of https://github.com/abseil/abseil-cpp/pull/889 PiperOrigin-RevId: 354344013 -- 0fc93d359cc1fb307552e917b37b7b2e7eed822f by Abseil Team <absl-team@google.com>: Integrate CordRepRing logic into cord (but do not enable it) PiperOrigin-RevId: 354312238 -- eda05622f7da71466723acb33403f783529df24b by Abseil Team <absl-team@google.com>: Protect ignore diagnostic with "__has_warning". PiperOrigin-RevId: 354112334 -- 47716c5d8fb10efa4fdd801d28bac414c6f8ec32 by Abseil Team <absl-team@google.com>: Rearrange InlinedVector copy constructor and destructor to treat a few special cases inline and then tail-call a non-inlined routine for the rest. In particular, we optimize for empty vectors in both cases. Added a couple of benchmarks that copy either an InlVec<int64> or an InlVec<InlVec<int64>>. Speed difference: ``` BM_CopyTrivial/0 0.92ns +- 0% 0.47ns +- 0% -48.91% (p=0.000 n=11+12) BM_CopyTrivial/1 0.92ns +- 0% 1.15ns +- 0% +25.00% (p=0.000 n=10+9) BM_CopyTrivial/8 8.57ns +- 0% 10.72ns +- 1% +25.16% (p=0.000 n=10+12) BM_CopyNonTrivial/0 3.21ns +- 0% 0.70ns +- 0% -78.23% (p=0.000 n=12+10) BM_CopyNonTrivial/1 5.88ns +- 1% 5.51ns +- 0% -6.28% (p=0.000 n=10+8) BM_CopyNonTrivial/8 21.5ns +- 1% 15.2ns +- 2% -29.23% (p=0.000 n=12+12) ``` Note: the slowdowns are a few cycles which is expected given the procedure call added in that case. We decided this is a good tradeoff given the code size reductions and the more significant speedups for empty vectors. Size difference (as measured by nm): ``` BM_CopyTrivial from 1048 bytes to 326 bytes. BM_CopyNonTrivial from 749 bytes to 470 bytes. ``` Code size for a large binary drops by ~500KB (from 349415719 to 348906015 348906191). All of the benchmarks that showed a significant difference: Ones that improve with this CL: ``` BM_CopyNonTrivial/0 3.21ns +- 0% 0.70ns +- 0% -78.23% (p=0.000 n=12+10) BM_InlinedVectorFillString/0 0.93ns +- 0% 0.24ns +- 0% -74.19% (p=0.000 n=12+10) BM_InlinedVectorAssignments/1 10.5ns +- 0% 4.1ns +- 0% -60.64% (p=0.000 n=11+10) BM_InlinedVectorAssignments/2 10.7ns +- 0% 4.4ns +- 0% -59.08% (p=0.000 n=11+11) BM_CopyTrivial/0 0.92ns +- 0% 0.47ns +- 0% -48.91% (p=0.000 n=11+12) BM_CopyNonTrivial/8 21.5ns +- 1% 15.2ns +- 2% -29.23% (p=0.000 n=12+12) BM_StdVectorEmpty 0.47ns +- 1% 0.35ns +- 0% -24.73% (p=0.000 n=12+12) BM_StdVectorSize 0.46ns +- 2% 0.35ns +- 0% -24.32% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableOnly>/0 3.44ns +- 0% 2.76ns +- 1% -19.83% (p=0.000 n=11+11) BM_InlinedVectorFillRange/256 20.7ns +- 1% 17.8ns +- 0% -14.08% (p=0.000 n=12+9) BM_CopyNonTrivial/1 5.88ns +- 1% 5.51ns +- 0% -6.28% (p=0.000 n=10+8) BM_SwapElements<LargeCopyableMovable>/1 4.19ns +- 0% 3.95ns +- 1% -5.63% (p=0.000 n=11+12) BM_SwapElements<LargeCopyableMovableSwappable>/1 4.18ns +- 0% 3.99ns +- 0% -4.70% (p=0.000 n=9+11) BM_SwapElements<LargeCopyableMovable>/0 2.41ns +- 0% 2.31ns +- 0% -4.45% (p=0.000 n=12+12) BM_InlinedVectorFillRange/64 8.25ns +- 0% 8.04ns +- 0% -2.51% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableOnly>/1 82.4ns +- 0% 81.5ns +- 0% -1.06% (p=0.000 n=12+12) ``` Ones that get worse with this CL: ``` BM_CopyTrivial/1 0.92ns +- 0% 1.15ns +- 0% +25.00% (p=0.000 n=10+9) BM_CopyTrivial/8 8.57ns +- 0% 10.72ns +- 1% +25.16% (p=0.000 n=10+12) BM_SwapElements<LargeCopyableMovableSwappable>/512 1.48ns +- 1% 1.66ns +- 1% +11.88% (p=0.000 n=12+12) BM_InlinedVectorFillString/1 11.5ns +- 0% 12.8ns +- 1% +11.62% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableMovableSwappable>/64 1.48ns +- 2% 1.66ns +- 1% +11.66% (p=0.000 n=12+11) BM_SwapElements<LargeCopyableMovableSwappable>/1k 1.48ns +- 1% 1.65ns +- 2% +11.32% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableMovable>/512 1.48ns +- 2% 1.58ns +- 4% +6.62% (p=0.000 n=11+12) BM_SwapElements<LargeCopyableMovable>/1k 1.49ns +- 2% 1.58ns +- 3% +6.05% (p=0.000 n=12+12) BM_SwapElements<LargeCopyableMovable>/64 1.48ns +- 2% 1.57ns +- 4% +6.04% (p=0.000 n=11+12) BM_InlinedVectorFillRange/1 4.81ns +- 0% 5.05ns +- 0% +4.83% (p=0.000 n=11+11) BM_InlinedVectorFillString/8 79.4ns +- 1% 83.1ns +- 1% +4.64% (p=0.000 n=10+12) BM_StdVectorFillString/1 16.3ns +- 0% 16.6ns +- 0% +2.13% (p=0.000 n=11+8) ``` PiperOrigin-RevId: 353906786 -- 8e26518b3cec9c598e5e9573c46c3bd1b03a67ef by Abseil Team <absl-team@google.com>: Internal change PiperOrigin-RevId: 353737330 -- f206ae0983e58c9904ed8b8f05f9caf564a446be by Matt Kulukundis <kfm@google.com>: Import of CCTZ from GitHub. PiperOrigin-RevId: 353682256 GitOrigin-RevId: c68f1886f5e8fd90eb0c2d2e68feaf00a7cdacda Change-Id: I5790c1036c4f543c701d1039848fabf7ae881ad8
4 years ago
template <typename T, size_t N, typename A>
void Storage<T, N, A>::DestroyContents() {
pointer data = GetIsAllocated() ? GetAllocatedData() : GetInlinedData();
inlined_vector_internal::DestroyElements(GetAllocPtr(), data, GetSize());
DeallocateIfAllocated();
}
template <typename T, size_t N, typename A>
void Storage<T, N, A>::InitFrom(const Storage& other) {
const auto n = other.GetSize();
assert(n > 0); // Empty sources handled handled in caller.
const_pointer src;
pointer dst;
if (!other.GetIsAllocated()) {
dst = GetInlinedData();
src = other.GetInlinedData();
} else {
// Because this is only called from the `InlinedVector` constructors, it's
// safe to take on the allocation with size `0`. If `ConstructElements(...)`
// throws, deallocation will be automatically handled by `~Storage()`.
size_type new_capacity = ComputeCapacity(GetInlinedCapacity(), n);
dst = AllocatorTraits::allocate(*GetAllocPtr(), new_capacity);
SetAllocatedData(dst, new_capacity);
src = other.GetAllocatedData();
}
if (IsMemcpyOk::value) {
MemcpyIfAllowed<IsMemcpyOk::value>(dst, src, sizeof(dst[0]) * n);
} else {
auto values = IteratorValueAdapter<const_pointer>(src);
inlined_vector_internal::ConstructElements(GetAllocPtr(), dst, &values, n);
}
GetSizeAndIsAllocated() = other.GetSizeAndIsAllocated();
}
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
template <typename T, size_t N, typename A>
template <typename ValueAdapter>
auto Storage<T, N, A>::Initialize(ValueAdapter values, size_type new_size)
-> void {
// Only callable from constructors!
assert(!GetIsAllocated());
assert(GetSize() == 0);
pointer construct_data;
if (new_size > GetInlinedCapacity()) {
// Because this is only called from the `InlinedVector` constructors, it's
// safe to take on the allocation with size `0`. If `ConstructElements(...)`
// throws, deallocation will be automatically handled by `~Storage()`.
size_type new_capacity = ComputeCapacity(GetInlinedCapacity(), new_size);
construct_data = AllocatorTraits::allocate(*GetAllocPtr(), new_capacity);
SetAllocatedData(construct_data, new_capacity);
SetIsAllocated();
} else {
construct_data = GetInlinedData();
}
inlined_vector_internal::ConstructElements(GetAllocPtr(), construct_data,
&values, new_size);
// Since the initial size was guaranteed to be `0` and the allocated bit is
// already correct for either case, *adding* `new_size` gives us the correct
// result faster than setting it directly.
AddSize(new_size);
}
template <typename T, size_t N, typename A>
template <typename ValueAdapter>
auto Storage<T, N, A>::Assign(ValueAdapter values, size_type new_size) -> void {
StorageView storage_view = MakeStorageView();
AllocationTransaction allocation_tx(GetAllocPtr());
absl::Span<value_type> assign_loop;
absl::Span<value_type> construct_loop;
absl::Span<value_type> destroy_loop;
if (new_size > storage_view.capacity) {
size_type new_capacity = ComputeCapacity(storage_view.capacity, new_size);
construct_loop = {allocation_tx.Allocate(new_capacity), new_size};
destroy_loop = {storage_view.data, storage_view.size};
} else if (new_size > storage_view.size) {
assign_loop = {storage_view.data, storage_view.size};
construct_loop = {storage_view.data + storage_view.size,
new_size - storage_view.size};
} else {
assign_loop = {storage_view.data, new_size};
destroy_loop = {storage_view.data + new_size, storage_view.size - new_size};
}
inlined_vector_internal::AssignElements(assign_loop.data(), &values,
assign_loop.size());
inlined_vector_internal::ConstructElements(
GetAllocPtr(), construct_loop.data(), &values, construct_loop.size());
inlined_vector_internal::DestroyElements(GetAllocPtr(), destroy_loop.data(),
destroy_loop.size());
if (allocation_tx.DidAllocate()) {
DeallocateIfAllocated();
AcquireAllocatedData(&allocation_tx);
SetIsAllocated();
}
SetSize(new_size);
}
template <typename T, size_t N, typename A>
template <typename ValueAdapter>
auto Storage<T, N, A>::Resize(ValueAdapter values, size_type new_size) -> void {
StorageView storage_view = MakeStorageView();
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
auto* const base = storage_view.data;
const size_type size = storage_view.size;
auto* alloc = GetAllocPtr();
if (new_size <= size) {
// Destroy extra old elements.
inlined_vector_internal::DestroyElements(alloc, base + new_size,
size - new_size);
} else if (new_size <= storage_view.capacity) {
// Construct new elements in place.
inlined_vector_internal::ConstructElements(alloc, base + size, &values,
new_size - size);
} else {
// Steps:
// a. Allocate new backing store.
// b. Construct new elements in new backing store.
// c. Move existing elements from old backing store to now.
// d. Destroy all elements in old backing store.
// Use transactional wrappers for the first two steps so we can roll
// back if necessary due to exceptions.
AllocationTransaction allocation_tx(alloc);
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
size_type new_capacity = ComputeCapacity(storage_view.capacity, new_size);
pointer new_data = allocation_tx.Allocate(new_capacity);
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
ConstructionTransaction construction_tx(alloc);
construction_tx.Construct(new_data + size, &values, new_size - size);
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
IteratorValueAdapter<MoveIterator> move_values((MoveIterator(base)));
inlined_vector_internal::ConstructElements(alloc, new_data, &move_values,
size);
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
inlined_vector_internal::DestroyElements(alloc, base, size);
construction_tx.Commit();
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
DeallocateIfAllocated();
AcquireAllocatedData(&allocation_tx);
SetIsAllocated();
}
SetSize(new_size);
}
template <typename T, size_t N, typename A>
template <typename ValueAdapter>
auto Storage<T, N, A>::Insert(const_iterator pos, ValueAdapter values,
size_type insert_count) -> iterator {
StorageView storage_view = MakeStorageView();
size_type insert_index =
std::distance(const_iterator(storage_view.data), pos);
size_type insert_end_index = insert_index + insert_count;
size_type new_size = storage_view.size + insert_count;
if (new_size > storage_view.capacity) {
AllocationTransaction allocation_tx(GetAllocPtr());
ConstructionTransaction construction_tx(GetAllocPtr());
ConstructionTransaction move_construciton_tx(GetAllocPtr());
IteratorValueAdapter<MoveIterator> move_values(
MoveIterator(storage_view.data));
size_type new_capacity = ComputeCapacity(storage_view.capacity, new_size);
pointer new_data = allocation_tx.Allocate(new_capacity);
construction_tx.Construct(new_data + insert_index, &values, insert_count);
move_construciton_tx.Construct(new_data, &move_values, insert_index);
inlined_vector_internal::ConstructElements(
GetAllocPtr(), new_data + insert_end_index, &move_values,
storage_view.size - insert_index);
inlined_vector_internal::DestroyElements(GetAllocPtr(), storage_view.data,
storage_view.size);
construction_tx.Commit();
move_construciton_tx.Commit();
DeallocateIfAllocated();
AcquireAllocatedData(&allocation_tx);
SetAllocatedSize(new_size);
return iterator(new_data + insert_index);
} else {
size_type move_construction_destination_index =
(std::max)(insert_end_index, storage_view.size);
ConstructionTransaction move_construction_tx(GetAllocPtr());
IteratorValueAdapter<MoveIterator> move_construction_values(
MoveIterator(storage_view.data +
(move_construction_destination_index - insert_count)));
absl::Span<value_type> move_construction = {
storage_view.data + move_construction_destination_index,
new_size - move_construction_destination_index};
pointer move_assignment_values = storage_view.data + insert_index;
absl::Span<value_type> move_assignment = {
storage_view.data + insert_end_index,
move_construction_destination_index - insert_end_index};
absl::Span<value_type> insert_assignment = {move_assignment_values,
move_construction.size()};
absl::Span<value_type> insert_construction = {
insert_assignment.data() + insert_assignment.size(),
insert_count - insert_assignment.size()};
move_construction_tx.Construct(move_construction.data(),
&move_construction_values,
move_construction.size());
for (pointer destination = move_assignment.data() + move_assignment.size(),
last_destination = move_assignment.data(),
source = move_assignment_values + move_assignment.size();
;) {
--destination;
--source;
if (destination < last_destination) break;
*destination = std::move(*source);
}
inlined_vector_internal::AssignElements(insert_assignment.data(), &values,
insert_assignment.size());
inlined_vector_internal::ConstructElements(
GetAllocPtr(), insert_construction.data(), &values,
insert_construction.size());
move_construction_tx.Commit();
AddSize(insert_count);
return iterator(storage_view.data + insert_index);
}
}
template <typename T, size_t N, typename A>
template <typename... Args>
auto Storage<T, N, A>::EmplaceBack(Args&&... args) -> reference {
StorageView storage_view = MakeStorageView();
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
const auto n = storage_view.size;
if (ABSL_PREDICT_TRUE(n != storage_view.capacity)) {
// Fast path; new element fits.
pointer last_ptr = storage_view.data + n;
AllocatorTraits::construct(*GetAllocPtr(), last_ptr,
std::forward<Args>(args)...);
AddSize(1);
return *last_ptr;
}
// TODO(b/173712035): Annotate with musttail attribute to prevent regression.
return EmplaceBackSlow(std::forward<Args>(args)...);
}
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
template <typename T, size_t N, typename A>
template <typename... Args>
auto Storage<T, N, A>::EmplaceBackSlow(Args&&... args) -> reference {
StorageView storage_view = MakeStorageView();
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
AllocationTransaction allocation_tx(GetAllocPtr());
IteratorValueAdapter<MoveIterator> move_values(
MoveIterator(storage_view.data));
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
size_type new_capacity = NextCapacity(storage_view.capacity);
pointer construct_data = allocation_tx.Allocate(new_capacity);
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
pointer last_ptr = construct_data + storage_view.size;
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
// Construct new element.
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
AllocatorTraits::construct(*GetAllocPtr(), last_ptr,
std::forward<Args>(args)...);
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
// Move elements from old backing store to new backing store.
ABSL_INTERNAL_TRY {
inlined_vector_internal::ConstructElements(
GetAllocPtr(), allocation_tx.GetData(), &move_values,
storage_view.size);
}
ABSL_INTERNAL_CATCH_ANY {
AllocatorTraits::destroy(*GetAllocPtr(), last_ptr);
ABSL_INTERNAL_RETHROW;
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
}
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
// Destroy elements in old backing store.
inlined_vector_internal::DestroyElements(GetAllocPtr(), storage_view.data,
storage_view.size);
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
Export of internal Abseil changes -- 03700706d80f0939e2b5b8c02a326f045b643730 by Abseil Team <absl-team@google.com>: Reduced latency and code-size of some InlinedVector methods: 1. Simpler fast path for push_back/emplace_back. 2. Do not inline slow path of push-back/emplace_back. 3. Simplify resize implementation. Performance: A simple benchmark that does the following per iteration: ``` push_back on an InlinedVector<int64> push_back on an InlinedVector<bool> ``` Sees iteration time go from 4.3ns to 2.8ns and code size shrink from 1129 bytes to 175 bytes. PiperOrigin-RevId: 343335635 -- 16f74277a9e8bf228c164b053da8b8098f76de62 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 343332753 -- 886b6d5d0244783d309e34f03c21710f411e3cb3 by Abseil Team <absl-team@google.com>: Optimize `Status::Status`: When creating a status, we currently create an empty struct first, then assign fields. This is suboptimal: https://screenshot.googleplex.com/5HqDuFBKUEqrVgy. Relevant Benchmarks: ``` BM_StatusCopyError_Deep/threads:1 26.9ns ±13% 21.2ns ±16% -21.46% (p=0.000 n=15+15) BM_StatusCopyError_Deep/threads:2 32.0ns ±30% 25.6ns ±37% -20.17% (p=0.004 n=15+14) BM_StatusCopyError_Deep/threads:4 37.4ns ±84% 30.6ns ±58% -18.26% (p=0.029 n=15+15) BM_StatusCopyError_Deep/threads:8 47.2ns ±33% 33.5ns ±56% -28.91% (p=0.000 n=15+14) ``` PiperOrigin-RevId: 343303312 -- 2f9d945654292e8e52cad410fa41dae794cff42c by Abseil Team <absl-team@google.com>: Set SOVERSION for the installed libraries PiperOrigin-RevId: 343287682 -- 600bbfffe91cfbdc60b43cdad5619258298d0b0d by Abseil Team <absl-team@google.com>: Fix a typo in a comment (than -> that) PiperOrigin-RevId: 343187724 -- 310c82cd97b3f1f0d1ee93a0ee2b0aee828b2a93 by Abseil Team <absl-team@google.com>: Simplify unaligned memory access functions. The #ifdef to produce calls to __sanitizer_unaligned_load16 etc were needed in past versions of this code, when we were lying to the compiler about the alignment of the loads/stores, by using a reinterpret_cast. However, a year ago, absl switched to simply use memcpy. Sanitizers support this correctly by default, nothing extra is required. PiperOrigin-RevId: 343159883 -- bdf6fcf99180c371fda6ba8af82fd44656e372fa by Gennadiy Rozental <rogeeff@google.com>: Migrate usage flags to global variables instead of modeling them as Abseil Flags. Also introduce new semantic for --help=substring command line argument. PiperOrigin-RevId: 343019883 GitOrigin-RevId: 03700706d80f0939e2b5b8c02a326f045b643730 Change-Id: I4ad40dfa9606f8b8bfb2d91fd09e327105311bfb
4 years ago
DeallocateIfAllocated();
AcquireAllocatedData(&allocation_tx);
SetIsAllocated();
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
AddSize(1);
return *last_ptr;
}
template <typename T, size_t N, typename A>
auto Storage<T, N, A>::Erase(const_iterator from, const_iterator to)
-> iterator {
StorageView storage_view = MakeStorageView();
size_type erase_size = std::distance(from, to);
size_type erase_index =
std::distance(const_iterator(storage_view.data), from);
size_type erase_end_index = erase_index + erase_size;
IteratorValueAdapter<MoveIterator> move_values(
MoveIterator(storage_view.data + erase_end_index));
inlined_vector_internal::AssignElements(storage_view.data + erase_index,
&move_values,
storage_view.size - erase_end_index);
inlined_vector_internal::DestroyElements(
GetAllocPtr(), storage_view.data + (storage_view.size - erase_size),
erase_size);
SubtractSize(erase_size);
return iterator(storage_view.data + erase_index);
}
template <typename T, size_t N, typename A>
auto Storage<T, N, A>::Reserve(size_type requested_capacity) -> void {
StorageView storage_view = MakeStorageView();
if (ABSL_PREDICT_FALSE(requested_capacity <= storage_view.capacity)) return;
AllocationTransaction allocation_tx(GetAllocPtr());
IteratorValueAdapter<MoveIterator> move_values(
MoveIterator(storage_view.data));
size_type new_capacity =
ComputeCapacity(storage_view.capacity, requested_capacity);
pointer new_data = allocation_tx.Allocate(new_capacity);
inlined_vector_internal::ConstructElements(GetAllocPtr(), new_data,
&move_values, storage_view.size);
inlined_vector_internal::DestroyElements(GetAllocPtr(), storage_view.data,
storage_view.size);
DeallocateIfAllocated();
AcquireAllocatedData(&allocation_tx);
SetIsAllocated();
}
template <typename T, size_t N, typename A>
auto Storage<T, N, A>::ShrinkToFit() -> void {
// May only be called on allocated instances!
assert(GetIsAllocated());
StorageView storage_view{GetAllocatedData(), GetSize(),
GetAllocatedCapacity()};
if (ABSL_PREDICT_FALSE(storage_view.size == storage_view.capacity)) return;
AllocationTransaction allocation_tx(GetAllocPtr());
IteratorValueAdapter<MoveIterator> move_values(
MoveIterator(storage_view.data));
pointer construct_data;
if (storage_view.size > GetInlinedCapacity()) {
size_type new_capacity = storage_view.size;
construct_data = allocation_tx.Allocate(new_capacity);
} else {
construct_data = GetInlinedData();
}
ABSL_INTERNAL_TRY {
inlined_vector_internal::ConstructElements(GetAllocPtr(), construct_data,
&move_values, storage_view.size);
}
ABSL_INTERNAL_CATCH_ANY {
SetAllocatedData(storage_view.data, storage_view.capacity);
ABSL_INTERNAL_RETHROW;
}
inlined_vector_internal::DestroyElements(GetAllocPtr(), storage_view.data,
storage_view.size);
AllocatorTraits::deallocate(*GetAllocPtr(), storage_view.data,
storage_view.capacity);
if (allocation_tx.DidAllocate()) {
AcquireAllocatedData(&allocation_tx);
} else {
UnsetIsAllocated();
}
}
template <typename T, size_t N, typename A>
auto Storage<T, N, A>::Swap(Storage* other_storage_ptr) -> void {
using std::swap;
assert(this != other_storage_ptr);
if (GetIsAllocated() && other_storage_ptr->GetIsAllocated()) {
swap(data_.allocated, other_storage_ptr->data_.allocated);
} else if (!GetIsAllocated() && !other_storage_ptr->GetIsAllocated()) {
Storage* small_ptr = this;
Storage* large_ptr = other_storage_ptr;
if (small_ptr->GetSize() > large_ptr->GetSize()) swap(small_ptr, large_ptr);
for (size_type i = 0; i < small_ptr->GetSize(); ++i) {
swap(small_ptr->GetInlinedData()[i], large_ptr->GetInlinedData()[i]);
}
IteratorValueAdapter<MoveIterator> move_values(
MoveIterator(large_ptr->GetInlinedData() + small_ptr->GetSize()));
inlined_vector_internal::ConstructElements(
large_ptr->GetAllocPtr(),
small_ptr->GetInlinedData() + small_ptr->GetSize(), &move_values,
large_ptr->GetSize() - small_ptr->GetSize());
inlined_vector_internal::DestroyElements(
large_ptr->GetAllocPtr(),
large_ptr->GetInlinedData() + small_ptr->GetSize(),
large_ptr->GetSize() - small_ptr->GetSize());
} else {
Storage* allocated_ptr = this;
Storage* inlined_ptr = other_storage_ptr;
if (!allocated_ptr->GetIsAllocated()) swap(allocated_ptr, inlined_ptr);
StorageView allocated_storage_view{allocated_ptr->GetAllocatedData(),
allocated_ptr->GetSize(),
allocated_ptr->GetAllocatedCapacity()};
IteratorValueAdapter<MoveIterator> move_values(
MoveIterator(inlined_ptr->GetInlinedData()));
ABSL_INTERNAL_TRY {
inlined_vector_internal::ConstructElements(
inlined_ptr->GetAllocPtr(), allocated_ptr->GetInlinedData(),
&move_values, inlined_ptr->GetSize());
}
ABSL_INTERNAL_CATCH_ANY {
allocated_ptr->SetAllocatedData(allocated_storage_view.data,
allocated_storage_view.capacity);
ABSL_INTERNAL_RETHROW;
}
inlined_vector_internal::DestroyElements(inlined_ptr->GetAllocPtr(),
inlined_ptr->GetInlinedData(),
inlined_ptr->GetSize());
inlined_ptr->SetAllocatedData(allocated_storage_view.data,
allocated_storage_view.capacity);
}
swap(GetSizeAndIsAllocated(), other_storage_ptr->GetSizeAndIsAllocated());
swap(*GetAllocPtr(), *other_storage_ptr->GetAllocPtr());
}
Export of internal Abseil changes -- 0bfa836596a9c787a2f0bdc283011dd1f6810c6e by Benjamin Barenblat <bbaren@google.com>: Ignore missing CPU frequency on more architectures Linux on MIPS, PA-RISC, RISC-V, and SystemZ doesn’t expose the nominal CPU frequency via /sys, so don’t worry if `NominalCPUFrequency` returns 1.0 on those platforms. Some POWER machines expose the CPU frequency; others do not. Since we can’t predict which type of machine the tests will run on, simply disable testing for `NominalCPUFrequency` on POWER. PiperOrigin-RevId: 347079873 -- 492b6834ed4a07cbc3abccd846f7e37d8c556ee5 by Benjamin Barenblat <bbaren@google.com>: Use ABSL_HAVE_THREAD_LOCAL macro instead of copying code Reduce code duplication by checking the ABSL_HAVE_THREAD_LOCAL macro instead of copying code from base/config.h. PiperOrigin-RevId: 347079561 -- 8d656efce4da9cb032094377e58493d98427a536 by Abseil Team <absl-team@google.com>: Rollback PiperOrigin-RevId: 347078779 -- 221bc69ec6dd7e2777ffcff6942584f979ef6382 by Abseil Team <absl-team@google.com>: Add flag for 'shallow subcord' feature for experimental ring buffer rollout There is a potential trade-off of CPU cost vs over-sharing cord data for subcord of large cords. This flag allows making subcords shallow for ringbuffers (with a potential larger waste of referenced source cords), which allows us to make subcord fast for this apps that do no persist (unmodified / plain copied) sub cords. This change also introduces constants for the default settings, intended to keep the internal cord settings concistent with external flags. PiperOrigin-RevId: 347053271 -- 00a56c24293566734009f6bf2169a83fb37a35ba by Abseil Team <absl-team@google.com>: Revert the usage of variant<> in Cord iterator and reader. The introduction of the variant may lead to some missed compiler optimizations. PiperOrigin-RevId: 347053041 -- c7b7b5ed7e3ab46b1e75b80f1a7de0bda26c8f70 by Chris Kennelly <ckennelly@google.com>: Release library for integer power-of-2 functions and bit counting. PiperOrigin-RevId: 347035065 -- 5a035c0d9840b251967f9e7039fc6a4e01dd52f3 by Abseil Team <absl-team@google.com>: Restructure Cord::ChunkIterator for future ring buffer support. PiperOrigin-RevId: 346890054 GitOrigin-RevId: 0bfa836596a9c787a2f0bdc283011dd1f6810c6e Change-Id: I3a58e2a44cb4c6f2116c43e2a4ccbc319d3ccecf
4 years ago
// End ignore "maybe-uninitialized"
#if !defined(__clang__) && defined(__GNUC__)
#pragma GCC diagnostic pop
#endif
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
} // namespace inlined_vector_internal
ABSL_NAMESPACE_END
} // namespace absl
#endif // ABSL_CONTAINER_INTERNAL_INLINED_VECTOR_INTERNAL_H_