brintos

brintos / llvm-project-archived public Read only

0
0
Text · 4.5 KiB · 4f97740 Raw
140 lines · c
1//===-- Data structures for sorting routines --------------------*- C++ -*-===//2//3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.4// See https://llvm.org/LICENSE.txt for license information.5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception6//7//===----------------------------------------------------------------------===//8 9#ifndef LLVM_LIBC_SRC_STDLIB_QSORT_DATA_H10#define LLVM_LIBC_SRC_STDLIB_QSORT_DATA_H11 12#include "hdr/stdint_proxy.h"13#include "src/__support/CPP/cstddef.h"14#include "src/__support/macros/config.h"15#include "src/string/memory_utils/inline_memcpy.h"16 17namespace LIBC_NAMESPACE_DECL {18namespace internal {19 20class ArrayGenericSize {21  cpp::byte *array_base;22  size_t array_len;23  size_t elem_size;24 25  LIBC_INLINE cpp::byte *get_internal(size_t i) const {26    return array_base + (i * elem_size);27  }28 29public:30  LIBC_INLINE ArrayGenericSize(void *a, size_t s, size_t e)31      : array_base(reinterpret_cast<cpp::byte *>(a)), array_len(s),32        elem_size(e) {}33 34  static constexpr bool has_fixed_size() { return false; }35 36  LIBC_INLINE void *get(size_t i) const { return get_internal(i); }37 38  LIBC_INLINE void swap(size_t i, size_t j) const {39    // It's possible to use 8 byte blocks with `uint64_t`, but that40    // generates more machine code as the remainder loop gets41    // unrolled, plus 4 byte operations are more likely to be42    // efficient on a wider variety of hardware. On x86 LLVM tends43    // to unroll the block loop again into 2 16 byte swaps per44    // iteration which is another reason that 4 byte blocks yields45    // good performance even for big types.46    using block_t = uint32_t;47    constexpr size_t BLOCK_SIZE = sizeof(block_t);48 49    alignas(block_t) cpp::byte tmp_block[BLOCK_SIZE];50 51    cpp::byte *elem_i = get_internal(i);52    cpp::byte *elem_j = get_internal(j);53 54    const size_t elem_size_rem = elem_size % BLOCK_SIZE;55    const cpp::byte *elem_i_block_end = elem_i + (elem_size - elem_size_rem);56 57    while (elem_i != elem_i_block_end) {58      inline_memcpy(tmp_block, elem_i, BLOCK_SIZE);59      inline_memcpy(elem_i, elem_j, BLOCK_SIZE);60      inline_memcpy(elem_j, tmp_block, BLOCK_SIZE);61 62      elem_i += BLOCK_SIZE;63      elem_j += BLOCK_SIZE;64    }65 66    for (size_t n = 0; n < elem_size_rem; ++n) {67      cpp::byte tmp = elem_i[n];68      elem_i[n] = elem_j[n];69      elem_j[n] = tmp;70    }71  }72 73  LIBC_INLINE size_t len() const { return array_len; }74 75  // Make an Array starting at index |i| and length |s|.76  LIBC_INLINE ArrayGenericSize make_array(size_t i, size_t s) const {77    return ArrayGenericSize(get_internal(i), s, elem_size);78  }79 80  // Reset this Array to point at a different interval of the same81  // items starting at index |i|.82  LIBC_INLINE void reset_bounds(size_t i, size_t s) {83    array_base = get_internal(i);84    array_len = s;85  }86};87 88// Having a specialized Array type for sorting that knows at89// compile-time what the size of the element is, allows for much more90// efficient swapping and for cheaper offset calculations.91template <size_t ELEM_SIZE> class ArrayFixedSize {92  cpp::byte *array_base;93  size_t array_len;94 95  LIBC_INLINE cpp::byte *get_internal(size_t i) const {96    return array_base + (i * ELEM_SIZE);97  }98 99public:100  LIBC_INLINE ArrayFixedSize(void *a, size_t s)101      : array_base(reinterpret_cast<cpp::byte *>(a)), array_len(s) {}102 103  // Beware this function is used a heuristic for cheap to swap types, so104  // instantiating `ArrayFixedSize` with `ELEM_SIZE > 100` is probably a bad105  // idea perf wise.106  static constexpr bool has_fixed_size() { return true; }107 108  LIBC_INLINE void *get(size_t i) const { return get_internal(i); }109 110  LIBC_INLINE void swap(size_t i, size_t j) const {111    alignas(32) cpp::byte tmp[ELEM_SIZE];112 113    cpp::byte *elem_i = get_internal(i);114    cpp::byte *elem_j = get_internal(j);115 116    inline_memcpy(tmp, elem_i, ELEM_SIZE);117    __builtin_memmove(elem_i, elem_j, ELEM_SIZE);118    inline_memcpy(elem_j, tmp, ELEM_SIZE);119  }120 121  LIBC_INLINE size_t len() const { return array_len; }122 123  // Make an Array starting at index |i| and length |s|.124  LIBC_INLINE ArrayFixedSize<ELEM_SIZE> make_array(size_t i, size_t s) const {125    return ArrayFixedSize<ELEM_SIZE>(get_internal(i), s);126  }127 128  // Reset this Array to point at a different interval of the same129  // items starting at index |i|.130  LIBC_INLINE void reset_bounds(size_t i, size_t s) {131    array_base = get_internal(i);132    array_len = s;133  }134};135 136} // namespace internal137} // namespace LIBC_NAMESPACE_DECL138 139#endif // LLVM_LIBC_SRC_STDLIB_QSORT_DATA_H140