daw_json_simd_avx2.h

Go to the documentation of this file.

// Copyright (c) Darrell Wright
//
// Distributed under the Boost Software License, Version 1.0. (See accompanying
// file LICENSE or copy at http://www.boost.org/LICENSE_1_0.txt)
//
// Official repository: https://github.com/beached/daw_json_link
//
 
#pragma once
 
#include <daw/daw_cpp_feature_check.h>
 
#define DAW_JSON_HAS_AVX2_SIMD 1
 
#include <immintrin.h>
 
#include <cstddef>
#include <cstdint>
#include <cstring>
#include <type_traits>
 
namespace daw::simd_impl::avx2 {
  class mask {
    __m256i m_value;
 
  public:
    explicit mask( __m256i value ) noexcept
      : m_value( value ) {}
 
    [[nodiscard]] std::uint64_t to_ullong( ) const noexcept {
      return static_cast<std::uint32_t>( _mm256_movemask_epi8( m_value ) );
    }
 
    [[nodiscard]] friend mask operator|( mask lhs, mask rhs ) noexcept {
      return mask( _mm256_or_si256( lhs.m_value, rhs.m_value ) );
    }
 
    [[nodiscard]] friend mask operator&( mask lhs, mask rhs ) noexcept {
      return mask( _mm256_and_si256( lhs.m_value, rhs.m_value ) );
    }
 
    [[nodiscard]] friend mask operator!( mask value ) noexcept {
      return mask( _mm256_xor_si256(
        value.m_value, _mm256_set1_epi8( static_cast<char>( 0xFF ) ) ) );
    }
  };
 
  template<typename T, std::size_t /*MaximumSize*/ = 32U>
  class vec {
    static_assert( sizeof( T ) == 1,
                   "The AVX2 JSON SIMD implementation supports byte types" );
    static_assert(
      std::is_integral_v<T>,
      "The AVX2 JSON SIMD implementation requires an integral type" );
 
    __m256i m_value;
 
    template<typename U, std::size_t MaximumSize>
    friend class vec;
 
    template<typename Simd, typename Range, typename Flag>
    friend Simd unchecked_load( Range values, Flag );
 
    template<typename Simd, typename Range, typename Flag>
    friend Simd partial_load( Range values, Flag );
 
    explicit vec( __m256i value ) noexcept
      : m_value( value ) {}
 
    [[nodiscard]] static __m256i order_bytes( __m256i value ) noexcept {
      return _mm256_xor_si256(
        value, _mm256_set1_epi8( static_cast<char>( 0x80 ) ) );
    }
 
  public:
    using value_type = T;
    using mask_type = mask;
 
    explicit vec( value_type value ) noexcept
      : m_value( _mm256_set1_epi8( static_cast<char>( value ) ) ) {}
 
    [[nodiscard]] static constexpr std::size_t size( ) noexcept {
      return 32;
    }
 
    [[nodiscard]] friend mask operator==( vec lhs, vec rhs ) noexcept {
      return mask( _mm256_cmpeq_epi8( lhs.m_value, rhs.m_value ) );
    }
 
    [[nodiscard]] friend mask operator>=( vec lhs, vec rhs ) noexcept {
      auto const ordered_lhs = order_bytes( lhs.m_value );
      auto const ordered_rhs = order_bytes( rhs.m_value );
      return mask( _mm256_or_si256(
        _mm256_cmpgt_epi8( ordered_lhs, ordered_rhs ),
        _mm256_cmpeq_epi8( lhs.m_value, rhs.m_value ) ) );
    }
 
    [[nodiscard]] friend mask operator<=( vec lhs, vec rhs ) noexcept {
      auto const ordered_lhs = order_bytes( lhs.m_value );
      auto const ordered_rhs = order_bytes( rhs.m_value );
      return mask( _mm256_or_si256(
        _mm256_cmpgt_epi8( ordered_rhs, ordered_lhs ),
        _mm256_cmpeq_epi8( lhs.m_value, rhs.m_value ) ) );
    }
  };
 
  struct flag_default_t {};
  struct flag_convert_t {};
 
  inline constexpr flag_default_t flag_default = { };
  inline constexpr flag_convert_t flag_convert = { };
 
  template<typename Simd, typename Range, typename Flag = flag_default_t>
  [[nodiscard]] inline Simd unchecked_load( Range values,
                                            Flag = flag_default ) {
    static_assert( Simd::size( ) == 32 );
    return Simd( _mm256_loadu_si256(
      reinterpret_cast<__m256i const *>( values.data( ) ) ) );
  }
 
  template<typename Simd, typename Range, typename Flag = flag_default_t>
  [[nodiscard]] inline Simd partial_load( Range values, Flag = flag_default ) {
    static_assert( Simd::size( ) == 32 );
    auto result = _mm256_setzero_si256( );
    std::memcpy( &result, values.data( ), values.size( ) );
    return Simd( result );
  }
} // namespace daw::simd_impl::avx2