更新最后添加比较和基准

这是另一个基于完美哈希的方法.完美的哈希是使用 gperf 生成的(如下所述:10 小时前一时兴起.如果您真的想要高吞吐量，完美的哈希是一个不错的开始，但可以考虑手动滚动基于 SIMD 的解决方案

When changing a Binary string to Hex, I could only do it to a certain size based off of the answers I found. But I want to change MASSIVE Binary strings into their complete Hex counterpart in a more efficient way than this which is the only way I've come across that does it completely:

for(size_t i = 0; i < (binarySubVec.size() - 1); i++){
    string binToHex, tmp = "0000";
    for (size_t j = 0; j < binaryVecStr[i].size(); j += 4){
        tmp = binaryVecStr[i].substr(j, 4);
        if      (!tmp.compare("0000")) binToHex += "0";
        else if (!tmp.compare("0001")) binToHex += "1";
        else if (!tmp.compare("0010")) binToHex += "2";
        else if (!tmp.compare("0011")) binToHex += "3";
        else if (!tmp.compare("0100")) binToHex += "4";
        else if (!tmp.compare("0101")) binToHex += "5";
        else if (!tmp.compare("0110")) binToHex += "6";
        else if (!tmp.compare("0111")) binToHex += "7";
        else if (!tmp.compare("1000")) binToHex += "8";
        else if (!tmp.compare("1001")) binToHex += "9";
        else if (!tmp.compare("1010")) binToHex += "A";
        else if (!tmp.compare("1011")) binToHex += "B";
        else if (!tmp.compare("1100")) binToHex += "C";
        else if (!tmp.compare("1101")) binToHex += "D";
        else if (!tmp.compare("1110")) binToHex += "E";
        else if (!tmp.compare("1111")) binToHex += "F";
        else continue;
    }
    hexOStr << binToHex;
    hexOStr << " ";
}

Its thorough and absolute, but slow.

Is there a simpler way of doing this?

UPDATE Added comparison and benchmarks at the end

Here's another take, based on a perfect hash. The perfect hash was generated using gperf (like described here: Is it possible to map string to int faster than using hashmap?).

I've further optimized by moving function local statics out of the way and marking hexdigit() and hash() as constexpr. This removes unnecessary any initialization overhead and gives the compiler full room for optimization/

I don't expect things to get much faster than this.

You could try reading e.g. 1024 nibbles at a time if possible, and give the compiler a chance to vectorize the operations using AVX/SSE instruction sets. (I have not inspected the generated code to see whether this would happen.)

The full sample code to translate std::cin to std::cout in streaming mode is:

#include <iostream>

int main()
{
    char buffer[4096];
    while (std::cin.read(buffer, sizeof(buffer)), std::cin.gcount())
    {
        size_t got = std::cin.gcount();
        char* out = buffer;

        for (auto it = buffer; it < buffer+got; it += 4)
            *out++ = Perfect_Hash::hexchar(it);

        std::cout.write(buffer, got/4);
    }
}

Here's the Perfect_Hash class, slightly redacted and extended with the hexchar lookup. Note that it does validate input in DEBUG builds using the assert:

Live On Coliru

#include <array>
#include <algorithm>
#include <cassert>

class Perfect_Hash {
    /* C++ code produced by gperf version 3.0.4 */
    /* Command-line: gperf -L C++ -7 -C -E -m 100 table  */
    /* Computed positions: -k'1-4' */

    /* maximum key range = 16, duplicates = 0 */
  private:
      static constexpr unsigned char asso_values[] = {
          27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27,
          27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 15, 7,  3,  1,  0,  27,
          27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27,
          27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27,
          27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27};
      template <typename It>
      static constexpr unsigned int hash(It str)
      {
          return 
              asso_values[(unsigned char)str[3] + 2] + asso_values[(unsigned char)str[2] + 1] +
              asso_values[(unsigned char)str[1] + 3] + asso_values[(unsigned char)str[0]];
      }

      static constexpr char hex_lut[] = "???????????fbead9c873625140";
  public:
#ifdef DEBUG
    template <typename It>
    static char hexchar(It binary_nibble)
    {
        assert(Perfect_Hash::validate(binary_nibble)); // for DEBUG only
        return hex_lut[hash(binary_nibble)]; // no validation!
    }
#else
    template <typename It>
    static constexpr char hexchar(It binary_nibble)
    {
        return hex_lut[hash(binary_nibble)]; // no validation!
    }
#endif
    template <typename It>
    static bool validate(It str)
    {
        static constexpr std::array<char, 4> vocab[] = {
            {{'?', '?', '?', '?'}}, {{'?', '?', '?', '?'}}, {{'?', '?', '?', '?'}},
            {{'?', '?', '?', '?'}}, {{'?', '?', '?', '?'}}, {{'?', '?', '?', '?'}},
            {{'?', '?', '?', '?'}}, {{'?', '?', '?', '?'}}, {{'?', '?', '?', '?'}},
            {{'?', '?', '?', '?'}}, {{'?', '?', '?', '?'}},
            {{'1', '1', '1', '1'}}, {{'1', '0', '1', '1'}},
            {{'1', '1', '1', '0'}}, {{'1', '0', '1', '0'}},
            {{'1', '1', '0', '1'}}, {{'1', '0', '0', '1'}},
            {{'1', '1', '0', '0'}}, {{'1', '0', '0', '0'}},
            {{'0', '1', '1', '1'}}, {{'0', '0', '1', '1'}},
            {{'0', '1', '1', '0'}}, {{'0', '0', '1', '0'}},
            {{'0', '1', '0', '1'}}, {{'0', '0', '0', '1'}},
            {{'0', '1', '0', '0'}}, {{'0', '0', '0', '0'}},
        }; 
        int key = hash(str);

        if (key <= 26 && key >= 0)
            return std::equal(str, str+4, vocab[key].begin());
        else
            return false;
    }
};

constexpr unsigned char Perfect_Hash::asso_values[];
constexpr char Perfect_Hash::hex_lut[];

#include <iostream>

int main()
{
    char buffer[4096];
    while (std::cin.read(buffer, sizeof(buffer)), std::cin.gcount())
    {
        size_t got = std::cin.gcount();
        char* out = buffer;

        for (auto it = buffer; it < buffer+got; it += 4)
            *out++ = Perfect_Hash::hexchar(it);

        std::cout.write(buffer, got/4);
    }
}

Demo output for e.g. od -A none -t o /dev/urandom | tr -cd '01' | dd bs=1 count=4096 | ./test

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

BENCHMARKS

I came up with three different approaches:

1. naive.cpp (no hacks, no libraries); Live disassembly on Godbolt
2. spirit.cpp (Trie); Live disassembly on pastebin
3. and this answer: perfect.cpp hash based; Live disassembly on Godbolt

In order to do some comparisons, I've

• compiled them all with the same compiler (GCC 4.9) and flags (-O3 -march=native -g0 -DNDEBUG)
• optimized input/output so it doesn't read by 4 chars/write single characters
• created a large input file (1 Gigabyte)

Here are the results:

• Surprisingly, the naive approach from the first answer does rather well
• Spirit does really badly here; it nets 3.4MB/s so that the whole file would take at 294 seconds (!!!). We've left it off the charts
• The average throughputs are ~720MB/s for naive.cpp and ~1.14GB/s forperfect.cpp
• This makes the perfect hash approach roughly 50% faster than the naive approach.

*Summary I'd say the naive approach was plenty good as I posted it on whim 10 hours ago. If you really want high throughput, the perfect hash is a nice start, but consider hand-rolling a SIMD based solution