$ NUM_THREADS=1  ./rtoa-pgatram-allinone2b t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.498 secs
737201628 75552000

$ NUM_THREADS=4  ./rtoa-pgatram-allinone2b t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.176 secs
737201628 75552000

$ NUM_THREADS=8  ./rtoa-pgatram-allinone2b t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.124 secs
737201628 75552000

$ NUM_THREADS=16 ./rtoa-pgatram-allinone2b t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.096 secs
737201628 75552000

##</code><code>##

$ NUM_THREADS=1  ./rtoa-pgatram-allinone2c t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.488 secs
737201628 75552000

$ NUM_THREADS=4  ./rtoa-pgatram-allinone2c t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.143 secs
737201628 75552000

$ NUM_THREADS=8  ./rtoa-pgatram-allinone2c t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.091 secs
737201628 75552000

$ NUM_THREADS=16 ./rtoa-pgatram-allinone2c t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.065 secs
737201628 75552000

##</code><code>##

$ NUM_THREADS=1  ./rtoa-pgatram-allinone2d t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.489 secs
737201628 75552000

$ NUM_THREADS=4  ./rtoa-pgatram-allinone2d t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.144 secs
737201628 75552000

$ NUM_THREADS=8  ./rtoa-pgatram-allinone2d t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.075 secs
737201628 75552000

$ NUM_THREADS=16 ./rtoa-pgatram-allinone2d t1.txt t1.txt t1.txt t1.txt | cksum
do_it_all time    :    0.048 secs
737201628 75552000

##</code><code>##

// rtoa-pgatram-allinone2d.cpp. Crude allinone version.
//   based on rtoa-pgatram-allinone2.cpp https://perlmonks.org/?node_id=11152186
//
// Obtain the fast_io library (required dependency):
//   git clone --depth=1 https://github.com/cppfastio/fast_io
//
// Compile with g++ or clang++:
//   clang++ -o rtoa-pgatram-allinone2d -std=c++20 -fopenmp -Wall -O3 -I fast_io/include rtoa-pgatram-allinone2d.cpp
//
// OpenMP Little Book:
//   https://nanxiao.gitbooks.io/openmp-little-book/content/

#include <chrono>
#include <cstring>
#include <string>
#include <string_view>
#include <numeric>
#include <thread>

#ifdef _OPENMP
#include <omp.h>
#endif

#include <fstream>
#include <iostream>
#include <iomanip>
#include <semaphore>

// See [id://11149504] for more info on the fast_io C++ library
#include <fast_io.h>
#include <fast_io_legacy.h>

// ---------------------------------------------------------------

typedef std::chrono::high_resolution_clock high_resolution_clock;
typedef std::chrono::high_resolution_clock::time_point time_point;
typedef std::chrono::milliseconds milliseconds;

double elaspe_time(
   time_point cend,
   time_point cstart)
{
   return double (
      std::chrono::duration_cast<milliseconds>(cend - cstart).count()
   ) * 1e-3;
}

// ---------------------------------------------------------------

// Though there are less than 256 initializers in this ascii table,
// the others are guaranteed by ANSI C to be initialized to zero.
static const int romtab[256] = {
    0,0,0,0,0,0,   0,   0,   0,   0,  //  00- 09
    0,0,0,0,0,0,   0,   0,   0,   0,  //  10- 19
    0,0,0,0,0,0,   0,   0,   0,   0,  //  20- 29
    0,0,0,0,0,0,   0,   0,   0,   0,  //  30- 39
    0,0,0,0,0,0,   0,   0,   0,   0,  //  40- 49
    0,0,0,0,0,0,   0,   0,   0,   0,  //  50- 59
    0,0,0,0,0,0,   0, 100, 500,   0,  //  60- 69
    0,0,0,1,0,0,  50,1000,   0,   0,  //  70- 79
    0,0,0,0,0,0,   5,   0,  10,   0,  //  80- 89
    0,0,0,0,0,0,   0,   0,   0, 100,  //  90- 99
  500,0,0,0,0,1,   0,   0,  50,1000,  // 100-109
    0,0,0,0,0,0,   0,   0,   5,   0,  // 110-119
   10,0,0,0,0,0,   0,   0,   0,   0   // 120-129
};

// Return the arabic number for a roman letter c.
// Return zero if the roman letter c is invalid.
inline int urtoa(int c) { return romtab[c]; }

inline int accfn(int t, char c) {
   return t + urtoa(c) - t % urtoa(c) * 2;
}

inline int roman_to_dec(std::string_view st) {
   return std::accumulate(st.begin(), st.end(), 0, accfn);
}

// ---------------------------------------------------------------

inline constexpr auto CHUNK_SIZE  = 1048576;
inline constexpr auto LINE_LENGTH = 255;

// Helper function to find '\n'.
inline constexpr char const* find_lf(char const* first, char const* last)
{
   while (first != last) {
      if (*first == '\n') break;
      ++first;
   }
   return first;
}

// Read an input file of Roman Numerals and do it all.
static void do_it_all(
   std::ifstream&  fin,     // in: file input stream containing a list of Roman Numerals
   int             nthds    // in: number of threads
)
{
   fast_io::out_buf_type obf{fast_io::out()};
   std::binary_semaphore *sem[nthds];
   size_t next_chunk_id = 0;

   // Create semaphores for orderly output.
   for (int i = 0; i < nthds; ++i)
      sem[i] = new std::binary_semaphore{0};

   #pragma omp parallel
   {
      std::string buf; buf.resize(CHUNK_SIZE + LINE_LENGTH + 1, '\0');
      const char *first, *last;
      size_t chunk_id, len;

      while (fin.good()) {
         std::string output;
         len = 0;

         // Read the next chunk serially.
         //
         #pragma omp critical
         {
            fin.read(&buf[0], CHUNK_SIZE);
            if ((len = fin.gcount()) > 0) {
               chunk_id = ++next_chunk_id;
               if (buf[len - 1] != '\n' && fin.getline(&buf[len], LINE_LENGTH)) {
                  // Getline discards the newline char and appends null char.
                  // Therefore, change '\0' to '\n'.
                  len += fin.gcount();
                  buf[len - 1] = '\n';
               }
            }
         }

         if (!len) break;

         buf[len] = '\0';
         first    = &buf[0];
         last     = &buf[len];

         // Process max Nthreads chunks concurrently.
         //
         while (first < last) {
            auto beg_ptr{first}; first = find_lf(first, last);
            auto end_ptr{first};

            int dec = roman_to_dec(std::string_view(beg_ptr, end_ptr - beg_ptr));
            output.append(fast_io::concatln(dec));

            ++first;
         }

         // Output completed chunk, orderly by chunk_id.
         //
         if (nthds > 1 && chunk_id > 1)
            sem[chunk_id % nthds]->acquire();

         fast_io::io::print(obf, output);

         if (nthds > 1)
            sem[(chunk_id + 1) % nthds]->release();
      }
   }

   // Destroy the dynamically allocated semaphores.
   for (int i = 0; i < nthds; ++i)
      delete sem[i];
}

int main(int argc, char* argv[])
{
   if (argc < 2) {
      if (argc > 0)
         std::cerr << "Usage: rtoa-pgatram-allinone2d file... >out.txt\n";
      return 1;
   }

   std::cerr << std::setprecision(3) << std::setiosflags(std::ios::fixed);
   time_point cstartall, cendall;
   cstartall = high_resolution_clock::now();

#ifdef _OPENMP
   // Determine the number of threads.
   int nthds = std::thread::hardware_concurrency();
   const char* env_nthds1 = std::getenv("OMP_NUM_THREADS");
   const char* env_nthds2 = std::getenv("NUM_THREADS");

   if (env_nthds1 && strlen(env_nthds1))
      nthds = ::atoi(env_nthds1);
   else if (env_nthds2 && strlen(env_nthds2))
      nthds = ::atoi(env_nthds2);

   omp_set_dynamic(false);
   omp_set_num_threads(nthds);
#else
   int nthds = 1;
#endif

   int    nfiles = argc - 1;
   char** fname  = &argv[1];

   for (int i = 0; i < nfiles; ++i) {
      std::ifstream fin(fname[i], std::ifstream::binary);
      if (!fin.is_open()) {
         std::cerr << "Error opening '" << fname[i] << "' : " << strerror(errno) << '\n';
         continue;
      }
      do_it_all(fin, nthds);
      fin.close();
   }

   cendall = high_resolution_clock::now();
   double ctakenall = elaspe_time(cendall, cstartall);
   std::cerr << "do_it_all time    : " << std::setw(8) << ctakenall << " secs\n";

   return 0;
}