2

我在CRAN上发布了一个包，它通过 RcppParallel 框架使用多个内核。它在r-devel-linux-x86_64-fedora-clang和r-patched-solaris-x86上安装有问题。我收到以下错误消息（有几条与 std::transform 相关的类似消息，因此为简洁起见，我只介绍其中一条）：

1.对于 r-patched-solaris-x86：

ParallelFunctions.cpp: In member function ‘virtual void ParallelVectorExpStruct::operator()(std::size_t, std::size_t)’:
ParallelFunctions.cpp:134:27: error: no matching function for call to ‘transform(RcppParallel::RVector<double>::const_iterator, RcppParallel::RVector<double>::const_iterator, RcppParallel::RVector<double>::iterator, <unresolved overloaded function type>)’
                      ::exp);
                           ^
In file included from /opt/csw/include/c++/5.2.0/algorithm:62:0,
                 from /home/ripley/R/Lib32/Rcpp/include/RcppCommon.h:63,
                 from /home/ripley/R/Lib32/RcppArmadillo/include/RcppArmadilloForward.h:26,
                 from /home/ripley/R/Lib32/RcppArmadillo/include/RcppArmadillo.h:31,
                 from ParallelFunctions.h:4,
                 from ParallelFunctions.cpp:1:
/opt/csw/include/c++/5.2.0/bits/stl_algo.h:4164:5: note: candidate: template<class _IIter, class _OIter, class _UnaryOperation> _OIter std::transform(_IIter, _IIter, _OIter, _UnaryOperation)
     transform(_InputIterator __first, _InputIterator __last,
     ^
/opt/csw/include/c++/5.2.0/bits/stl_algo.h:4164:5: note:   template argument deduction/substitution failed:
ParallelFunctions.cpp:134:27: note:   couldn't deduce template parameter ‘_UnaryOperation’
                      ::exp);
                           ^
In file included from /opt/csw/include/c++/5.2.0/algorithm:62:0,
                 from /home/ripley/R/Lib32/Rcpp/include/RcppCommon.h:63,
                 from /home/ripley/R/Lib32/RcppArmadillo/include/RcppArmadilloForward.h:26,
                 from /home/ripley/R/Lib32/RcppArmadillo/include/RcppArmadillo.h:31,
                 from ParallelFunctions.h:4,
                 from ParallelFunctions.cpp:1:
/opt/csw/include/c++/5.2.0/bits/stl_algo.h:4201:5: note: candidate: template<class _IIter1, class _IIter2, class _OIter, class _BinaryOperation> _OIter std::transform(_IIter1, _IIter1, _IIter2, _OIter, _BinaryOperation)
     transform(_InputIterator1 __first1, _InputIterator1 __last1,
     ^
/opt/csw/include/c++/5.2.0/bits/stl_algo.h:4201:5: note:   template argument deduction/substitution failed:
ParallelFunctions.cpp:134:27: note:   candidate expects 5 arguments, 4 provided
                      ::exp);
                           ^

2.对于 r-devel-linux-x86_64-fedora-clang：

hpaML.cpp:754:45: warning: explicitly assigning value of variable of type 'Rcpp::NumericVector' (aka 'Vector<14>') to itself [-Wself-assign-overloaded]
                           mean_ind, sd_ind = sd_ind,
                                     ~~~~~~ ^ ~~~~~~
ParallelFunctions.cpp:46:7: error: no matching function for call to 'transform'
      std::transform(input.begin() + begin, 
      ^~~~~~~~~~~~~~
/usr/local/bin/../include/c++/v1/algorithm:1955:1: note: candidate template ignored: couldn't infer template argument '_BinaryOperation'
transform(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2,
^
/usr/local/bin/../include/c++/v1/algorithm:1945:1: note: candidate function template not viable: requires 4 arguments, but 5 were provided
transform(_InputIterator __first, _InputIterator __last, _OutputIterator __result, _UnaryOperation __op)
^

这是调用 std::tansform 和 std::exp 函数的函数的代码：

// Parallel exp of vectors
struct ParallelVectorExpStruct : public Worker
{
  // source matrix
  const RVector<double> input;
  
  // destination matrix
  RVector<double> output;
  
  // initialize with source and destination
  ParallelVectorExpStruct(const NumericVector input, NumericVector output) 
    : input(input), output(output) {}
  
  // take the exponents of the range of elements requested
  void operator()(std::size_t begin, std::size_t end) {
      std::transform(input.begin() + begin, 
                     input.begin() + end,
                     output.begin() + begin, 
                     ::exp);
  }
};

// Parallel exponent of vector elements
NumericVector ParallelVectorExp(NumericVector x) 
{
  // allocate the output matrix
  NumericVector output(x.size());
  
  // ParallelVectorPowStruct functor
  ParallelVectorExpStruct parallelVectorExpStruct(x, output);
  
  // call parallelFor to do the work
  parallelFor(0, x.length(), parallelVectorExpStruct);
  
  // return the output matrix
  return (output);
}

我的描述文件包括 SystemRequirements: GNU make

我的 makevars 文件具有以下标志

CXX_STD = CXX11
PKG_CXXFLAGS = $(SHLIB_OPENMP_CXXFLAGS) 
PKG_LIBS = $(SHLIB_OPENMP_CXXFLAGS) $(LAPACK_LIBS) $(BLAS_LIBS) $(FLIBS)
PKG_LIBS += $(shell ${R_HOME}/bin/Rscript -e "RcppParallel::RcppParallelLibs()")

请帮我弄清楚如何解决错误。将是非常伟大的帮助！

4

1 回答 1

2

我找到了一个奇怪的解决方案。这个想法是制作以下包装函数：

double exp_parallel(double x)
{
  return std::exp(x);
}

然后我用exp_parallel替换exp产生：

// Parallel exp of vectors
struct ParallelVectorExpStruct : public Worker
{
  // source matrix
  const RVector<double> input;
  
  // destination matrix
  RVector<double> output;
  
  // initialize with source and destination
  ParallelVectorExpStruct(const NumericVector input, NumericVector output) 
    : input(input), output(output) {}
  
  // take the exponents of the range of elements requested
  void operator()(std::size_t begin, std::size_t end) {
      std::transform(input.begin() + begin, 
                     input.begin() + end,
                     output.begin() + begin, 
                     ::exp_parallel);
  }
};

可能原因是某些系统无法区分std::exp和Rcpp::exp函数。

于 2020-06-28T11:42:39.287 回答