Question

我有一个函数名.execute()，其中两个参数是markovchainListRcpp，另一个是我们想要的list。

让我们看一个例子

number of sequences

查看> mclist [[1]] Unnamed Markov chain A 3 - dimensional discrete Markov Chain with following states: a, b, c The transition matrix (by rows) is defined as follows: a b c a 0.2 0.5 0.3 b 0.0 0.2 0.8 c 0.1 0.8 0.1 [[2]] Unnamed Markov chain A 3 - dimensional discrete Markov Chain with following states: a, b, c The transition matrix (by rows) is defined as follows: a b c a 0.2 0.5 0.3 b 0.0 0.2 0.8 c 0.1 0.8 0.1 [[3]] Unnamed Markov chain A 3 - dimensional discrete Markov Chain with following states: a, b, c The transition matrix (by rows) is defined as follows: a b c a 0.2 0.5 0.3 b 0.0 0.2 0.8 c 0.1 0.8 0.1

的输出

markovchainListRcpp

我们得到两个序列：

> markovchainListRcpp(n = 2, object = mclist)
[[1]]
[1] 1 1 1 2 2 2

[[2]]
[1] "b" "c" "c" "a" "b" "b"

这些序列彼此独立。所以我想使用"b" "c" "c" (1) "a" "b" "b" (2)并行生成这些序列。

以下是我RcppParallel实现并行生成n个序列

RCppParallel

看看当我调用并行功能时会发生什么。

struct SequenceMC : public Worker
{   
  // source
  const List tmatList;
  const int num_tmat;

  // store results
  List output;

  // constructors
  SequenceMC(const List tmat, const int n) : tmatList(tmat), num_tmat(n), output() {}
  SequenceMC(SequenceMC &sequenceMC, Split) : tmatList(sequenceMC.tmatList), num_tmat(sequenceMC.num_tmat), output(){}

  // process just the elements of the range I've been asked to
  void operator()(std::size_t begin, std::size_t end) {
    List temp;
    CharacterVector helper;
    int n = end-begin;
    int m = num_tmat;

    temp = markovchainListRcpp(n, tmatList, false); 
    helper = temp[1];

    for(int i = 0;i < n;i++) {
      CharacterVector help(num_tmat);
      for(int j = 0;j < m;j++) {
        help[j] = helper[m*i+j];
      }  
      output.push_back(help);
    }
  }

  void join(SequenceMC& rhs) { 
    List temp = rhs.output;
    for(int i = 0;i < temp.size();i++) {
      output.push_back(temp[i]);
    }
  }
};

// [[Rcpp::export]]
List markovchainSequenceParallelRcpp(List object, int n) {

  List tmatList = object;

  SequenceMC sequenceMC(tmatList, tmatList.size());

  // call it with parallelReduce
  parallelReduce(0, n, sequenceMC);

  return sequenceMC.output;
}

我得到了结果。

> markovchainSequenceParallelRcpp(mclist, 5)
[[1]]
[1] "b" "c" "c"

[[2]]
[1] "b" "c" "c"

[[3]]
[1] "b" "c" "b"

[[4]]
[1] "b" "c" "c"

[[5]]
[1] "b" "c" "c"

这次RStudio崩溃 R会话中止。但是有时候> markovchainSequenceParallelRcpp(mclist, 15)会崩溃，有时它甚至不会崩溃n = 5我没有得到什么。任何想法。

请查看我今天观察到的内容。

n = 20.

仍然无法解决问题。请建议任何方法来克服这个问题。我正在研究它的最后一周。现在似乎浪费时间了。请帮忙。

R在运行并行程序时遇到致命错误

0 个答案: