openmp

问题 I have a function which has an unordered set as a parameter . Since I am using openmp I am converting this unordered set to vector . I use a std::copy for this conversion . //pseudo code func( std::unorderedset s1) begin vector v1; std::copy(s1.begin,s2.end,std::back_inserter(v1.end()); #openmp scope for( i = 0 ; i < v1.size(); i++ ) { //accessing v1(i) } end However I feel std::copy is a costly operation . So what I think is, if I create a class variable vector and I keep populating this

问题 The following is defined in the makefile: CXX = g++ CXXFLAGS = -std=c++11 I would like to compile my code with OpenMP directives without changing the original makefile (which runs perfectly). The manual suggests a way to do this, is by changing it on the command line. However, when I run, make CXXFLAGS=-std=c++11 -fopenmp the error mentioned before pops up. Can someone help me understand what I'm doing wrong? 回答1: The problem here is the space between -std=c++11 and -fopenmp . It splits these

问题 As I know, false sharing occurs when two cpu cores access different parts of a single memory block. In this case, the L1 caches in each core contains different values. What is the effect of false sharing? Do commodity CPUs always detect a false sharing? Can it lead to something like a race condition (persisting different cache versions in the memory)? 回答1: False sharing occurs when one core modifies some data, and another reads some unrelated data that happens to lie in the same cache line.

问题 I am trying to get some code paralleled, and algorithm must be remained. Here is the original Code: #include <stdio.h> #define N 50 main() { int prime[N] ; int j ; int k ; int n ; int quo,rem ; P1: prime[0] = 2 ; n = 3 ; j = 0 ; P2: j = j+1 ; prime[j] = n ; P3: if (j == (N-1)) goto P9 ; P4: n = n + 2 ; P5: k = 1 ; P6: quo = n / prime[k] ; rem = n % prime[k] ; if (rem == 0) goto P4 ; P7: if (quo <= prime[k]) goto P2 ; P8: k = k+1 ; goto P6 ; P9: for(j=0 ; j < N ; j++) printf("%d ",prime[j]) ;

问题 I was wondering how OpenMP directives are handled by compiler, such as gcc? For example, in this code int main(int argc, char *argv[]) { #pragma omp parallel printf("Hello, world.\n"); return 0; } Does preprocessor of gcc modify the C code by replacing the OpenMP directive with some other code? What is the code like after preprocessing and right before being assembled? Thanks and regards! 回答1: You can do a web search and find papers discussing this topic. I hate to give links because they

问题 I have a do loop which updates T value and calculates the maximum difference during the iteration, called dumax. I need to initialize dumax, so I set it to be firstprivate Then I will not be able to use: reduction(max: dumax) Reduction operator seems to accept private variable. Then how can I get the maximum value of dumax before I end the parallel? My program is shown below: DUMAX=0.0D0 !$OMP PARALLEL DEFAULT(PRIVATE), SHARED(T_R, T_B), FIRSTPRIVATE(DUMAX) !$OMP DO DO I=2, N-1, 2 DO J=2, N-1

问题 I have implemented the Jacobi algorithm based on the routine described in the book Numerical Recipes but since I plan to work with very large matrices I am trying to parallelize it using openmp. void ROTATE(MatrixXd &a, int i, int j, int k, int l, double s, double tau) { double g,h; g=a(i,j); h=a(k,l); a(i,j)=g-s*(h+g*tau); a(k,l)=h+s*(g-h*tau); } void jacobi(int n, MatrixXd &a, MatrixXd &v, VectorXd &d ) { int j,iq,ip,i; double tresh,theta,tau,t,sm,s,h,g,c; VectorXd b(n); VectorXd z(n); v

问题 #include <iostream> #include <iomanip> #include <fstream> #include <sstream> #include <string> #include <stdio.h> #include <stdlib.h> #include <math.h> #include <omp.h> using namespace std; void output(float a[], float X[], float Y[], int I, int J) { ofstream ft; int i; ft.open("flow.dat"); ft<<"variables=\"x\",\"y\",\"a\""<<"\n" <<"zone f=point"<<"\n" <<"I="<<I<<",J="<<J<<"\n" <<endl; for(int i=0;i<I*J;i++) { ft<<setiosflags(ios::scientific) <<X[i]<<" "<<Y[i]<<" "<<a[i]<<endl; } ft.close();

问题 I'm using PGI to compile the following program which uses OpenMP's target directives to offload work to a GPU: #include <iostream> #include <cmath> int main(){ const int SIZE = 400000; double *m; m = new double[SIZE]; #pragma omp target teams distribute parallel for for(int i=0;i<SIZE;i++) m[i] = std::sin((double)i); for(int i=0;i<SIZE;i++) std::cout<<m[i]<<"\n"; } My compilation string is as follows: pgc++ -omp -ta=tesla,pinned,cc60 -Minfo=accel -fast test2.cpp Compilation succeeds, but it

问题 I have mpf_t omp_mpf_add(mpf_t out, mpf_t in) { mpf_add(out, out, in); return out; } And i want to make this function as an openmp reduction # pragma omp declare reduction (mpf_add:mpf_t:omp_mpf_add(omp_out,omp_in)) \ initializer(omp_priv=omp_orig) But i get error: reduction type cannot be an array type error. what should i change so that the function will work and i can use it with #pragma omp parallel for reduction(mpf_add:x) 来源： https://stackoverflow.com/questions/47617628/openmp-with