GPU programming with Thrust

Thrust (http://code.google.com/p/thrust/) is a fantastic CUDA C++ interface matching within the C++ STL paradigm (templates, containers, iterators, generic algorithms,…).

Just an example from Thrust web site:

#include <thrust/device_vector.h>
#include <thrust/transform.h>
#include <thrust/sequence.h>
#include <thrust/copy.h>
#include <thrust/fill.h>
#include <thrust/replace.h>
#include <thrust/functional.h>
#include <iostream>
 
int main(void)
{
    // allocate three device_vectors with 10 elements
    thrust::device_vector<int> X(10);
    thrust::device_vector<int> Y(10);
    thrust::device_vector<int> Z(10);
 
   //STL like vectors but allocated on the GPU.All commands in the thrust namespace are excuted on GPU.
 
    // initialize X to 0,1,2,3, ....
    thrust::sequence(X.begin(), X.end());
 
    // compute Y = -X
    thrust::transform(X.begin(), X.end(), Y.begin(), thrust::negate<int>());
 
    // fill Z with twos
    thrust::fill(Z.begin(), Z.end(), 2);
 
    // compute Y = X mod 2
    thrust::transform(X.begin(), X.end(), Z.begin(), Y.begin(), thrust::modulus<int>());
 
    // replace all the ones in Y with tens
    thrust::replace(Y.begin(), Y.end(), 1, 10);
 
    // print Y
    thrust::copy(Y.begin(), Y.end(), std::ostream_iterator<int>(std::cout, "\n"));
 
    return 0;    
}

To compile this vector_example.cu file on Ubuntu 64bits (see Cuda on Ubuntu 64bits), you just have to use the NVidia nvcc compiler:

nvcc vector_example.cu -o vector_example --compiler-bindir=/usr/bin/g++-4.3

A couple of related links:

• An introduction to Thrust (PDF)
• Quick Start Guide