Simple CUDA Thrust Program Error

Question

I just write an simple CUDA Thrust program, but when I run it. I got this error: thrust::system::system_error at position 0x0037f99c .

Can someone help me to figure out why this happen?

#include<thrust\host_vector.h>
#include<thrust\device_vector.h>
#include<iostream>
using namespace std;
using namespace thrust;
int main()
{

    thrust::host_vector<int> h_vec(3);
    h_vec[0]=1;h_vec[1]=2;h_vec[2]=3;
    thrust::device_vector<int> d_vec(3) ;
    d_vec= h_vec;
    int h_sum = thrust::reduce(h_vec.begin(), h_vec.end());
    int d_sum = thrust::reduce(d_vec.begin(), d_vec.end());
return 0;
}

Probably you're not generating the right kind of code for your gpu's architecture. Catch the system_error and print out its .what() message to find out what the error is. — Jared Hoberock
I was able to successfully compile and run your code. Please check the suggestion given by JaredHoberock in above. — Sagar Masuti
I use CUDA Wizard to generate the project. I think the generation of code have no problem. The system error is generate by "int d_sum = thrust::reduce(d_vec.begin(), d_vec.end());" The error information is:synchronize: launch_closure_by_values:unspecified launch failure. BTW, while compile the program, I receive a lot warning like: Cannot tell what pointer points to, assuming global memory space. I use memory-check to check the program. It crashed. My GPU is Quadro K2000,CPU is Xeon E5. What's wrong? When I use CUSP library. It also crash when generating sparse matrix. --@Jared Hoberock — user3072616
Oh the problem solved. As you suggested, I change the kind of device code generation mode to "compute_20,sm_20". It works. Thank you very much @Jared Hoberock — user3072616
Does someone want to add this as an answer so we can getbthis off the unanswered list? — talonmies

Robert Crovella Robert Crovella · Accepted Answer · 2013-12-11T00:20:17

A few suggestions with Thrust:

If you are compiling your code with -G and having trouble, try compiling without -G
You can catch the errors that thrust throws, to get more information.
It's always recommended to compile your code for the architecture of the GPU you are using. So if you are on a cc2.0 GPU, compile with -arch=sm_20. If you are on a cc3.0 GPU, compile with -arch=sm_30 etc.
Finally, it's recommended to build a 64-bit project. On windows you would select a release/x64 project.

Simple CUDA Thrust Program Error

1 Answers