Would be appreciate if somebody can help. I have a small kernel that always finished with CL_INVALID_COMMAND_QUEUE error. I’ve tried on different hardware gtx 765m or gtx 980, the result is the same.
Here is the code host + kernel:
//get all platforms (drivers)
std::vector<cl::Platform> all_platforms;
cl::Platform::get(&all_platforms);
if(all_platforms.size()==0){
std::cout<<" No platforms found. Check OpenCL installation!
";
exit(1);
}
cl::Platform default_platform=all_platforms[1];
std::cout << "Using platform: “<<default_platform.getInfo<CL_PLATFORM_NAME>()<<”
";
//get default device of the default platform
std::vector<cl::Device> all_devices;
default_platform.getDevices(CL_DEVICE_TYPE_ALL, &all_devices);
if(all_devices.size()==0){
std::cout<<" No devices found. Check OpenCL installation!
";
exit(1);
}
cl::Device default_device=all_devices[0];
std::cout<< "Using device: “<<default_device.getInfo<CL_DEVICE_NAME>()<<”
";
cl::Context context({default_device});
cl::Program::Sources sources;
std::string kernel_code=
"__kernel void test(__global float* A,__global float* R) {"
"int i = get_global_id(0);"
"if(i>=1075021) return;"
"if(i<60000) {"
"R[i]=0;"
"return;"
"};"
"float vm=0.f;"
"for(int j=i-60000;j<=i;++j)"
"vm+=A[j];"
"R[i]=vm;"
"};";
sources.push_back({kernel_code.c_str(),kernel_code.length()});
cl::Program program(context,sources);
if(program.build({default_device})!=CL_SUCCESS){
std::cout<<" Error building: "<<program.getBuildInfo<CL_PROGRAM_BUILD_LOG>(default_device)<<"
";
exit(1);
}
size_t n=1075021;
// create buffers on the device
cl::Buffer buffer_A(context,CL_MEM_READ_WRITE,sizeof(float)*n);
cl::Buffer buffer_R(context,CL_MEM_READ_WRITE,sizeof(float)*n);
float *A = new float[n];
float *R = new float[n];
srand (time(NULL));
for(size_t i=0;i<n;++i)
A[i]=rand()%10;
cl::CommandQueue queue(context,default_device);
cl_int ret;
ret=queue.enqueueWriteBuffer(buffer_A,CL_TRUE,0,sizeof(float)*n,A);
ret=queue.finish();
cl::Kernel kernel_test=cl::Kernel(program,"test");
kernel_test.setArg(0,buffer_A);
kernel_test.setArg(1,buffer_R);
size_t max_work_size=1024;
size_t num_work_groups = (n-1) / max_work_size + 1;
size_t global_size_padded = num_work_groups * max_work_size;
queue.enqueueNDRangeKernel(kernel_test,cl::NullRange,cl::NDRange(global_size_padded),cl::NDRange(max_work_size));
ret=queue.finish();
ret=queue.enqueueReadBuffer(buffer_R,CL_TRUE,0,sizeof(float)*n,R);