OpenCL - Empty profile timeline

Hi everyone,

I’m now working on an OpenCL project and have to optimize my program to get the best rate of speed.
So I got to use NSight software (under eclipse edition) to profile my project.

The problem is that no timeline is generated when I click on the “Profile” button.
But with a CUDA sample, it works great.
Have I missed something in the NSight settings ?

It would be usefull if someone has an idea to help me up.

The configuration I have :
Ubuntu 14
Nvidia GTX 960 (OpenCL 1.2)
NSight 7.5

An OpenCL Sample I try to profile but still empty :

#include <iostream>
#include <CL/cl.hpp>
using namespace std;

int main() {
    //get all platforms (drivers)
    std::vector<cl::Platform> all_platforms;
        std::cout<<" No platforms found. Check OpenCL installation!\n";
    cl::Platform default_platform=all_platforms[0];
    std::cout << "Using platform: "<<default_platform.getInfo<CL_PLATFORM_NAME>()<<"\n";

    //get default device of the default platform
    std::vector<cl::Device> all_devices;
    default_platform.getDevices(CL_DEVICE_TYPE_ALL, &all_devices);
        std::cout<<" No devices found. Check OpenCL installation!\n";
    cl::Device default_device=all_devices[0];
    std::cout<< "Using device: "<<default_device.getInfo<CL_DEVICE_NAME>()<<"\n";

    cl::Context context(CL_DEVICE_TYPE_GPU);

    cl::Program::Sources sources;

    // kernel calculates for each element C=A+B
    std::string kernel_code=
            "   void kernel simple_add(global const int* A, global const int* B, global int* C){       "
            "       C[get_global_id(0)]=A[get_global_id(0)]+B[get_global_id(0)];                 "
            "   }                                                                               ";

    cl::Program program(context, kernel_code, true);

    // create buffers on the device
    cl::Buffer buffer_A(context,CL_MEM_READ_WRITE,sizeof(int)*10);
    cl::Buffer buffer_B(context,CL_MEM_READ_WRITE,sizeof(int)*10);
    cl::Buffer buffer_C(context,CL_MEM_READ_WRITE,sizeof(int)*10);

    int A[] = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9};
    int B[] = {0, 1, 2, 0, 1, 2, 0, 1, 2, 0};

    //create queue to which we will push commands for the device.
    cl::CommandQueue queue(context,default_device, CL_QUEUE_PROFILING_ENABLE);

    //write arrays A and B to the device

    //run the kernel
     auto vadd = cl::make_kernel<cl::Buffer, cl::Buffer, cl::Buffer>(program, "simple_add");
     vadd(cl::EnqueueArgs(queue, cl::NDRange(10)), buffer_A, buffer_B, buffer_C);

    int C[10];
    //read result C from the device to array C


    std::cout<<" result: \n";
    for(int i=0;i<10;i++){
        std::cout<<C[i]<<" ";

    return 0;

Hi VladimirS,

Thank you for your asking, I just tested your OCL code and got no time line as well.

Nishgt profile uses nvprof to profile the GPU kernel and generate the time line, as far as I know nvprof doesn’t have support for opencl app

Best Regards

Thank you harryz_ for your answer,

I have tested this OCL code with NSight integrated in the Visual Studio solution. And it worked great !
So I’m wondering why the visual studio solution works instead of eclipse for an opencl application profiling.

Hi VladimirS,

You’re right, Performance Analysis in Nsight visual studio uses Nsight monitor and it can analysis OCL app, but the nvprof on windows also cannot work with OCL app, Nsight VSE works far different from the EE version.

By the way, NVIDIA Visual Profiler also cannot profile OCL app as it uses nvprof no matter on Windows or Linux.