Can't get boost's sample main.cpp to compile

[Joshua Laferriere]

I posted the issue on stackexchange: http://stackoverflow.com/questions/38045020/boost-compute-opencl-wrapper-initial-setup-problems-qt-g

copied here for reference

Been trying to compile this sample code: https://github.com/boostorg/compute/blob/master/README.md

I installed QT Creator 5.7 using mingw530

I compiled the boost libraries using

bootstrap.bat gcc
b2 install --prefix="C:\Boostbuild" --toolset=gcc
bjam --build-dir=c:/Dev/Boost/Boost_lib toolset=gcc stage

I installed AMD SDK 3.0, 2.9.1, and 2.9

I even downloaded opencl 1.1, 1.2, and 2.1 cl.hpp and tried to include that.

The compile starts, but I get a slew of errors

C:\Dev\Boost\compute-master\include\boost\compute\device.hpp:80: error: undefined reference to `clRetainDevice@4'

C:\Users\User\Documents\Projects\build-console-test-Desktop_Qt_5_7_0_MinGW_32bit-Debug\debug\main.o:-1: In function `ZN5boost7compute6deviceaSERKS1_':

I tried a simple qt console app, using the code supplied by boost compute

Note: this isn't specific to qt, I've also tried compiling this using

g++ -I/path/to/compute/include sort.cpp -lOpenCL

doing an -I to each of the include's in the main.cpp (see below)

Ideally, I'd like to know how to compile the example given on their page, with includes and all (and relevant amd sdk and/or opencl versions) along with the necessary included libraries.

My qt project file libraries

INCLUDEPATH += C:\Dev\Boost\compute-master\include
INCLUDEPATH += C:/Users/User/Downloads/dev/boost_1_61_0
INCLUDEPATH += "C:\Program Files (x86)\AMD APP SDK\2.9-1\include"

My main.cpp

#include <iostream>
#include <vector>
#include <algorithm>
#include <boost/compute.hpp>
//#define CL_USE_DEPRECATED_OPENCL_1_1_APIS
//#undef CL_VERSION_1_2
//#include <C:\Dev\OpenCL\2.1\cl.hpp>

namespace compute = boost::compute;

int main()
{

// get the default compute device
compute::device gpu = compute::system::default_device();

// create a compute context and command queue
compute::context ctx(gpu);
compute::command_queue queue(ctx, gpu);

// generate random numbers on the host
std::vector<float> host_vector(1000000);
std::generate(host_vector.begin(), host_vector.end(), rand);

// create vector on the device
compute::vector<float> device_vector(1000000, ctx);

// copy data to the device
compute::copy(
host_vector.begin(), host_vector.end(), device_vector.begin(), queue
);

// sort data on the device
compute::sort(
device_vector.begin(), device_vector.end(), queue
);

// copy data back to the host
compute::copy(
device_vector.begin(), device_vector.end(), host_vector.begin(), queue
);

return 0;
}

if I uncomment out the include cl.hpp, I get further

C:/Dev/Boost/compute-master/include/boost/compute/allocator/buffer_allocator.hpp:91: undefined reference to `clReleaseMemObject@4'

[Kyle Lutz]

Hi Joshua,

It looks like the OpenCL library isn't being properly linked in (which
leads to the "undefined reference to 'cl...'" errors). Also, make sure
that the OpenCL headers you're compiling against match the OpenCL
library version you're linking against (e.g. make sure to use the v1.2
headers with a v1.2 OpenCL library).

To help debug more, can you try this more reduced test case which just
uses OpenCL, no Boost.Compute:

#include <stdio.h>
#include <CL/cl.h>

int main()
{
cl_uint num_platforms = 0;
clGetPlatformIDs(0, 0, &num_platforms);
printf("found %u platforms\n", num_platforms);
return 0;
}

Let me know if that works for you.

-kyle

[Joshua Laferriere]

Also, make sure that the OpenCL headers you're compiling against match the OpenCL library version you're linking against (e.g. make sure to use the v1.2 headers with a v1.2 OpenCL library).

how do I verify that?  That is something I've been trying to figure out.  I thought that maybe AMD SDK 3.0 was backwards compatible (in the examples I used, I just happened to keep the DSK at 2.9.1)

[Joshua Laferriere]

To help debug more, can you try this more reduced test case which just
uses OpenCL, no Boost.Compute:

#include <stdio.h>
#include <CL/cl.h>

int main()
{
    cl_uint num_platforms = 0;
    clGetPlatformIDs(0, 0, &num_platforms);
    printf("found %u platforms\n", num_platforms);
    return 0;
}

Let me know if that works for you.

-kyle

project file:
QT += core

QT -= gui

CONFIG += c++11

TARGET = console-test

CONFIG += console

CONFIG -= app_bundle

OPENCL_ROOT = "C:\Program Files (x86)\AMD APP SDK\2.9-1"

isEmpty(OPENCL_ROOT) {

  error("Please set AMDAPPSDKROOT to the location of the AMD APP SDK")

} else {

  message(Using Boost from: $$OPENCL_ROOT)

}

INCLUDEPATH += $$OPENCL_ROOT/include

LIBS += -L$${OPENCL_ROOT}/lib/x86

LIBS += -LOpenCL

INCLUDEPATH += C:\Dev\Boost\compute-master\include

INCLUDEPATH += C:/Users/User/Downloads/dev/boost_1_61_0

INCLUDEPATH += "C:\Program Files (x86)\AMD APP SDK\2.9-1\include"

TEMPLATE = app

SOURCES += main.cpp

C:\Users\User\Documents\Projects\build-console-test-Desktop_Qt_5_7_0_MinGW_32bit-Debug/../console-test/main.cpp:7: undefined reference to `clGetPlatformIDs@12'

collect2.exe: error: ld returned 1 exit status

 

[Joshua Laferriere]

nevermind, figured that cl.hpp is included in the amd sdk, so no need for me to include it.

[Joshua Laferriere]

fixed it:

QT += core
QT -= gui


CONFIG += c++11


TARGET = console-test
CONFIG += console
CONFIG -= app_bundle


OPENCL_ROOT = "C:\Program Files (x86)\AMD APP SDK\2.9-1"
isEmpty(OPENCL_ROOT) {
  error("Please set AMDAPPSDKROOT to the location of the AMD APP SDK")
} else {
  message(Using Boost from: $$OPENCL_ROOT)
}


INCLUDEPATH += $$OPENCL_ROOT/include
LIBS += -L$${OPENCL_ROOT}/lib/x86

LIBS += -lOpenCL


#INCLUDEPATH += C:\Dev\Boost\compute-master\include
#INCLUDEPATH += C:/Users/User/Downloads/dev/boost_1_61_0
#INCLUDEPATH += "C:\Program Files (x86)\AMD APP SDK\2.9-1\include"