Loading multiple modules in JCuda does not work

Question

In jCuda, you can load cuda files in PTX or CUBIN format and call (run) __global__functions (kernels) from Java.

With this in mind, I want to develop a framework using JCuda that gets a user-defined function __device__in a file .cuat runtime, loads and runs it. And I have already implemented a function __global__in which each thread detects the starting point of its associated data, performs some calculations, initialization, and then calls the user function __device__.

Here is my kernel pseudo code:

extern "C" __device__ void userFunc(args);
extern "C" __global__ void kernel(){

    // initialize

    userFunc(args);

    // rest of the kernel
}

And user function __device__:

extern "C" __device__ void userFunc(args){
    // do something
}

Java, , ( ptx, cuda : nvcc -m64 -ptx path/to/cudaFile -o cudaFile.ptx)

CUmodule kernelModule = new CUmodule(); // 1 
CUmodule userFuncModule = new CUmodule(); // 2
cuModuleLoad(kernelModule, ptxKernelFileName); // 3 
cuModuleLoad(userFuncModule, ptxUserFuncFileName); // 4

, 3: CUDA_ERROR_NO_BINARY_FOR_GPU. , ptx . :

ptxas -arch=sm_30 kernel.ptx

:

ptxas fatal : Unresolved extern function 'userFunc'

3 4, userFunc , . . , JCuda? ?

Edit:

Answer 1

: , API .

, , JIT. , ( ) JCUDA, , API-. ...

​​ , :

// test_function.cu
#include <math.h>
__device__ float mathop(float &x, float &y, float &z)
{
        float res = sin(x) + cos(y) + sqrt(z);
        return res;
}

// test_kernel.cu
extern __device__ float mathop(float & x, float & y, float & z);

__global__ void kernel(float *xvals, float * yvals, float * zvals, float *res)
{

        int tid = threadIdx.x + blockIdx.x * blockDim.x;

        res[tid] = mathop(xvals[tid], yvals[tid], zvals[tid]);
}

PTX, :

$ nvcc -arch=sm_30 -ptx test_function.cu
$ nvcc -arch=sm_30 -ptx test_kernel.cu
$ head -14 test_kernel.ptx
//
// Generated by NVIDIA NVVM Compiler
//
// Compiler Build ID: CL-19324607
// Cuda compilation tools, release 7.0, V7.0.27
// Based on LLVM 3.4svn
//

.version 4.2
.target sm_30
.address_size 64

        // .globl       _Z6kernelPfS_S_S_
.extern .func  (.param .b32 func_retval0) _Z6mathopRfS_S_

JIT, PTX , . , , . API- :

#include <cstdio>
#include <cuda.h>

#define drvErrChk(ans) { drvAssert(ans, __FILE__, __LINE__); }

inline void drvAssert(CUresult code, const char *file, int line, bool abort=true)
{
    if (code != CUDA_SUCCESS) {
        fprintf(stderr, "Driver API Error %04d at %s %d\n", int(code), file, line);
        exit(-1);
    }
}

int main()
{
    cuInit(0);

    CUdevice device;
    drvErrChk( cuDeviceGet(&device, 0) );

    CUcontext context;
    drvErrChk( cuCtxCreate(&context, 0, device) );

    CUlinkState state;
    drvErrChk( cuLinkCreate(0, 0, 0, &state) );
    drvErrChk( cuLinkAddFile(state, CU_JIT_INPUT_PTX, "test_function.ptx", 0, 0, 0) );
    drvErrChk( cuLinkAddFile(state, CU_JIT_INPUT_PTX, "test_kernel.ptx" , 0, 0, 0) );

    size_t sz;
    char * image;
    drvErrChk( cuLinkComplete(state, (void **)&image, &sz) );

    CUmodule module;
    drvErrChk( cuModuleLoadData(&module, image) );

    drvErrChk( cuLinkDestroy(state) );

    CUfunction function;
    drvErrChk( cuModuleGetFunction(&function, module, "_Z6kernelPfS_S_S_") );

    return 0;
}

. JCUDA, JIT.