How do I put multiple parameters as arguments to the kernel function.

[uniquerussa]

test.cu

__global__ void helloWorld(char *data) {

change
test.cu

__global__ void helloWorld(char *data, char *data2) {

cuda.js

module.exports.launch = function () {
  var func = arguments[0];
  var gridDim = arguments[1];
  var blockDim = arguments[2];
  var args = arguments[3];
  var args2= arguments[4];

  args = module.exports.prepareArguments(args);
  args2 = module.exports.prepareArguments(args2);

  return func.launchKernel(gridDim, blockDim, args, args2 );
}

function.cpp

Local<Object> buf = args[2]->ToObject();
  char *pbuffer = Buffer::Data(buf);
  size_t bufferSize = Buffer::Length(buf);

Local<Object> buf 2= args[3]->ToObject();
  char *pbuffer2 = Buffer::Data(buf2);
  size_t bufferSize2 = Buffer::Length(buf2);


void *cuKernel_Parameters= { &pbuffer , & pbuffer2}
 /*
 void *cuExtra[] = {
    CU_LAUNCH_PARAM_BUFFER_POINTER, pbuffer,
    CU_LAUNCH_PARAM_BUFFER_SIZE,    &bufferSize,
    CU_LAUNCH_PARAM_END
  };*/

  CUresult error = cuLaunchKernel(pfunction->m_function,
      gridDimX, gridDimY, gridDimZ,
      blockDimX, blockDimY, blockDimZ,
      0, 0, cuKernel_Parameters, NULL);

If i have access to an array data2 in kernel function, This code returns a 700 error.

[kashif]

thanks for the issue @teriusbin I will have a look...

[uniquerussa]

Multi- parameter problem is solved. Another question is how can I bind runtime api?