ret = cl.clSetKernelArg(kernel, 3, SIZEOF_INT, wrap(ELEMENT_COUNT)); checkError("on clSetKernelArg3", ret);
out.println("used device memory: "+ (srcA.capacity()+srcB.capacity()+dest.capacity())/1000000 +"MB");
// Asynchronous write of data to GPU device
ret = cl.clEnqueueWriteBuffer(commandQueue, devSrcA, CL.CL_FALSE, 0, srcA.capacity(), srcA, 0, null, null);
checkError("on clEnqueueWriteBuffer", ret);
ret = cl.clEnqueueWriteBuffer(commandQueue, devSrcB, CL.CL_FALSE, 0, srcB.capacity(), srcB, 0, null, null);
checkError("on clEnqueueWriteBuffer", ret);
// Launch kernel