Dataflow Function OpenCL (OpenCL Kernel)

This is simple example of vector addition to demonstrate Dataflow functionality in OpenCL Kernel. OpenCL Dataflow allows user to run multiple functions together to achieve higher throughput.

KEY CONCEPTS: Function/Task Level Parallelism

KEYWORDS: xcl_dataflow, xclDataflowFifoDepth

This example demonstrates the use of xcl_dataflow attribute used to implement task level parallelism in OpenCL kernels i.e.  multiple functions can be pipelied to increase the throughput of the design.

Kernel adder uses 3 functions read_input to read inputs from global memory, compute_add for addition of these inputs and write_output to write the results back to the global memory.

__attribute__ ((xcl_dataflow))
void adder(__global int *in, __global int *out, int inc, int size)
{
    int buffer_in[BUFFER_SIZE];
    int buffer_out[BUFFER_SIZE];

    read_input(in,buffer_in,size);
    compute_add(buffer_in,buffer_out,inc,size);
    write_result(out,buffer_out,size);
}

EXCLUDED PLATFORMS:

• All NoDMA Platforms, i.e u50 nodma etc

DESIGN FILES

Application code is located in the src directory. Accelerator binary files will be compiled to the xclbin directory. The xclbin directory is required by the Makefile and its contents will be filled during compilation. A listing of all the files in this example is shown below

src/adder.cl
src/host.cpp

Access these files in the github repo by clicking here.

COMMAND LINE ARGUMENTS

Once the environment has been configured, the application can be executed by

./cl_dataflow_func <adder XCLBIN>