Currently, we are focusing on the calculation of Big Data distribution. Considering the adventage of FPGA card, the efficiency of parrallel compution and energy conservation, we are targeting at applying FPGA card in spark under the cluster computiong. However, opencl is temporarily short for the JAVA interface, or the function supporting JAVA or Scala, which results in the failure of data trasmission betwee the host and the device, and the failure of invoking the FPGA. We are looking forward to having your technological supports, developing FPGA’s strong computing capability. Thanks.
I recommend addressing this issue with our embedded design team. Please post this inquiry at https://embedded.communities.intel.com/community/en