I'm currently working on a originally GPU code that had MPI. I've removed MPI and selected an FPGA as the device. Each time I run the code with the same arguments , I end up with different results. Any solutions ?
I've attached a sample below. (from the out put - first 7 are arguments , 2 are time related outputs and last 4 are results)
you are given just you issue details but with that information not sufficient.
please tell me what what type of work are you doing, what arguments are you passing and why did you changed MPI to FPGA?
what type of output did you expecting?
please attaché screen shots
Thank you for taking the time to reply. I'm very new to FPGA programming and trying out different things to get things working.
I am currently working on a Computational Fluid Dynamics Algorithm having two kernels - one for initialization and the other for computation.
Originally the code was written for GPU and had MPI. To get it working on a single FPGA I had to remove the mpi code and select FPGA as the device. For parallel computation ND range was already used. So that was left unaltered. The algorithm deals with points, blocks and time steps. The first 6 parameters decide number of points and the 7th parameter is the amount of time steps. As you can see from the output, different results (the last 4 values of the output) are generated for the same input arguments (time parameter 0 and 10 ).
Compilation was done for Stratix10 as USM is being used.
Also, could you let me know the max parallel processes for ND range as well.
As we do not receive any response from you on the previous question/reply/answer that we have provided. Please login to ‘https://supporttickets.intel.com’, view details of the desire request, and post a feed/response within the next 15 days to allow me to continue to support you. After 15 days, this thread will be transitioned to community support. The community users will be able to help you on your follow-up questions.