Now that OpenMP 4 is fairly well supported on the Fortran side (except for simd reduction), I've been able to set up an example which can run in host, MIC native, and offload mode, simply by changing compile options.
We are planning to get the Xeon Phi working in our research environment which is a non-linux / non-windows operating system. Therefore we need to write our own drivers and our driver model is quite different to the one used in Linux / Windows. We have looked at the available documents for the Xeon Phi, however from a systems point of view they are pretty disappointing.
Is there something like a technical reference manual / register reference of the Xeon Phi containing all the available registers on host side / coprocessor side ?
We have some offload code that doesn't run because dlopen() fails with undefined symbols. One of the suggestions that I found was to add "-Wl,-zdefs" to the offload linker options, which might bring the undefined symbols to our attention at link time. However, when I do this, even the most simple offload code won't link.
I am trying to figure out the effect of using array alignment in vectorization of MIC code.
Here is a simple piece of offload code from Intel
I trying a run a simple file-open program on mic0 but the program fails (does not open file but terminates gracefully!) when i run it from the host machine and succeeds (program opens the file on mic0))when i run it directly from the mic card
I have cleared the firewalls, and i do
I_MPI_MIC = enable
% scp /opt/intel/impi/4.1.1.036/mic/bin/pmi_proxy mic0:/bin
% scp /opt/intel/impi/4.1.1.036/mic/lib/* mic0:/lib64/.
here is my run command:-
/opt/intel/impi/4.1.1.036/bin64/mpirun -host mic0 -n 1 ~/a.out
I have read in the FAQS that MKL has a random number generator included in the vector libraries.
Are there intrinsics for generating vector random numbers? or a small kernels that I could implement. I don't want to use MKL.
could someone help me with this issue? I try to compile some OpenMP code, but still wrote this for offload and native compilation:
I am running a MPI program which involves 5 ranks communicating with each other. The program runs fine when i run it on Xeon platform but when i put one of the process rank on MIC the program crashes giving me a segmentation fault
mpi version: intel Mpi, 4.1.1.036
This is how i am setting up the libraries and compiler settings for mic
scp /opt/intel/impi/4.1.1.036/mic/bin/pmi_proxy mic0:/bin
scp /opt/intel/impi/4.1.1.036/mic/lib/* mic0:/lib64/.