I am new to Intel MKL and I am sorry if my question seems out of line.
Is there any function or subroutine in MKL that computes the kronecker product of two matrix (like kron in matlab).
Also is there a function in MKL that returns the identity matrix of a specified dimension (like eye(n) in matlab).
I tried to compile in Visual studio 2013 for some program using mkl.
It works fine in 32bit environment. But in 64 bit they say
error LNK1181 :Can not open 'mkl_intel_s.lib'
error #11018: Cannot open mkl_intel_s_dll.lib
error $11018: Cannot open mkl_intel_s.lib
I already download all files from parallel studio
And when I copy and paste mkl_intel_s_dll.lib and mkl_intel_s.lib from C:\Program Files (x86)\Intel\Composer XE\mkl\lib\ia32
I'm using MKL to speed up numpy and doing some benchmark of matrix multiplication. Mostly I followed the instruction in https://software.intel.com/en-us/articles/numpyscipy-with-intel-mkl. It works great on my dev Mac machine but when I build it the same way on a ec2 c3.2xlarge instance, only one core is used with the benchmark script. I'm pretty sure it has 8 physical core...
Has anyone encounter this before or the processor on ec2 doesn't support MKL?
I'd like to accomplish the following on a cluster-computer with nproc processors:
1. define two blacs grids, the first of the form p x q = nprocs, the second of the form 1 x 1, which is also just the root process
2. use pdgemr2d to copy a distributed matrix from the first grid to the second
I am already stuck at point 1.
So here are my questions:
a) Do I have to initialize two MPI communicators and define the blacs contexts in these?
b) If not, how do I create two separate blacs contexts?
I have a problem with update my project from VS 2010 to VS 2015 widh mkl_rt.dll
I using a few function from mkl_rt.dll file version 10.3.4.1
VS 2010 - all Ok
VS 2013 - compiling and linking - are Ok. but in runtime the application was ended without crash or exeption on any function from mkl_rt.dll
I don't know what is wrong ?
May be I need new version of library ?
Recently I am using MKL FFT code to get the cycle count of DftiComputeForward. Form mkl documents, DFTI_NUMBER_OF_USER_THREADS is no longer used in latest MKL version. But I made a test.
Method is adding "status = DftiSetValue(FFT_desc, DFTI_NUMBER_OF_USER_THREADS, (1/2/3/4));" in my test code and result is:
FFT and thread setting
No setting thread
I am working on a 3D numerical integrator for a non-linear PDE using the parallel FFT library included in MKL.
My arrays consist of 2^30 data points which is much much larger than the cache. This results in ~50% of cache references being misses leading to a massive amount of execution time being purely accessing memory.
Is there a clever way I can deal with this? Is it expected to have 50% cache misses using an array this large?
Any help would be much appreciated.
Hi all, Please, are threaded the following MKL sparse BLAS functions? 'mkl_cspblas_?coosymv', 'mkl_?coosymv' and 'mkl_sparse_?_mv1 (11.3 beta) I'm not able to run with more than one thread these functions. Thanks in advance, Roberto Souto.