MAGMA Forum

We are going to release MAGMA 0.2 today or tomorrow and we will fix the problem of linking with "C" Projects. Thanks for pointing it out. Until we do that, the easiest fix for you would be if you rename the file from where you call magma_dpotrf from .c to .cpp and wrap the C function declarations th...

Hi, Yes, we knew Intel has removed some "dummy" libraries in MKL 10.2 http://software.intel.com/en-us/articles/dummy-libraries-have-been-removed/ They provide a tool on how to get the correct linking sequence http://software.intel.com/en-us/articles/intel-mkl-link-line-advisor/ We will also try the ...

Just for the record of this topic, recompiling the MAGMA CUDA kernels for this specific configuration fixed the problem.

Hi,
It looks like we have to recompile a few CUDA kernels for your system. To make sure that's the problem, are the other functions O.K., e.g. what do you get when running testing_sgeqrf?
Thanks,
Stan

Hello, A user had a similar problem before and in that case updating the driver fixed it. You can run an older cuda on a new driver ( for example CUDA 2.1 on 190 driver) but not vice-versa. For example CUDA 2.3 requires 190.xx CUDA 2.2 requires 185.xx CUDA 2.1 requires 180.xx You can check your driv...

Hi, The undefined functions are in the LAPACK library and should be recognized (in general) in the MKL library when you link with "-L/opt/intel/mkl721/lib/em64t -lmkl_em64t -lguide" What version of gcc and mkl are you using? We use mkl 10.0.1.014 and compile exactly in the way you are trying. Would ...

Hello,
We are preparing MAGMA version 0.2 to be released by November 14, along with the sources. We don't plan to directly support Apple OS X (but know of collaborators that have ported the current MAGMA version 0.1 and we will check with them if they want to do it).
Stan

Hi Michael, I am happy to hear you like the library. We are preparing MAGMA version 0.2 to be released by November 14, along with the sources. The library itself is "high" level, in the sense we try to avoid directly coding in CUDA, just using CUDA BLAS. There are particular BLAS cases (e.g. gemm on...

Malcolm, I don't see any technical reasons. As far as I know they are working on it and would have it soon. My guess is that the reason is combination of man-power needed to do it and priorities. There are many routines to do; others to optimize; maintain them for different platforms; the ones to be...

The complex version of the 3 one-sided factorizations will be included. We still don't have some BLAS in complex so if NVIDIA does not provide it until then we are going to provide wrappers for what we need. For example, to do a cherk on the GPU we will just copy the data needed for the operations o...