Search found 33 matches

by evanlezar
Tue Mar 08, 2011 2:51 pm
Forum: User discussion
Topic: Contributing code to MAGMA
Replies: 2
Views: 1849

Re: Contributing code to MAGMA

Thanks Mathieu

The changes I have made were made to rc3. I will move them to the rc4 source by the end of the week and send it through. Hopefully this will reduce your workload a little.
by evanlezar
Tue Mar 08, 2011 5:33 am
Forum: User discussion
Topic: Contributing code to MAGMA
Replies: 2
Views: 1849

Contributing code to MAGMA

Hi,

There are a few minor bugs/niggles that I came across in MAGMA 1.0 rc3, and from the rc4 source code these have not yet been addressed.

What is the best way to contribute patches to magma?

Thanks
by evanlezar
Fri Feb 25, 2011 9:45 am
Forum: User discussion
Topic: Increased GPU memory requirement with MAGMA 1.0
Replies: 4
Views: 4126

Re: Increased GPU memory requirement with MAGMA 1.0

Just a note. I have implemented a set of magma_Xgetrf_gpu_square routines, which work pretty much as the standard magma_Xgetrf_gpu routines, but don't have the requirement that the matrix sizes be a multiple of 32. What is required is that the leading dimension specified is a multiple of 32 and that...
by evanlezar
Thu Feb 24, 2011 4:51 am
Forum: User discussion
Topic: Increased GPU memory requirement with MAGMA 1.0
Replies: 4
Views: 4126

Re: Increased GPU memory requirement with MAGMA 1.0

Stan, Thanks for the reply. Sorry that it has taken me so long to respond. I understand fully that the in-place transpose is done for performance reasons. I just have a few quick questions. I have looked a the magma_cgetrf routine, and see that there is non %32 check there, so as long as enough devi...
by evanlezar
Mon Feb 07, 2011 10:22 am
Forum: User discussion
Topic: Increased GPU memory requirement with MAGMA 1.0
Replies: 4
Views: 4126

Increased GPU memory requirement with MAGMA 1.0

Hi, I have recently made the move from MAGMA 0.2 to 1.0 RC3, and have noticed that in certain cases, the GPU LU decomposition (device interface) such as is provided by magma_cgetrf_gpu(), the device memory used has increased by about a factor 2. Upon investigation of cgetrf_gpu.cpp I found the follo...
by evanlezar
Sun Oct 31, 2010 9:49 pm
Forum: User discussion
Topic: MAGMA 0.2 and CUDA 3.2 RC
Replies: 1
Views: 1890

Re: MAGMA 0.2 and CUDA 3.2 RC

Just a note.

Since the release of CUDA 3.2 RC2 it seems that MAGMA once again works. I was however not able to track down the problem.

Thanks.
by evanlezar
Wed Sep 29, 2010 7:18 am
Forum: User discussion
Topic: MAGMA 0.2 and CUDA 3.2 RC
Replies: 1
Views: 1890

MAGMA 0.2 and CUDA 3.2 RC

Hi

I recently upgraded to CUDA 3.2 RC (for the CUSPARSE functionality) and as a result MAGMA no longer works (I have been using the GETRF and GETRS functions specifically).

Is there someone else with a similar problem, or perhaps a timeline for the release of MAGMA 0.3?

Thanks

Evan
by evanlezar
Sun Nov 29, 2009 3:55 pm
Forum: User discussion
Topic: MAGMA Shared libraries
Replies: 4
Views: 17996

Re: MAGMA Shared libraries

Just an update. I found my error. I was not linking in lapack (DOH). What was throwing me off was the fact that when I included lapack the following error occured OSError: /home/elezar/src/cuda/magma_0.2/lib/libmagma.so: undefined symbol: magma_get_dgeqrf_nb Which I foolishly thought was due to lapa...
by evanlezar
Sat Nov 28, 2009 5:27 pm
Forum: User discussion
Topic: MAGMA Shared libraries
Replies: 4
Views: 17996

MAGMA Shared libraries

Hi I am trying to build certain MAGMA functions into a shared library. The idea behind this is that I want a number of backends (CPU blas, MAGMA, ... etc) that adhere to a common interface for a benchmarking application that I am developing. I have tried to do this using the static libraries supplie...
by evanlezar
Wed Nov 04, 2009 4:58 pm
Forum: User discussion
Topic: Complex types of the MAGMA routines
Replies: 10
Views: 34585

Re: Complex types of the MAGMA routines

It was not my intention to link spam. I have no affiliation with EM Photonics, and was just pointing it out to those readers that were not aware of it. It is my opinion, that although in its infancy Magma offers a much better solution - especially to academic developers such as myself - and I will c...