Stack Overflow
Stack Overflow

Questions tagged [magma]

MAGMA (Matrix Algebra on GPU and Multicore Architectures) is a collection of linear algebra libraries for heterogeneous processor architectures. If you have questions about the MAGMA computational algebra system, please use the [magma-ca] tag.

Designed by a team at the Innovative Computing Laboratory at the University of Tennessee, MAGMA allows applications to exploity the power of heterogeneous systems of multi-core CPU's and multi-GPU's to deliver the most accurate solution in the fastest time within specified energy constraints.

42 questions
1
vote
1 answer

magmablas_dgemm not working for larger grid size

I am new to using cuda and the magma libraries. I'm trying out some functions on a test problem, a 2D heat equation. The code I wrote seemed to work perfectly for grid sizes of 32, 64, and 128. But it produced wrong results for grid sizes of 256 or…
user3527862
  • 45
  • 1
  • 5
1
vote
1 answer

Standard Fortran interface for cuBLAS

I am using a commercial simulation software on Linux that does intensive matrix manipulation. The software uses Intel MKL by default, but it allows me to replace it with a custom BLAS/LAPACK library. This library must be a shared object (.so)…
Bichoy
  • 351
  • 2
  • 9
0
votes
0 answers

Why does multiplying the user-defined opt.tolerance with the machine precision (eps) give the tolerance threshold (tol) in MAGMA?

for example, the user-defined opt.tolerance is 10^-6, the machine eps is 10^-14, according to the code: double tol = opts.tolerance * lapackf77_dlamch("E"); the tol is 10^-20, which is far less than the user need. Is this necessary? I thought once…
ingridli
  • 5
  • 2
0
votes
0 answers

Is there an algorithm for finding every isolated singular point on an algebraic variety, or a programming language that implements this?

(note: this question was closed on math stack exchange before it was answered, so I'm asking it here) Suppose one wishes to test if a given surface f(x,y,z,w) = 0 in projective 3 space has singular points (i.e. df/dx = df/dy = df/dz = 0), and one…
drfpslegend
  • 1
0
votes
0 answers

How to access the results of a Magma Routine

I am trying to access the results of the eigenvalue decomposition of a general real matrix, using the magma_sgeev routine. My code is as follows - #include #include #include #include #include…
Atharva Dubey
  • 832
  • 1
  • 8
  • 25
0
votes
1 answer

MAGMA in google colab

i always get the following error "Error in magma_getdevice_arch: MAGMA not initialized (call magma_init() first) or bad device" whenever i run a very simple code using google colab with magma-2.6.1 installed. //////////////////////// #include…
Hisham Elzayadi
  • 1
  • 1
0
votes
1 answer

How to use MAGMA with NVIDIA GPU card instead of CPU LAPACKE to inverse large matrix

I need to inverse large matrices and I would like to modify my current LAPACKE version routine in order to exploit the powerfull of a GPU NVIDIA Card. Indeed, my LAPACKE routines works well for relative small matrices but not for large…
user1773603
0
votes
1 answer

Why my inversions of matrices are such slow with LAPACKE in C++ : MAGMA Alternative and set up

I am using LAPACK to inverse a matrix: I did a reference passing, i.e by working on the address. Here below the function with an input matrix and an output matrix referenced by their address. The issue is that I am obliged to convert the F_matrix…
user1773603
0
votes
1 answer

CMakeLists linkage error 'magma_opts::parse_opts' using MAGMA testing code

Currently I am using MAGMA 2.5.4 to solve a batched linear solver with tiny sizes of matrices. I want to involve magma_dgesv_batched in my project compiled via a CMakeLists file. The include_directories and target_link_libraries are shown…
chchien
  • 23
  • 3
0
votes
1 answer

How to perform basic operations (+ - * /) on GPU and store the result on it

I have the following code line, gamma is a CPU variable, that after i will need to copy to GPU. gamma_x and delta are also stored on CPU. Is there any way that i can execute the following line and store its result directly on GPU? So basically, host…
eldev09
  • 39
  • 8
0
votes
1 answer

magma_dpotrf_gpu gives segmentation fault

I have a cblas and lapack code that I'm trying to "convert" to magma code to take advantage of the GPU. This is my CPU code that works fine: cblas_dsyrk(CblasColMajor,CblasUpper,CblasTrans,n,m,1.0,A2,m,…
thejoker
  • 31
  • 8
0
votes
1 answer

how to install MAGMA? I `make install` is run well, but `make test` cause error

I want to install MAGMA in a server(centos) in order to use the GEMM function. I have installed CUDA and MKL in this server before i install MAGMA, and i check both is run well by runing a cuBLAS example program and a MKL example program. Before i…
nomadlx
  • 155
  • 2
  • 10
0
votes
1 answer

make install saying CUDADIR environment variable not set, when it is

I'm currently trying to install MAGMA, however trying to run "sudo make install" gives me the error: $ sudo make install prefix=/usr/local/magma make.check-cuda:7: *** Set $CUDADIR, preferably in your environment, e.g., run "export…
as646
  • 195
  • 1
  • 9
0
votes
1 answer

Compiling with icc and MAGMAmic cannot find "complex" system file

I am trying to compile a source file with icc compiler and MAGMAmic library. However I get the following error: icc -c -o direct.o direct.c -O3 -openmp -DADD_ -Wall -DHAVE_MIC -I/opt/intel/mic/coi/include -I/usr/include/intel-coi…
Dimitar Slavchev
  • 192
  • 2
  • 11
0
votes
1 answer

matrix max a la matlab on gpu

I'm porting some code from Matlab to run on an Nvidia GPU. I can't figure out a way to do the following: B = max(A, 0) where A and B are matrices. In words, I need to replace negative values in a matrix with zeros. I know how to write a kernel…
user2297560
  • 2,953
  • 1
  • 14
  • 11
1
2
3