Stack Overflow
Stack Overflow

Questions tagged [cusolver]

a library that is part of the NVIDIA CUDA toolkit to accelerate various dense and sparse LAPACK-like operations on GPUs

cusolver is a library included with CUDA 7.0 and beyond, to provide GPU acceleration for 3 categories of linear algebra problems:

  1. cuSolverDN deals with dense matrix factorization and solve routines such as LU, QR, SVD and LDLT

  2. cuSolverSP provides a set of sparse routines based on a sparse QR factorization

  3. cuSolverRF is a sparse re-factorization package for solving a sequence of matrices where only the coefficients are changed but the sparsity pattern remains the same

cusolver documentation is here.

56 questions
10
votes
1 answer

Solving dense linear systems AX = B with CUDA

Can I use the new cuSOLVER library (CUDA 7) to solve linear systems of the form AX = B where A, X and B are NxN dense matrices ?
Shay S.
  • 101
  • 1
  • 1
  • 5
7
votes
1 answer

Best way of solving sparse linear systems in C++ - GPU Possible?

I am currently working on a project where we need to solve |Ax - b|^2. In this case, A is a very sparse matrix and A'A has at most 5 nonzero elements in each row. We are working with images and the dimension of A'A is NxN where N is the number of…
El_Loco
  • 1,716
  • 4
  • 20
  • 35
6
votes
3 answers

batch CUDA solution of sparse banded Ax=b for various b's

I have a sparse banded matrix A and I'd like to (direct) solve Ax=b. I have about 500 vectors b, so I'd like to solve for the corresponding 500 x's. I'm brand new to CUDA, so I'm a little confused as to what options I have available. cuSOLVER has a…
james proctor
  • 115
  • 5
4
votes
4 answers

Solving sparse definite positive linear systems in CUDA

We are experiencing problems while using cuSOLVER's cusolverSpScsrlsvchol function, probably due to misunderstanding of the cuSOLVER library. Motivation: we are solving the Poisson equation -divgrad x = b on a rectangular grid. In 2 dimensions with…
AdrianO
  • 175
  • 1
  • 11
3
votes
2 answers

Cholesky decomposition with CUDA

I am trying to implement Cholesky decomposition using the cuSOLVER library. I am a beginner CUDA programmer and I have always specified block-sizes and grid-sizes, but I am not able to find out how this can be set explicitly by the programmer with…
yath
  • 355
  • 1
  • 4
  • 14
3
votes
2 answers

Singular values calculation only with CUDA

I'm trying to use the new cusolverDnSgesvd routine of CUDA 7.0 for the calculation of the singular values. The full code is reported below: #include "cuda_runtime.h" #include "device_launch_parameters.h" #include…
WestWizard
  • 51
  • 7
2
votes
0 answers

multiple SVDs of a matrix using CUDA

I am trying to do SVD using the code given below, but I do not get correct results for d_S, when num_Matrices > 128, lets say 256, I get correct results when num_Matrices = 128*128, I am referring to a code based on an answer to a question asked on…
Kavita Kaithwas
  • 33
  • 4
2
votes
1 answer

Testing CUDA 11 cusolverDnDSgels()

Trying to make sense of cusolverDnDSgels function. If I run it with simple 3x3 example as in the docs it works, but when I run it with my data then d_info returns -1 which as the docs says if d_info = -i then i-th argument is not valid. Bellow I…
megaflow
  • 51
  • 3
2
votes
1 answer

Undefined reference to `cusolverDn` functions

I need to calculate eigenvalue of a big matrix in parallel. For that purpose I use cuSolver. To test, how it works, I took code from documentation: #include #include #include #include void…
anonymus321
  • 33
  • 4
2
votes
1 answer

What is the most efficient way to compute the inverse of a general matrix using cuSolver?

I have in mind to to use getrf and getrs from the cuSolver package and to solve AB=X with B=I. Is this the most best way to solve this problem? If so, what is the best way to create the col-major identity matrix B in device memory? It can be done…
avgn
  • 982
  • 6
  • 19
2
votes
1 answer

Solve linear equation of AX=B

I currently solve Ax=b equation two times. where A is sparse matrix NxN x, b are vectors of size N. (I have b1 and b2) I want to reduce times by solving both of them in one shot using cusparse functions. so what I though is to build from the 2 b's I…
Rotem.O
  • 93
  • 1
  • 11
2
votes
1 answer

Interfacing cuSOLVER-sparse using PyCUDA

I'm trying to interface the sparse cuSOLVER routine cusolverSpDcsrlsvqr() (>= CUDA 7.0) using PyCUDA and am facing some difficulties: I have tried wrapping the methods the same way the dense cuSolver routines are wrapped in scikits-cuda…
Ste
  • 58
  • 7
2
votes
1 answer

How to repeat forward and backward solve using cuSolver by CUDA?

Recently I developed a new method. The new method works perfect with CUDA (at 20 to 40FPS) and I have already tested it successfully. The problem comes when I try to make a comparison with an old method. The old method was implemented on CPU. It…
whmin
  • 21
  • 3
2
votes
1 answer

eigenvalue and eigenvector calculation with cusolver from cuda 7.0 RC

I'm trying to calculate the largest eigenvalue/eigenvector pair with cuSolver released in CUDA 7.0 RC. The problem is that I'm getting CUSOLVER_INTERNAL_ERROR, and I don't know what can I do about it. This is my handy stuff, used to call…
Leonardo Miquelutti
  • 195
  • 2
  • 8
1
vote
1 answer

Segmentation fault when using cusolverSpScsrlsvchol in CUDA for sparse linear problems

I'm trying to port a linear problem to CUDA in order to speed up solving times. I have successfully used cusolverDn to handle dense problems on the GPU. However, when I attempted to apply it to sparse problems using cusolverSpScsrlsvchol, I keep…
Finn Eggers
  • 857
  • 8
  • 21
1
2 3 4