Questions tagged [cula]

CULA is a linear algebra library using NVIDIA's CUDA architecture. Like most GPU-based solutions, massive speed gains are possible for highly parallelizable problems.

CULA (http://www.culatools.com/) is a linear algebra library using NVIDIA's CUDA architecture. Like most GPU-based solutions, massive speed gains are possible for highly parallelizable problems.

25 questions
2
votes
1 answer

Can CULA routines be called from device kernels?

So I'm trying to see if I can get some significant speedup from using a GPU to solve a small overdetermined system of equations by solving a bunch at the same time. My current algorithm involves using an LU decomposition function from the CULA…
elswindle
  • 53
  • 4
2
votes
1 answer

MAGMA and Rcpp for linear algebra in R

I was wondering if anybody has tried to use Rcpp and MAGMA to accelerate linear algebra operations in R by using the CPU and GPU? I tried culatools last month and it worked with Rcpp (link), but culatools is a commercial product that costs money to…
chris
  • 461
  • 2
  • 10
2
votes
2 answers

Rcpp and CULA: segmentation fault

I extracted the relevant bits from the gputools R-package to run a QR decomposition on my GPU using Rcpp by dynamically loading a shared library that links to culatools. Everything runs smoothly in the terminal and R.app on my Mac. The results agree…
chris
  • 461
  • 2
  • 10
1
vote
2 answers

gputools: error in installation

I am setting up a new Dell Precision workstation with an NVidia Tesla 2050 GPU card. I would like to install R's package gputools. My OS is openSuse 11.3 with KDE 4.4. I downloaded NVidia's CUDA Toolkit 3.2 and installed it in /usr/local/cuda, I…
Michaela
  • 11
  • 2
1
vote
0 answers

Python CULA Sgesv Ax=B solving but gives negatives for positives? Why?

I am using Python CULA Sgesv to solve a for a matrix operation. When I compare the answer from CULA to linear algebra solution CULA returns the correct numbers, but reverses the sign of the number. So if the real solution is positive the CULA…
Jesse
  • 213
  • 3
  • 13
1
vote
1 answer

How to use 3D matrices with CULA on a GPU?

In a CPU version of some Code, I have many things that look like the following: for(int i =0;i
LindleyLentati
  • 161
  • 3
  • 12
1
vote
1 answer

How to copy a matrix in a bigger matrix in CUDA

I want to setup a big matrix on my GPU to solve the according equation system with CULA. Some numbers for you, to understand the problem: big matrix: 400x400 small matrices: 200x200 Now I want to copy every quarter (100x100) of the small matrix…
Stefan
  • 2,460
  • 1
  • 17
  • 33
1
vote
1 answer

Improving Matlab + CUSP MEX solution for A*x=B on CUDA GPU

Matlab still can't compute sparse matrices on CUDA GPU. There are no such toolboxes (Jacket is discontinued) for that as well. That's why I am using CUSP integrated to Matlab through MEX file. However, my developed tool has two problems: It is VERY…
0
votes
1 answer

Inversion of two matrices on a gpu using cublasSgetriBatched

I am new to cublas. I want to calculate the inverse of two matrices in parallel on a GPU. The matrices are [4 8;3 9] and [5 2;1 7]. Is it possible to do so using cublasSgetriBatched? Here is my code, I am getting incorrect result with this. Here I…
0
votes
0 answers

Cula routine sgbtrf undefined

When I try to call the CULA routine GBTRF I get this error identifier "culaSgbtrf" is undefined, but everything else (inizializaion, GetErrorInfoStrin etc ..,) seems to work fine. Does anybody know why and can help me?
Shad
  • 1
0
votes
1 answer

Theano: mixing CPU and GPU?

I built a neural network which needs to use the Cholesky decomposition and solve triangular systems as part of its computation. This means that I also need to compute the gradients of the whole computation, of course. When I try to compile the code…
Kiuhnm
  • 478
  • 3
  • 13
0
votes
1 answer

CULA - using Python solve() within CULA

How can I incorporate Python's solve() in a Python CULA program? I am using LA = libculaC.solve() Result: Traceback (most recent call last): File "culaTest.py", line 96, in LA = libculaC.solve(0) File…
Jesse
  • 213
  • 3
  • 13
0
votes
0 answers

CUDA lapack librairies (CULA & MAGMA) as device functions

Could calls to lapack routines in CUDA (CULA or MAGMA) be instantiated within a CUDA kernel and not from the Host? (__device functions and not __global functions) If it is not possible, how to therefore implement (_device) linear algebra routines…
Didon
  • 383
  • 2
  • 4
  • 13
0
votes
2 answers

Load CULA dense library into python on jetson tk1 board

Having successfully downloaded and installed the full CULA dense R18 package from this website (http://www.culatools.com/downloads/dense/) onto my jetson TK1 board, I haven't been able to load the library into python using ctypes. I've changed…
Dan Lenton
  • 105
  • 9
0
votes
2 answers

Different QR decomposition results with numpy and CULA

I'm performing QR decomposition in two different ways: using standard numpy method and using GEQRF LAPACK function implemented in CULA library. Here is simple example in python (PyCULA used to access CULA): from PyCULA.cula import…
grapescan
  • 251
  • 1
  • 5
  • 10
1
2