Questions tagged [cudalink]
Questions about using CUDALink, Mathematica's method to accelerate programs using NVidia graphics cards.
111
questions
28
votes
3
answers
5k
views
How to find the center of a circular pattern?
I am writing an algorithm that finds the center of 500 x 500 pixels images coming from a live feed from a camera at 50 Hz. So it has to be quick! The images are images of an optical field which looks ...
- 2,470
19
votes
1
answer
2k
views
How to use SparseArray in CUDA?
Sparse arrays are widely used in many scientific applications where memory and time are strictly limited. For a considerable speedup one can use CUDA. Since Mathematica 8 we can conveniently use CUDA ...
- 43.7k
19
votes
1
answer
2k
views
Why does CUDAQ (from CUDALink) download data from Wolfram servers?
I understand why functions like ElementData or CountryData need to download up-to-date data when first launched. However, when ...
- 10.8k
17
votes
2
answers
812
views
Combining images with CUDA
SMALL UPDATE: This question is left as unanswered, because none of the existing answers as of yet actually use CUDA (though, whilst being very useful).
2nd Edition:
To make it a bit more clear, and ...
- 1,652
15
votes
0
answers
161
views
How to use user-provided CUDA toolkits?
In the revision history, Mathematica v12.1 claims to have "support for user-provided CUDA Toolkits". But other than this line in the revision history, I cannot find any other reference to ...
- 1,176
13
votes
1
answer
1k
views
calling CUDA library functions
Mathematica has some predefined CUDA functions, like CUDAMap, CUDAFourier, etc. My question is how can one call functions which are part of CUDA library but not part of predefined Mathematica ...
- 487
12
votes
4
answers
762
views
CUDA on Tesla V100 card
Context
I am trying to run CUDALink with a
Centos 6.10 Mathematica 12.0 [cuda 10.1]. Host has NVIDIA Corporation GV100GL [Tesla V100 PCIe 32GB] (rev a1) card
...
- 22.9k
12
votes
3
answers
2k
views
Are you able to use CUDA on Linux?
I am using Mathematica 10.2 on Ubuntu 15.04 x64. At first, I try to use CUDALink:
Needs["CUDALink`"];
CUDAQ[]
It returns ...
- 677
11
votes
3
answers
2k
views
Mathematica 12, supported GPUs
It seems that Mathematica 12 doesn't support NVidia 2xxx (i.e 2080), 16xx or 1xxx series. Could it be that the CUDA section has not yet been rewritten for 12?
https://reference.wolfram.com/language/...
- 420
10
votes
1
answer
541
views
How to make a volume CUDA rendering isometric?
I just realized that the output of CUDAVolumetricRender[] uses (quite marked) perspective, which I do not want (for scientific data representation, it can be very ...
- 10.8k
10
votes
2
answers
3k
views
CUDA and GPU hardware and compatibility questions?
We have iMac desktops and have just begun to parallelize computations across them see: Wolfram Light Weight Grid and parallel computing.
This still does not look like it will give us the performance ...
- 14.3k
10
votes
1
answer
995
views
CUDALink not working on Mac OS X
On my MacBook Pro, I have a CUDA enabled video card, yet Mathematica does not seem to recognize this. For example, following the CUDALink Setup, I get
...
- 34k
9
votes
2
answers
2k
views
Mathematica cannot find CUDA driver
I have successfully installed CUDA 5.0 on my Ubuntu 12.10, with driver 304.54, toolkit and samples. Running deviceQuery gives me a successful output:
...
- 2,470
9
votes
0
answers
1k
views
Poor CUDALink performance, when calling a kernel multiple times
When working on a project of mine I noticed that CUDALink shows very poor performance when calling CUDA kernels multiple times in a loop-construct.
To demonstrate this I put together this little ...
8
votes
2
answers
573
views
Use of Cellular Automaton or Discrete Convolution to calculate Potts Energy of a 2D Matrix
In a project of Monte Carlo Modelling of Grain Growth (Link) one of the most costly subroutines is the calculation of the Potts-Energy of an Monte Carlo Grid cell by comparing the center cell value ...
- 2,891
8
votes
1
answer
268
views
Can CUDAVolumetricRender be combined with other 3D graphics?
I have volumetric data, which I can render using CUDAVolumetricRender[]. However, I cannot seem to find a way to combine it with other 3D graphics, such as adding a ...
- 10.8k
8
votes
1
answer
714
views
Can't use CUDA shared memory
CUDALink is properly installed and up to date, all built-in functions work fine, and all little functions I did worked fine... Until I tried using shared memory.
The only example I found on the help ...
- 42.6k
7
votes
2
answers
667
views
7
votes
2
answers
993
views
CUDAQ not working in Mathematica 10.2.0.0 on Windows 7 (using CUDA 7.5)
I have been trying to get CUDA to work in Mathematica for kernel development.
I downloaded and installed Visual Studio 2013 Community and the latest NVIDIA developers toolkit 7.5 (in that order).
...
- 345
7
votes
1
answer
1k
views
CUDA for linear equations
TLDR; Is there a way to solve linear equations of a sparse matrix (discretized laplace operator) efficiently using CUDALink in Mathematica? I didn't find a CUDALinearSolve or CUDAMatrixInverse or ...
- 73
7
votes
2
answers
983
views
CUDALink not working on Linux-64 and Asus GTX460
On my hardware I have a CUDA enabled video card an ASUS GTX460, yet Mathematica does not seem to recognize this.
I have:
CUDA Toolkit 4.1
CUDA SDK 4.0
NVIDIA Driver 295.49
on Sabayon Linux 64 bit ...
- 10.2k
7
votes
1
answer
1k
views
make mathematica recognize CCompiler and compile CUDA code on remote Windows machine
I have set up a Mathematica remote kernel on a Windows 7 64 bit machine, which has Microsoft Visual Studio 2010 installed. When I start Mathematica on the windows machine Mathematica recognizes the C-...
- 2,720
7
votes
1
answer
228
views
A simple experiment to understand CUDAFunctionLoad
In an attempt to answer my own question (Beginner CUDA: two dimensional blocks and two dimensional threads?).
Please excuse the long post, it contains for examples I constructed as an experiment to ...
- 4,415
7
votes
1
answer
1k
views
CUDALink is not initialized
I've upgraded to Mathematica 9.0 on Ubuntu 12.04 with the latest Nvidia driver (304.64).
I load CUDA by using, as usual, Needs["CUDALink`"].
...
- 71
7
votes
0
answers
191
views
Extremely Long GPU Initialization Times on Mathematica 12.2 and RTX3090 [closed]
I'm having trouble getting my new GPU system to work with Mathematica 12.2. I have two RTX3090 cards in my setup. My NVIDIA drivers are version 460.89 (the latest).
I have installed CUDAToolkit 11.0 ...
7
votes
0
answers
241
views
Getting CUDALink (Toolkit 7.5) to work on Linux, Mathematica 10.0.2, GTX750
I want to make use of some CUDA functionality on my machine. It is running Gentoo Linux (64bit), Mathematica 10.0.2 on an Intel Xeon E5-1620v3 with a Geforce GTX750. So everything on the hardware/OS ...
- 2,702
6
votes
3
answers
818
views
How to properly remove the CUDAResources paclet?
I have Mathematica 8.0.0.0 installed on my Mac, and while I tried to see what it could do with CUDA, it downloaded a large directory (~1 GB) named ...
- 10.8k
6
votes
2
answers
738
views
CUDA: setting grid dimensions
I'm new to CUDA, and CUDALink.
I see how CUDAFunctionLoad allows one to set the block dimensions. On calling the function, you can also either pass as argument or have MMA automatically select the ...
- 42.6k
6
votes
1
answer
782
views
CUDALink support-Mathematica 11
Greetings & Salutations,
I have Mathematica 11.0 on a Windows 10(Home) system with a Nvidia GTX1050Ti gpu, driver version 382.33 and CUDA version 8.0.61; I am getting CUDALink is not supported ...
- 197
6
votes
1
answer
5k
views
How is the dimensionality of an OpenCL problem specified?
When writing an OpenCL kernel, calls to get_global_id() are used to determine the "position" within the problem that a particular instance of the kernel is working ...
- 163
6
votes
1
answer
433
views
CUDALink Error: A CUDALink kernel timed out?
I'm following the CUDALink tutorial [ http://reference.wolfram.com/language/CUDALink/tutorial/Introduction.html ] and everything is going fine until I get to the part that calls CUDADot on the ...
- 42.3k
6
votes
0
answers
338
views
Does OpenCL/Cuda memory leak if not Unloaded? Problem with allocating/loading GPU memory in Module
The GPU memory manager documentation is not too clear, does GPU memory leak if not explicitly unloaded or if the symbol that holds the reference to the GPU memory gets collected, the GPU memory gets ...
- 161
5
votes
1
answer
513
views
Volumetric Rendering of cylinder symmetric radial image
I would like to render a cylinder symmetric radial cross-section with Image3D. As an example one could use the propagation of a Bessel Beam through a Lens:
Here the cylinder axis is the bottom of the ...
- 2,891
5
votes
1
answer
412
views
cannot load CUDAFunction from cubin, ptx or library file
According to the documentation of CUDAFunctionLoad it should be easy to specify a compiled file (cubin, ptx, dll should all work) as the source for loading a ...
- 2,720
5
votes
2
answers
566
views
CUDA 6.5, Mathematica 10.0.2, and Ubuntu 14.04
I have CUDA 6.5 working fine on Ubuntu 14.04. Mathematica recognizes the driver (Nvidia 340), but CUDAQ[] returns False and I cannot get any CUDA functions to work in Mathematica.
I have created ...
- 183
5
votes
0
answers
168
views
Target Device->"GPU" Not Working For Machine Learning: LibraryFunction error
I'm running Wolfram Desktop version 13.2.1, and after downloading Visual Studio, Nvidia Toolkit version 12.1 and 11.2, and verifying that
...
- 683
5
votes
0
answers
174
views
Is there a Mathematica CUDA ode solver?
I have a system of ODEs which I'd like to solve with many (thousands) different initial conditions. In the net there are articles that CUDA can speed up such computations, but I haven't found ...
- 2,513
5
votes
0
answers
205
views
Fast N-Body Simulation with CUDA [closed]
There is a Mathematica code for 3 body simulation
...
- 8,958
5
votes
0
answers
700
views
How to invert a matrix with 100 trillion elements?
I have a $10^{7}\times 10^{7}$ matrix with the following properties:
it is sparse
it is gamma-5
it is almost Hermitian.
If $M$ is the matrix we are trying to invert, we know $M=A M^{\dagger} A$ ...
- 31.5k
5
votes
0
answers
988
views
Beginner CUDA: two dimensional blocks and two dimensional threads?
This is a follow up question to: CUDA: setting grid dimensions. The purpose of the question is to understand how Mathematica is interfacing with CUDA's architecture. I have a related question/request ...
- 4,415
4
votes
1
answer
634
views
Help with CUDA nvcc error in Mandelbulb example Mathematica 11.0 NVIDIA GTX 1070 on Ubuntu
I am exploring the Nvidia GTX 1070 with Mathematica, I tried this example from the Help guide and it fails with a nvcc complier error. A little google research seems to indicate that I need the nvcc ...
- 258
4
votes
1
answer
763
views
CreateLibrary for CUDA with NVCCCompiler
I am currently learning to use CUDA to speed up calculations. The book I am working with uses C code on the CPU to supply certain variables etc.. So I would like to compile the give code in ...
- 1,476
4
votes
1
answer
591
views
CUDAQ[] returns False in version 9 while True in version 8
CUDA works in version 8 but not in version 9. Is there a way to fix the problem in version 9? I've installed the latest CUDA from the official website https://developer.nvidia.com/cuda-downloads
- 27.6k
4
votes
1
answer
345
views
How to determine MMA/CUDA support for specific GPUs?
This question asked how to target an Nvidia GTX 1080 with MMA 11.0.1 (which is also my current version) and from the comments it is clear that the user was then able to use this card successfully even ...
- 2,560
4
votes
1
answer
438
views
How to troubleshoot CUDALink?
This is a follow-up question from linking the libraries for CUDALlink in Ubuntu 12.10.
The tests that I execute (CUDAQ[], ...
- 2,470
4
votes
2
answers
482
views
CUDALink gives wrong results but no error on Mathematica 13.2
I installed Mathematica 13.2 under Linux Mint 21.1 (Ubuntu 22.04). After installing cuda-toolkit-11-8. I get True as answer from CUDAQ[], but if I start some calculations I get wrong results.
...
- 877
4
votes
0
answers
215
views
Interactive CUDAFluidDynamics example without CUDA?
On Macbooks OpenCL works but not CUDA. I saw this fun example CUDAFluidDynamics in the CUDALink package and I wanted to play with it, but couldn't.
Perhaps someone ...
- 31.5k
4
votes
0
answers
507
views
NVCC CUDA Compiler not showing up
I am trying to follow instructions for "Compiling for CUDA", so that I can write my own functions to run on the GPU.
However, CUDAFunctionLoad gives me errors.
...
- 141
4
votes
0
answers
151
views
CUDA not recognized on MacBookPro
I am running
Mma 11.1.0 on
MacOS 10.12.4 on
MacBook Pro (Retina, 15-inch, Mid 2014) with
NVIDIA GeForce GT 750M 2048 MB
I installed the Nvidia CUDA driver 8.0.71.
I evaluate
...
- 1,919
4
votes
0
answers
308
views
CUDADot runs slowly and writes "CUDALink experienced a kernel launch failure" with matrix size > 3300
I am trying to learn CUDA and start from the help examples. It turnes out that CPU calculation is much faster for matrix sizes < 3200 than CUDA, while CUDADot gives the error for sizes > 3300. I ...
- 215