GPU Coder Interface for Deep Learning

Use GPU Coder to generate optimized CUDA code for deep learning networks

Deep Learning Toolbox, use this support package.GPU Coder generates optimized CUDA code from MATLAB code and Simulink models for deep learning, embedded vision, and autonomous systems. You can deploy a

MATLAB Coder Support Package for NVIDIA Jetson and NVIDIA DRIVE Platforms

NVIDIA hardware target support from MATLAB Coder and GPU Coder

the generated code on the target hardware board. It enables you to remotely communicate with the NVIDIA target and control the peripheral devices for prototyping and testing purposes.When used with GPU

GPUBench

Compare GPUs using standard numerical benchmarks in MATLAB.

GPUBENCH times different MATLAB GPU tasks and estimates the peak performance of your GPU in floating-point operations per second (FLOP/s). It produces a detailed HTML report showing how your GPU

A GPU Mandelbrot Set

Explore the Mandelbrot Set using MATLAB and a GPU.

This application allows you to explore the wonders of the Mandelbrot Set in MATLAB with the help of a capable GPU. It is primarily intended as a demonstration of the different ways in which a MATLAB

GPU Julia Set Explorer

Explore the Julia Set of the Mandelbrot Set using MATLAB and a capable GPU.

This application allows you to explore the Julia Set of the Mandelbrot Set in MATLAB with the help of a capable GPU. It is primarily intended as a demonstration of element-wise calculations using

GPU sparse, accumarray, non-uniform grid

performance gains above 1000x over matlab spmv can be observed with cuda compatible GPU.

bare-bone interface with cusp sparse class for gpus, support for single precision, real/complex. Usage:A=gcsparse(B,[format: 0=coo, 1=csr]);or A=gcsparse(col,row,val,[nrows,[ncols,[format]]]);input B

Color-balance demo with GPU computing

Use GPU in MATLAB to perform white-balance operation to input image.

This demo shows how to identify bottlenecks in code that is run on a CPU using the MATLAB Profiler. The computations forming the bottleneck in this example are then executed on the system's GPU

3D pseudo polar Fourier and Radon Transforms

Forward and inverse 3D pseudo polar Fourier transforms and Radon transforms

the PPFT/Radon transforms for GPU computation.4) Direct inversion - Fast direct inversion for the PPFT/Radon transforms, running time is independent on the input data, no convergence criterion is

Accelerating Heat Equation on the GPU

A GUI for comparing the performance of different implementations of the heat equation on CPU and GPU

HEATEQGUI is an interactive GUI that lets you compare different ways for solving the heat equation. Only minor code changes were necessary to run the code on the GPU instead of the CPU. *

GPU Implementation for GVF Force Field

We implement the GVF force field on GeForce GPU using CUDA.

time, in this project we implement the GVF algorithm with GPU, which will accelerate the algorithm to a great extent.

Tutorials: Parallel and GPU Computing with MATLAB: All in one (9 parts)

Tutorials on Parallel and GPU Computing with MATLAB

This submission contains all code examples used in tutorial series for Parallel and GPU Computing with MATLAB available here: http://www.mathworks.com/products/parallel-computing/tutorials.htmlTopics

NVSMI Toolbox

Monitoring of NVIDIA GPU devices

The NVSMI Toolbox is a set of functions that wrap the nvidia-smi executable shipped with the NVIDIA display driver (Linux and with 64bit Windows). It adds monitoring features of the NVIDIA GPU(s

Fast 3D Non Local Means Filtering using Convolutions CPU/GPU

Domain Decomposed CPU/GPU implementation of Non Local Means Filtering in 3D using Convolution

serialGPUNLMF(domain,locSize,DoS,k,npasses,gpuFlag)nlmf: the output domaindomain: the input 3D arraylocSize: the subdomain size in X,Y,Z. e.g. locsize=256 will partition the domain into subvolumes of 256^DoS: the degree of smoothing (commonly 0.05 to

Huygens-Fresnel integral approximation on the GPU

Simulate electromagnetic wave propagation through free-form apertures, or off rough surfaces. Speed up the computation by using the GPU.

Huygens-Fresnel integral.Toolbox features are:* GPGPU computing, using Nvidia graphics cards with CUDA* fallback to CPU if no GPU is found* rough surface generation via spatial frequency filters and surface

Weeks' Method for Numerical Laplace transform inversion with GPU acceleration

Implementation of the Weeks method for numerical Laplace transform inversion with GPU acceleration.

Weeks method. Particularly new here is the use of graphics processing unit [GPU] computing to accelerate the method.To assist the user, a wrapper (WeeksMethod.m) to the core inversion functions is

Spectral Visual Saliency Toolbox

Implementations of several state-of-the-art visual saliency detection algorithms.

Real-Time Object Detection with YOLO v2 Using GPU Coder

Example of real-time object detection using YOLO v2 on NVIDIA GPUs

You can use GPU Coder™ in tandem with the Deep Learning Toolbox™ to generate code and deploy deep learning networks on embedded platforms that use NVIDIA^® Jetson and Drive platforms. The pretrained

Large Data in MATLAB: A Seismic Data Processing Case Study

These are the files used in the webinar on Feb. 23, 2011.

to seismic analysis (Kirchhoff migration, reverse time migration) 2 - Large data extension of the functionality shown in (1) and parallel computing for speeding up the processing time 3 - GPU extension

TIGRE: Tomographic Iterative GPU-based Reconstruction Toolbox

3D tomographic reconstruction software

TIGRE: Tomographic Iterative GPU-based Reconstruction ToolboxTIGRE is a GPU accelerated software for big scale 3D tomographic reconstruction, being capable of reconstructing geometries such as Cone

GPU Centerline Extraction (Skeletonization) in 2D or 3D

Extracts the centerlines (skeleton) of binary 2D images or 3D volumes using bit encoded thinning on the GPU.

This code provides implementation of the real-time thinning / centerline extraction techniquesproposed in "Real-time thinning algorithms for 2D and 3D images using GPU processors" (Wagner, 2019

GPU Accelerated Edge-Region Based Level Set Evolution Constrained by 2D Gray-Scale Histogram

Here are the serial and the GPU based implementation of our paper IEEE TIP.2013.2255304

our work. The serial implementation can be runned even if you don't have an NVIDIA GPU. But this is not the case of the parallel implementation. Everything is done as simple as possible in order to make

Digital Image Correlation and Tracking

Calculate displacement, strain and stress from image sequences

Quantized Deep Neural Network on Jetson AGX Xavier

How to create, train and quantize network, then generate CUDA C++ code for targeting Jetson AGX Xavier

data types. And then you can use GPU Coder to generate optimized CUDA code for the quantized network.This example shows how to create, train and quantize a simple convolutional neural network for defect

fastRPCA

Code for Robust PCA

vLSTM

Vectorized multimodal LSTM using Matlab and GPU

Vectorized Long Short-term Memory (LSTM) using Matlab and GPU

Hardware accelerated ray-box intersection

GPU portable implementation of the ray-box intersection method of Smits (1998)

.% Call with gpuarray and arrayfun to execute on the GPU: this % may give two orders of magnitude speed up over vectorized % cpu based implementation % Ray box intersection is typically used to locate (axis

immiscible LB

Implements Immiscible Lattice Boltzmann (ILB, D2Q9) method for two phase flows

GPU vs CPU speed test of finite difference equation

Compares the speed of the parallel computing toolbox functions vs CPU for finite difference

Simulates the heat equation, with constant heat capacity and thermal conductivity, using GPU (parallel computing toolbox) or CPU (matrix calculations). Includes results from Nvidia titan and i5-2500k

Zeffiro Forward and Inverse Interface for Complex Geometries

Interface for using finite elements in inverse problems with complex domains

beaccelerated significantly in a computer equipped with a graphics computing unit (GPU). It isespecially recommendable to perform the forward simulation process, i.e., to generate the finiteelement mesh, the lead

Fast Checkerboard Demodulation for digital Schlieren imaging

Track optical distortions in a checkerboard pattern with high accuracy in real-time using the FCD method

FFT- Includes a live preview function- Runs on GPU without modificationsMore info: https://arxiv.org/abs/1712.05679

Suite2P

Fast, complete two-photon pipeline

Hardware accelerated ray-triangle intersection

GPU portable implementation of the ray-triangle intersection method of Moller and Trumbore (1997)

% Ray-triangle intersection algorithm of Muller and Trumbore (1997)% formatted for arrayfun to allow hardware acceleration% Call with gpuarray and arrayfun to execute on the GPU: thjs% may give two

Tutorials: Parallel and GPU Computing with MATLAB (9 of 9): GPU Computing with MATLAB

Tutorial on Parallel and GPU Computing with MATLAB (9 of 9)

This submission contains code examples used in part 9 of tutorial series on Parallel and GPU Computing with MATLAB. This part covers using GPU-enabled MATLAB functions, executing NVIDIA^® CUDA™ code

whosGPU

Provides WHOS-like output for gpuArray variables in the current workspace.

Anyone who has worked with gpuArray objects through Matlab's Parallel Computing Toolbox may have found it frustrating that the whos command does not reveal as much information about data size and

Fast 2D GPU-based convolution

Graphics chip assisted fast 2d convolution

SiftGPU- SIFT enabled on GPU

implementation of SIFT compiled on graphics card

WarpFactory

WarpFactory is a numerical toolkit for analyzing warp drive spacetimes.

Energy ConditionsMetric scalar evaluation for the shear, expansion, and vorticityMomentum flow visualizationsGPU utilization for accelerated computationsAnimation of the stress-energy tensor. Lighter

GPU CUDA convolution 2D 3D

matlab wrapper for CUDA 2D and 3D GPU-accelerated convolution

C++/CUDA GPU-accelerated convolution in 2D and 3D. Based on NVIDIA cuda-samples convolutionFFT2D combined with matlab

inpolygon_for_gpu

inpolygon function that works using gpuArray

This is a point-in-polygon function that can run on a gpu using large test point array sizes. It uses a simple ray-casting algorithm without pre-processing or "on" tolerance checks. Therefore it may

drawVector- draws 2D or 3D vectors from specified points

Draws 3 arrows representing the basis vectors of an R3 coordinate system

FDTD - 1D, 2D, 3D Simple Free Space Examples

Vectorized FDTD code with GPU functionality for the 3D case. Code is nicely organized and easy to understand.

A point source located at the center of the simulation domain generates electromagnetic radiation which then propagates through vacuum.Using a GPU for the 3D case, one can realize the performance

Deploying Deep Neural Networks to Embedded GPUs

How to create, train and deploy deep neural networks for embedded GPUs

Designing and deploying deep learning and computer vision applications to embedded GPU platforms is challenging because of resource constraints inherent in embedded devices. This demo shows how to

Interactive Incompressible Fluids

A GPU-enabled interactive demo of Navier-Stokes equations for incompressible fluids.

3D Linear Interpolation for GPU

3D Linear Interpolation for GPU

This function is faster than MATLAB's griddedInterpolant function for the CPU, but slower than MATLAB's interpn function for the GPU. However, I've coded this using arrayfun. Since MATLAB does not

MATLABによる画像処理・コンピュータービジョン・点群処理評価キット

MATLAB image processing, computer vision, and point cloud processing evaluation kit in Japanese

Driving ToolboxRoadRunnerSensor Fusion and Tracking ToolboxNavigation ToolboxRobotics System ToolboxUAV ToolboxROS ToolboxMATLAB CoderSimulink CoderGPU Coder

usefulSnippets

Collection of some "little" functions I wrote to make my life easier.

Efficient three-dimensional (3D) Gaussian smoothing using convolution via frequency domain

Native Fourier implementation, support GPU computation and anisotropic voxel.

(z*pixelspacing(3)).^2/sigma(3)^2/2)); Remarks The outputs of gauss3filter(I), gauss3filter(I, 1) and gauss3filter(I, 1, [1 1 1]) are identical. To enable GPU computation (Matlab 2012a or later, CUDA 1.3 GPU are required), use

GPU Fractal Explorer

View and investigate various fractals using MATLAB and an NVIDIA GPU

This application allows you to explore several different fractals in MATLAB with the help of a capable NVIDIA GPU. It was inspired by Cleve Moler's article GPU Enables Obsession with Fractals, where

MATLAB for CUDA Programmers

This is the code used during the MATLAB for CUDA Programmers webinar

Runnable demos showcasing the GPU computing capabilities of Parallel Computing Toolbox. Comes with a reference implementation in non-GPU MATLAB used to verify the correctness of the GPU

Metal for MATLAB

Apple Metal GPU processing toolbox for MATLAB on macOS

MATLABMetalApple Metal GPU processing toolbox for MATLAB on macOS.Matlab runs extremely well on the new Apple Silicon Macs, but if you want the best possible performance from these new processors

GPU from MATLAB(R) example

Simple example showing how to access the GPU from MATLAB

This is a simple example showing how to access the GPU from MATLAB. This example requires a Freeware library called GPUmat, that can be downloaded from http://gp-you.orgThe example shows the

Image white balancing

White balance camera-rendered sRGB images (CVPR 2019)

chart for Set1 images.Graphical user interfaceWe provide a Matlab GUI to help tuning our parameters in an interactive way. Please, check demo_GPU.m. Code/GUI parameters and optionsK: Number of nearest

NVIDIA CUDA-based bilinear (2D) interpolation

Incredible speed boost in comparison to the Matlab implementation. (interp2)

This code was inspired by Alexander Huth's bilinear interpolation approach( http://www.mathworks.com/matlabcentral/fileexchange/20248 )also using the GPU's built-in bilinear texture interpolation

Modified-MannKendall-Test

A modified version of the Mann-Kendall Test that works with autocorrelated data.

Fast Continuous Max-Flow Algorithm to 2D/3D Image Segmentation

Fast continuous max-flow algorithm to 2D/3D image segmentation developed in matlab, C and GPU

This software implements the fast continuous max-flow algorithm to 2D/3D image segmentation. It provides three implementations: matlab, C and GPU (cuda based). All the source files are provided. So

GPUCONV2

Example, Matlab R2010B Cuda CONV2 on GPU using Cuda kernels

GPUCONV2 Two dimensional convolution on the GPU using Cuda. C = GPUCONV2(A, B) performs the 2-D convolution of matrices A and B. If [ma,na] = size(A), [mb,nb] = size(B), and [mc,nc] = size(C

Assign GPU tokens to workers in spmd

This piece of code tries to give a frame work to allowing the number of workers who can use the GPU synchronously.

This piece of code tries to give a frame work to allowing the number of workers who can use the GPU synchronously, since oftentimes the use of gpu in all workers may lead to outage of gpu memory

Mac GPT GPU Benchmark Explorer

Live Script that benchmarks a small GPT on Apple Silicon with MATLAB-CPU, PyTorch-CPU/MPS, and MLX.

GPT training is usually relegated to GPUs, often intentionally without full floating point precision. MATLAB has no GPU backend on Apple Silicon: gpuArray and the Deep Learning Toolbox training path

GPU-accelerated 3D Stationary Wavelet-based Frame

3D Stationary wavelet-based transform using GPU-accelerated separable convolution

This implements a GPU-accelerated version of the 3D frame used in "Alp Kucukelbir, Fred J. Sigworth, Hemant D. Tagare, A Bayesian adaptive basis algorithm for single particle reconstruction, Journal

gpuSiftMatch(d1,d2,thresh,bunchSize,gpuEn,precision)

Fast SIFT feature matching using GPU

This feature matching code is designed to replace the ubc_match function from vl_feat (http://www.vlfeat.org/overview/sift.html) This code is significantly faster when run on a compatible GPU.