CUDA Samples

The reference guide for the CUDA Samples.

1. Release Notes

This section describes the release notes for the CUDA Samples only. For the release notes for the whole CUDA Toolkit, please see CUDA Toolkit Release Notes.

1.1. CUDA 11.6

  • All CUDA samples are now only available on GitHub repository. They are no longer available via CUDA toolkit.
  • Added new folder structure for samples.
  • Added Visual Studio 2022 support to all the samples.

1.2. CUDA 11.5

  • Added 4_CUDA_Libraries/cuDLAHybridMode. Demonstrate usage of cuDLA in hybrid mode. (available only on GitHub repository)
  • Added 4_CUDA_Libraries/cuDLAStandaloneMode. Demonstrate usage of cuDLA in standalone mode. (available only on GitHub repository)
  • Added 4_CUDA_Libraries/cuDLAErrorReporting. Demonstrate DLA error detection via CUDA. (available only on GitHub repository)
  • Added 3_CUDA_Features/graphMemoryNodes. Demonstrates memory allocations and frees within CUDA graphs using Graph APIs and Stream Capture APIs. (available only on GitHub repository)
  • Added 3_CUDA_Features/graphMemoryFootprint. Demonstrates how graph memory nodes re-use virtual addresses and physical memory. (available only on GitHub repository)

1.3. CUDA 11.4 Update 1

  • Added support for VS Code on linux platform.

1.4. CUDA 11.4

  • Added 7_CUDALibraries/simpleCUBLAS_LU. Demonstrates batched matrix LU decomposition using cuBLAS API cublas<t>getrfBatched().
  • Updated 2_Graphics/simpleVulkan, 2_Graphics/simpleVulkanMMAP and 3_Imaging/vulkanImageCUDA. Demonstrates use of SPIR-V shaders.
  • Removed 7_CUDALibraries/boundSegmentsNPP.

1.5. CUDA 11.3

  • Added 0_Simple/streamOrderedAllocationIPC. Demonstrates IPC pools of stream ordered memory allocated using cudaMallocAsync and cudaMemPool family of APIs.
  • Updated 2_Graphics/simpleVulkan. Demonstrates use of timeline semaphore.
  • Updated 0_Simple/globalToShmemAsyncCopy with a partitioned cuda pipeline producer-consumer GEMM kernel.
  • Updated multiple samples to use pinned memory using cudaMallocHost().

1.6. CUDA 11.2

  • FreeImage is no longer distributed with the CUDA Samples. On Windows, see the Dependencies section for more details on how to set up FreeImage. On Linux, it is recommended to install FreeImage with your distribution's package manager.

1.7. CUDA 11.1

  • Added 2_Graphics/simpleVulkanMMAP. Demonstrates Vulkan CUDA Interop via cuMemMap APIs where CUDA buffer is imported in vulkan.
  • Added 7_CUDALibraries/watershedSegmentationNPP. Demonstrates how to use the NPP watershed segmentation function.
  • Added 7_CUDALibraries/batchedLabelMarkersAndLabelCompressionNPP. Demonstrates how to use the NPP label markers generation and label compression functions based on a Union Find (UF) algorithm including both single image and batched image versions.
  • Deprecated Visual Studio 2015 support for all Windows supported samples.
  • Dropped Visual Studio 2012, 2013 support from all the Windows supported samples.

1.8. CUDA 11.0

  • Added 0_Simple/globalToShmemAsyncCopy. Demonstrates asynchronous copy of data from global to shared memory using cuda pipeline. Also demonstrates arrive-wait barrier for synchronization.
  • Added 0_Simple/simpleAttributes. Demonstrates the stream attributes that affect L2 locality.
  • Added 0_Simple/dmmaTensorCoreGemm. Demonstrates double precision GEMM computation using the WMMA API for double precision employing the Tensor Cores. Also makes use of asynchronous copy from global to shared memory using cuda pipeline which leads to further performance gain.
  • Added 0_Simple/bf16TensorCoreGemm. Demonstrates __nv_bfloat16 (e8m7) GEMM computation using the WMMA API for __nv_bfloat16 employing the Tensor Cores. Also makes use of asynchronous copy from global to shared memory using cuda pipeline which leads to further performance gain.
  • Added 0_Simple/tf32TensorCoreGemm. Demonstrates tf32 (e8m10) GEMM computation using the WMMA API for tf32 employing the Tensor Cores. Also makes use of asynchronous copy from global to shared memory using cuda pipeline which leads to further performance gain.
  • Added 0_Simple/simpleAWBarrier. Demonstrates the arrive wait barriers.
  • Added warp aggregated atomic multi bucket increments kernel using labeled_partition cooperative groups in 6_Advanced/warpAggregatedAtomicsCG which can be used on compute capability 7.0 and above GPU architectures.
  • Added 0_Simple/binaryPartitionCG. Demonstrates binary_partition cooperative groups creation and usage in divergent path.
  • Added 6_Advanced/cudaCompressibleMemory. Demonstrates compressible memory allocation using cuMemMap API.
  • Removed 7_CUDALibraries/nvgraph_Pagerank, 7_CUDALibraries/nvgraph_SemiRingSpMV, 7_CUDALibraries/nvgraph_SpectralClustering, 7_CUDALibraries/nvgraph_SSSP as the NVGRAPH library is dropped from CUDA Toolkit 11.0.
  • Added two new reduction kernels in 6_Advanced/reduction one which demonstrates reduce_add_sync intrinstic supported on compute capability 8.0 and another which uses cooperative_groups::reduce function which does thread_block_tile level reduction introduced from CUDA 11.0
  • Added windows support to 6_Advanced/c++11_cuda.

1.9. CUDA 10.2

  • Added 6_Advanced/jacobiCudaGraphs. Demonstrates Instantiated CUDA Graph Update usage.
  • Added 0_Simple/memMapIPCDrv. Demonstrates Inter Process Communication using cuMemMap APIs with one process per GPU for computation.
  • Added 0_Simple/vectorAddMMAP. Demonstrates how cuMemMap API allows the user to specify the physical properties of their memory while retaining the contiguous nature of their access, thus not requiring a change in their program structure.
  • Added 0_Simple/simpleDrvRuntime. Demonstrates how CUDA Driver and Runtime APIs can work together to load cuda fatbinary of vector add kernel.
  • Added 0_Simple/cudaNvSci. Demonstrates CUDA-NvSciBuf/NvSciSync Interop.

1.10. CUDA 10.1 Update 2

  • Added 3_Imaging/vulkanImageCUDA. Demonstrates how to perform Vulkan Image-CUDA Interop.
  • Added 7_CUDALibraries/nvJPEG_encoder. Demonstrates encoding of jpeg images using NVJPEG Library.
  • Added Windows support to 7_CUDALibraries/nvJPEG.
  • Removed DirectX SDK (June 2010 or newer) installation requirement, all the DirectX-CUDA samples now use DirectX from Windows SDK shipped with Microsoft Visual Studio 2012 or higher

1.11. CUDA 10.1 Update 1

  • Added 3_Imaging/NV12toBGRandResize. Demonstrates how to convert and resize NV12 frames to BGR planars frames using CUDA in batch.
  • Added Visual Studio 2019 support to all the samples.

1.12. CUDA 10.1

  • Added 0_Simple/immaTensorCoreGemm. Demonstrates integer GEMM computation using the Warp Matrix Multiply and Accumulate (WMMA) API for integers employing the Tensor Cores.
  • Added 2_Graphics/simpleD3D12. Demonstrates Direct3D12 interoperability with CUDA.
  • Added 7_CUDALibraries/nvJPEG. Demonstrates single and batched decoding of jpeg images using NVJPEG Library.
  • Added 7_CUDALibraries/conjugateGradientCudaGraphs. Demonstrates conjugate gradient solver on GPU using CUBLAS/CUSPARSE library calls captured and called using CUDA Graph APIs.
  • Updated 0_Simple/simpleIPC to work on Windows OS as well with TCC enabled GPUs.

1.13. CUDA 10.0

  • Added 1_Utilities/UnifiedMemoryPerf. Demonstrates the performance comparision of Unified Memory and other types of memory like zero copy buffers, pageable, pagelocked memory on a single GPU.
  • Added 2_Graphics/simpleVulkan. Demonstrates the Vulkan-CUDA Interop. CUDA imports the Vulkan vertex buffer and operates on it to create sinewave, and synchronizes with Vulkan through vulkan semaphores imported by CUDA.
  • Added 0_Simple/simpleCudaGraphs. Demonstrates how to use CUDA Graphs through Graphs APIs and Stream Capture APIs.
  • Removed 3_Imaging/cudaDecodeGL, 3_Imaging/cudaDecodeD3D9 as the cuvid library is dropped from CUDA Toolkit 10.0.
  • Removed 6_Advanced/cdpLUDecomposition, 7_CUDALibraries/simpleDevLibCUBLAS as the CUBLAS Device library is dropped from CUDA Toolkit 10.0.

1.14. CUDA 9.2

  • Added 7_CUDALibraries/boundSegmentsNPP. Demonstrates nppiLabelMarkers to generate connected region segment labels.
  • Added 6_Advanced/conjugateGradientMultiDeviceCG. Demonstrates a conjugate gradient solver on multiple GPUs using Multi Device Cooperative Groups, also uses Unified Memory optimized using prefetching and usage hints.
  • Updated 0_Simple/fp16ScalarProduct to use fp16 native operators for half2 and other fp16 features, it also compare results of using native vs intrinsics fp16 operations.

1.15. CUDA 9.0

  • Added 7_CUDALibraries/nvgraph_SpectralClustering. Demonstrates Spectral Clustering using NVGRAPH Library.
  • Added 6_Advanced/warpAggregatedAtomicsCG. Demonstrates warp aggregated atomics using Cooperative Groups.
  • Added 6_Advanced/reductionMultiBlockCG. Demonstrates single pass reduction using Multi Block Cooperative Groups.
  • Added 6_Advanced/conjugateGradientMultiBlockCG. Demonstrates a conjugate gradient solver on GPU using Multi Block Cooperative Groups.
  • Added Cooperative Groups(CG) support to several samples notable ones to name are 6_Advanced/cdpQuadtree, 6_Advanced/cdpAdvancedQuicksort, 6_Advanced/threadFenceReduction, 3_Imaging/dxtc, 4_Finance/MonteCarloMultiGPU, 0_Simple/matrixMul_nvrtc.
  • Added 0_Simple/simpleCooperativeGroups. Illustrates basic usage of Cooperative Groups within the thread block.
  • Added 0_Simple/cudaTensorCoreGemm. Demonstrates a GEMM computation using the Warp Matrix Multiply and Accumulate (WMMA) API introduced in CUDA 9, as well as the new Tensor Cores introduced in the Volta chip family.
  • Updated 0_Simple/simpleVoteIntrinsics to use newly added *_sync equivalent of the vote intrinsics _any, _all.
  • Updated 6_Advanced/shfl_scan to use newly added *_sync equivalent of the shfl intrinsics.

1.16. CUDA 8.0

  • Added 7_CUDALibraries/FilterBorderControlNPP. Demonstrates how any border version of an NPP filtering function can be used in the most common mode (with border control enabled), can be used to duplicate the results of the equivalent non-border version of the NPP function, and can be used to enable and disable border control on various source image edges depending on what portion of the source image is being used as input.
  • Added 7_CUDALibraries/cannyEdgeDetectorNPP. Demonstrates the recommended parameters to use with the nppiFilterCannyBorder_8u_C1R Canny Edge Detection image filter function. This function expects a single channel 8-bit grayscale input image. You can generate a grayscale image from a color image by first calling nppiColorToGray() or nppiRGBToGray(). The Canny Edge Detection function combines and improves on the techniques required to produce an edge detection image using multiple steps.
  • Added 7_CUDALibraries/cuSolverSp_LowlevelCholesky. Demonstrates Cholesky factorization using cuSolverSP's low level APIs.
  • Added 7_CUDALibraries/cuSolverSp_LowlevelQR. Demonstrates QR factorization using cuSolverSP's low level APIs.
  • Added 7_CUDALibraries/BiCGStab. Demonstrates Bi-Conjugate Gradient Stabilized (BiCGStab) iterative method for nonsymmetric and symmetric positive definite linear systems using CUSPARSE and CUBLAS
  • Added 7_CUDALibraries/nvgraph_Pagerank. Demonstrates Page Rank computation using nvGRAPH Library.
  • Added 7_CUDALibraries/nvgraph_SemiRingSpMV. Demonstrates Semi-Ring SpMV using nvGRAPH Library.
  • Added 7_CUDALibraries/nvgraph_SSSP. Demonstrates Single Source Shortest Path(SSSP) computation using nvGRAPH Library.
  • Added 7_CUDALibraries/simpleCUBLASXT. Demonstrates simple example to use CUBLAS-XT library.
  • Added 6_Advanced/c++11_cuda. Demonstrates C++11 feature support in CUDA.
  • Added 1_Utilities/topologyQuery. Demonstrates how to query the topology of a system with multiple GPU.
  • Added 0_Simple/fp16ScalarProduct. Demonstrates scalar product calculation of two vectors of FP16 numbers.
  • Added 0_Simple/systemWideAtomics. Demonstrates system wide atomic instructions on migratable memory.
  • Removed 0_Simple/template_runtime. Its purpose is served by 0_Simple/template.

1.17. CUDA 7.5

  • Added 7_CUDALibraries/cuSolverDn_LinearSolver. Demonstrates how to use the CUSOLVER library for performing dense matrix factorization using cuSolverDN's LU, QR and Cholesky factorization functions.
  • Added 7_CUDALibraries/cuSolverRf. Demonstrates how to use cuSolverRF, a sparse re-factorization package of the CUSOLVER library.
  • Added 7_CUDALibraries/cuSolverSp_LinearSolver. Demonstrates how to use cuSolverSP which provides sparse set of routines for sparse matrix factorization.
  • The 2_Graphics/simpleD3D9, 2_Graphics/simpleD3D9Texture, 3_Imaging/cudaDecodeD3D9, and 5_Simulations/fluidsD3D9 samples have been modified to use the Direct3D 9Ex API instead of the Direct3D 9 API.
  • The 7_CUDALibraries/grabcutNPP and 7_CUDALibraries/imageSegmentationNPP samples have been removed. These samples used the NPP graphcut APIs, which have been deprecated in CUDA 7.5.

1.18. CUDA 7.0

  • Removed support for Windows 32-bit builds.
  • The Makefile x86_64=1 and ARMv7=1 options have been deprecated. Please use TARGET_ARCH to set the targeted build architecture instead.
  • The Makefile GCC option has been deprecated. Please use HOST_COMPILER to set the host compiler instead.
  • The CUDA Samples are no longer shipped as prebuilt binaries on Windows. Please use VS Solution files provided to build respective executable.
  • Added 0_Simple/clock_nvrtc. Demonstrates how to compile clock function kernel at runtime using libNVRTC to measure the performance of kernel accurately.
  • Added 0_Simple/inlinePTX_nvrtc. Demonstrates compilation of CUDA kernel having PTX embedded at runtime using libNVRTC.
  • Added 0_Simple/matrixMul_nvrtc. Demonstrates compilation of matrix multiplication CUDA kernel at runtime using libNVRTC.
  • Added 0_Simple/simpleAssert_nvrtc. Demonstrates compilation of CUDA kernel having assert() at runtime using libNVRTC.
  • Added 0_Simple/simpleAtomicIntrinsics_nvrtc. Demonstrates compilation of CUDA kernel performing atomic operations at runtime using libNVRTC.
  • Added 0_Simple/simpleTemplates_nvrtc. Demonstrates compilation of templatized dynamically allocated shared memory arrays CUDA kernel at runtime using libNVRTC.
  • Added 0_Simple/simpleVoteIntrinsics_nvrtc. Demonstrates compilation of CUDA kernel which uses vote intrinsics at runtime using libNVRTC.
  • Added 0_Simple/vectorAdd_nvrtc. Demonstrates compilation of CUDA kernel performing vector addition at runtime using libNVRTC.
  • Added 4_Finance/binomialOptions_nvrtc. Demonstrates runtime compilation using libNVRTC of CUDA kernel which evaluates fair call price for a given set of European options under binomial model.
  • Added 4_Finance/BlackScholes_nvrtc. Demonstrates runtime compilation using libNVRTC of CUDA kernel which evaluates fair call and put prices for a given set of European options by Black-Scholes formula.
  • Added 4_Finance/quasirandomGenerator_nvrtc. Demonstrates runtime compilation using libNVRTC of CUDA kernel which implements Niederreiter Quasirandom Sequence Generator and Inverse Cumulative Normal Distribution functions for the generation of Standard Normal Distributions.

1.19. CUDA 6.5

  • Added 7_CUDALibraries/cuHook. Demonstrates how to build and use an intercept library with CUDA.
  • Added 7_CUDALibraries/simpleCUFFT_callback. Demonstrates how to compute a 1D-convolution of a signal with a filter using a user-supplied CUFFT callback routine, rather than a separate kernel call.
  • Added 7_CUDALibraries/simpleCUFFT_MGPU. Demonstrates how to compute a 1D-convolution of a signal with a filter by transforming both into frequency domain, multiplying them together, and transforming the signal back to time domain on Multiple GPUs.
  • Added 7_CUDALibraries/simpleCUFFT_2d_MGPU. Demonstrates how to compute a 2D-convolution of a signal with a filter by transforming both into frequency domain, multiplying them together, and transforming the signal back to time domain on Multiple GPUs.
  • Removed 3_Imaging/cudaEncode. Support for the CUDA Video Encoder (NVCUVENC) has been removed.
  • Removed 4_Finance/ExcelCUDA2007. The topic will be covered in a blog post at Parallel Forall.
  • Removed 4_Finance/ExcelCUDA2010. The topic will be covered in a blog post at Parallel Forall.
  • The 4_Finance/binomialOptions sample is now restricted to running on GPUs with SM architecture 2.0 or greater.
  • The 4_Finance/quasirandomGenerator sample is now restricted to running on GPUs with SM architecture 2.0 or greater.
  • The 7_CUDALibraries/boxFilterNPP sample now demonstrates how to use the static NPP libraries on Linux and Mac.
  • The 7_CUDALibraries/conjugateGradient sample now demonstrates how to use the static CUBLAS and CUSPARSE libraries on Linux and Mac.
  • The 7_CUDALibraries/MersenneTwisterGP11213 sample now demonstrates how to use the static CURAND library on Linux and Mac.

1.20. CUDA 6.0

  • New featured samples that support a new CUDA 6.0 feature called UVM-Lite
  • Added 0_Simple/UnifiedMemoryStreams - new CUDA sample that demonstrates the use of OpenMP and CUDA streams with Unified Memory on a single GPU.
  • Added 1_Utilities/p2pBandwidthTestLatency - new CUDA sample that demonstrates how measure latency between pairs of GPUs with P2P enabled and P2P disabled.
  • Added 6_Advanced/StreamPriorities - This sample demonstrates basic use of the new CUDA 6.0 feature stream priorities.
  • Added 7_CUDALibraries/ConjugateGradientUM - This sample implements a conjugate gradient solver on GPU using cuBLAS and cuSPARSE library, using Unified Memory.

1.21. CUDA 5.5

  • Linux makefiles have been updated to generate code for the AMRv7 architecture. Only the ARM hard-float floating point ABI is supported. Both native ARMv7 compilation and cross compilation from x86 is supported
  • Performance improvements in CUDA toolkit for Kepler GPUs (SM 3.0 and SM 3.5)
  • Makefiles projects have been updated to properly find search default paths for OpenGL, CUDA, MPI, and OpenMP libraries for all OS Platforms (Mac, Linux x86, Linux ARM).
  • Linux and Mac project Makefiles now invoke NVCC for building and linking projects.
  • Added 0_Simple/cppOverload - new CUDA sample that demonstrates how to use C++ overloading with CUDA.
  • Added 6_Advanced/cdpBezierTessellation - new CUDA sample that demonstrates an advanced method of implementing Bezier Line Tessellation using CUDA Dynamic Parallelism. Requires compute capability 3.5 or higher.
  • Added 7_CUDALibrariess/jpegNPP - new CUDA sample that demonstrates how to use NPP for JPEG compression on the GPU.
  • CUDA Samples now have better integration with Nsight Eclipse IDE.
  • 6_Advanced/ptxjit sample now includes a new API to demonstrate PTX linking at the driver level.

1.22. CUDA 5.0

  • New directory structure for CUDA samples. Samples are classified accordingly to categories: 0_Simple, 1_Utilities, 2_Graphics, 3_Imaging, 4_Finance, 5_Simulations, 6_Advanced, and 7_CUDALibraries
  • Added 0_Simple/simpleIPC - CUDA Runtime API sample is a very basic sample that demonstrates Inter Process Communication with one process per GPU for computation. Requires Compute Capability 2.0 or higher and a Linux Operating System.
  • Added 0_Simple/simpleSeparateCompilation - demonstrates a CUDA 5.0 feature, the ability to create a GPU device static library and use it within another CUDA kernel. This example demonstrates how to pass in a GPU device function (from the GPU device static library) as a function pointer to be called. Requires Compute Capability 2.0 or higher.
  • Added 2_Graphics/bindlessTexture - demonstrates use of cudaSurfaceObject, cudaTextureObject, and MipMap support in CUDA. Requires Compute Capability 3.0 or higher.
  • Added 3_Imaging/stereoDisparity - demonstrates how to compute a stereo disparity map using SIMD SAD (Sum of Absolute Difference) intrinsics. Requires Compute Capability 2.0 or higher.
  • Added 0_Simple/cdpSimpleQuicksort - demonstrates a simple quicksort implemented using CUDA Dynamic Parallelism. This sample requires devices with compute capability 3.5 or higher.
  • Added 0_Simple/cdpSimplePrint - demonstrates simple printf implemented using CUDA Dynamic Parallelism. This sample requires devices with compute capability 3.5 or higher.
  • Added 6_Advanced/cdpLUDecomposition - demonstrates LU Decomposition implemented using CUDA Dynamic Parallelism. This sample requires devices with compute capability 3.5 or higher.
  • Added 6_Advanced/cdpAdvancedQuicksort - demonstrates an advanced quicksort implemented using CUDA Dynamic Parallelism. This sample requires devices with compute capability 3.5 or higher.
  • Added 6_Advanced/cdpQuadtree - demonstrates Quad Trees implemented using CUDA Dynamic Parallelism. This sample requires devices with compute capability 3.5 or higher.
  • Added 7_CUDALibraries/simpleDevLibCUBLAS - implements a simple cuBLAS function calls that call GPU device API library running cuBLAS functions. cuBLAS device code functions take advantage of CUDA Dynamic Parallelism and requires compute capability of 3.5 or higher.

1.23. CUDA 4.2

  • Added segmentationTreeThrust - demonstrates a method to build image segmentation trees using Thrust. This algorithm is based on Boruvka's MST algorithm.

1.24. CUDA 4.1

  • Added MersenneTwisterGP11213 - implements Mersenne Twister GP11213, a pseudorandom number generator using the cuRAND library.
  • Added HSOpticalFlow - When working with image sequences or video it's often useful to have information about objects movement. Optical flow describes apparent motion of objects in image sequence. This sample is a Horn-Schunck method for optical flow written using CUDA.
  • Added volumeFiltering - demonstrates basic volume rendering and filtering using 3D textures.
  • Added simpleCubeMapTexture - demonstrates how to use texcubemap fetch instruction in a CUDA C program.
  • Added simpleAssert - demonstrates how to use GPU assert in a CUDA C program.
  • Added grabcutNPP - CUDA implementation of Rother et al. GrabCut approach using the 8 neighborhood NPP Graphcut primitive introduced in CUDA 4.1. (C. Rother, V. Kolmogorov, A. Blake. GrabCut: Interactive Foreground Extraction Using Iterated Graph Cuts. ACM Transactions on Graphics (SIGGRAPH'04), 2004).

2. Getting Started

The CUDA Samples are an educational resource provided to teach CUDA programming concepts. The CUDA Samples are not meant to be used for performance measurements.

For system requirements and installation instructions, please refer to the Linux Installation Guide and the Windows Installation Guide.

2.1. Getting CUDA Samples

Windows

On Windows, the CUDA Samples are installed using the CUDA Toolkit Windows Installer. By default, the CUDA Samples are installed in: 
C:\ProgramData\NVIDIA Corporation\CUDA Samples\v11.6\
The installation location can be changed at installation time.

Linux

On Linux, to install the CUDA Samples, the CUDA toolkit must first be installed. See the Linux Installation Guide for more information on how to install the CUDA Toolkit.

Then the CUDA Samples can be installed by running the following command, where <target_path> is the location where to install the samples: 
$ cuda-install-samples-11.6.sh <target_path>

2.2. Building Samples

Windows

The Windows samples are built using the Visual Studio IDE. Solution files (.sln) are provided for each supported version of Visual Studio, using the format: 
*_vs<version>.sln - for Visual Studio <version>
Complete samples solution files exist at:
C:\ProgramData\NVIDIA Corporation\CUDA Samples\v11.6\
Each individual sample has its own set of solution files at:
C:\ProgramData\NVIDIA Corporation\CUDA Samples\v11.6\<sample_dir>\
To build/examine all the samples at once, the complete solution files should be used. To build/examine a single sample, the individual sample solution files should be used.

Linux

The Linux samples are built using makefiles. To use the makefiles, change the current directory to the sample directory you wish to build, and run make: 
$ cd <sample_dir>
$ make
The samples makefiles can take advantage of certain options:

  • TARGET_ARCH=<arch> - cross-compile targeting a specific architecture. Allowed architectures are x86_64, armv7l, aarch64, sbsa, and ppc64le.

    By default, TARGET_ARCH is set to HOST_ARCH. On a x86_64 machine, not setting TARGET_ARCH is the equvalent of setting TARGET_ARCH=x86_64.

    $ make TARGET_ARCH=x86_64
$ make TARGET_ARCH=armv7l
$ make TARGET_ARCH=aarch64
$ make TARGET_ARCH=sbsa
$ make TARGET_ARCH=ppc64le
    See here for more details.
  • dbg=1 - build with debug symbols 
    $ make dbg=1
  • SMS="A B ..." - override the SM architectures for which the sample will be built, where "A B ..." is a space-delimited list of SM architectures. For example, to generate SASS for SM 35 and SM 50, use SMS="35 50". 
    $ make SMS="35 50"
  • HOST_COMPILER=<host_compiler> - override the default g++ host compiler. See the Linux Installation Guide for a list of supported host compilers. 
    $ make HOST_COMPILER=g++

2.3. CUDA Cross-Platform Samples

CUDA Samples are now located in https://github.com/nvidia/cuda-samples, which includes instructions for obtaining, building, and running the samples.

2.4. Using CUDA Samples to Create Your Own CUDA Projects

2.4.1. Creating CUDA Projects for Windows

Creating a new CUDA Program using the CUDA Samples infrastructure is easy. We have provided a template project that you can copy and modify to suit your needs. Just follow these steps:

(<category> refers to one of the following folders: 0_Simple, 1_Utilities, 2_Graphics, 3_Imaging, 4_Finance, 5_Simulations, 6_Advanced, 7_CUDALibraries.)

  1. Copy the content of:
    C:\ProgramData\NVIDIA Corporation\CUDA Samples\v11.6\<category>\template
    to a directory of your own: 
    C:\ProgramData\NVIDIA Corporation\CUDA Samples\v11.6\<category>\myproject
  2. Edit the filenames of the project to suit your needs.
  3. Edit the *.sln, *.vcproj and source files. Just search and replace all occurrences of template with myproject.
  4. Build the 64-bit, release or debug configurations using:
    • myproject_vs<version>.sln
  5. Run myproject.exe from the release or debug directories located in:
    C:\ProgramData\NVIDIA Corporation\CUDA Samples\v11.6\bin\win64\[release|debug]
  6. Now modify the code to perform the computation you require. See the CUDA Programming Guide for details of programming in CUDA.

2.4.2. Creating CUDA Projects for Linux

Note: The default installation folder <SAMPLES_INSTALL_PATH> is NVIDIA_CUDA_11.6_Samples and <category> is one of the following: 0_Simple, 1_Utilities, 2_Graphics, 3_Imaging, 4_Finance, 5_Simulations, 6_Advanced, 7_CUDALibraries.
Creating a new CUDA Program using the NVIDIA CUDA Samples infrastructure is easy. We have provided a template project that you can copy and modify to suit your needs. Just follow these steps:
  1. Copy the template project:
    cd <SAMPLES_INSTALL_PATH>/<category>
cp -r template <myproject>
cd <SAMPLES_INSTALL_PATH>/<category>
  2. Edit the filenames of the project to suit your needs:
    mv template.cu myproject.cu
mv template_cpu.cpp myproject_cpu.cpp
  3. Edit the Makefile and source files. Just search and replace all occurrences of template with myproject.
  4. Build the project as (release):
    make
    To build the project as (debug), use "make dbg=1": 
    make dbg=1
  5. Run the program:
    ../../bin/x86_64/linux/release/myproject
  6. Now modify the code to perform the computation you require. See the CUDA Programming Guide for details of programming in CUDA.

3. Samples Reference

This document contains a complete listing of the code samples that are included with the NVIDIA CUDA Toolkit. It describes each code sample, lists the minimum GPU specification, and provides links to the source code and white papers if available.

The code samples are divided into the following categories:
Introduction Reference
Basic CUDA samples for beginners that illustrate key concepts with using CUDA and CUDA runtime APIs.
Utilities Reference
Utility samples that demonstrate how to query device capabilities and measure GPU/CPU bandwidth.
Concepts and Techniques Reference
Samples that demonstrate CUDA related concepts and common problem solving techniques.
CUDA Features Reference
Samples that demonstrate CUDA Features.
CUDA Libraries Reference
Samples that demonstrate how to use CUDA platform libraries (NPP, NVJPEG, NVGRAPH cuBLAS, cuFFT, cuSPARSE, cuSOLVER and cuRAND).
Domain Specific Reference
Samples that are specific to domain (Graphics, Finance, Image Processing).
Performance Reference
Samples that demonstrate performance optimization.

4. Dependencies

Some CUDA Samples rely on third-party applications and/or libraries, or features provided by the CUDA Toolkit and Driver, to either build or execute. These dependencies are listed below.

If a sample has a dependency that is not available on the system, the sample will not be installed. If a sample has a third-party dependency that is available on the system, but is not installed, the sample will waive itself at build time.

Each sample's dependencies are listed in the Samples Reference section.

Third-Party Dependencies

These third-party dependencies are required by some CUDA samples. If available, these dependencies are either installed on your system automatically, or are installable via your system's package manager (Linux) or a third-party website.

FreeImage

FreeImage is an open source imaging library. FreeImage can usually be installed on Linux using your distribution's package manager system. FreeImage can also be downloaded from the FreeImage website.

To set up FreeImage on a Windows system, extract the FreeImage DLL distribution into the 7_CUDALibraries/common/ folder such that 7_CUDALibraries/common/FreeImage/Dist/x64/ contains the .h, .dll, and .lib files.

Message Passing Interface

MPI (Message Passing Interface) is an API for communicating data between distributed processes. A MPI compiler can be installed using your Linux distribution's package manager system. It is also available on some online resources, such as Open MPI. On Windows, to build and run MPI-CUDA applications one can install MS-MPI SDK.

Only 64-Bit

Some samples can only be run on a 64-bit operating system.

DirectX

DirectX is a collection of APIs designed to allow development of multimedia applications on Microsoft platforms. For Microsoft platforms, NVIDIA's CUDA Driver supports DirectX. Several CUDA Samples for Windows demonstrates CUDA-DirectX Interoperability, for building such samples one needs to install Microsoft Visual Studio 2012 or higher which provides Microsoft Windows SDK for Windows 8.

DirectX 12

DirectX 12 is a collection of advanced low-level programming APIs which can reduce driver overhead, designed to allow development of multimedia applications on Microsoft platforms starting with Windows 10 OS onwards. For Microsoft platforms, NVIDIA's CUDA Driver supports DirectX. Few CUDA Samples for Windows demonstrates CUDA-DirectX12 Interoperability, for building such samples one needs to install Windows 10 SDK or higher , with VS 2015 or VS 2017.

OpenGL

OpenGL is a graphics library used for 2D and 3D rendering. On systems which support OpenGL, NVIDIA's OpenGL implementation is provided with the CUDA Driver.

OpenGL ES

OpenGL ES is an embedded systems graphics library used for 2D and 3D rendering. On systems which support OpenGL ES, NVIDIA's OpenGL ES implementation is provided with the CUDA Driver.

Vulkan

Vulkan is a low-overhead, cross-platform 3D graphics and compute API. Vulkan targets high-performance realtime 3D graphics applications such as video games and interactive media across all platforms. On systems which support Vulkan, NVIDIA's Vulkan implementation is provided with the CUDA Driver. For building and running Vulkan applications one needs to install the Vulkan SDK.

OpenMP

OpenMP is an API for multiprocessing programming. OpenMP can be installed using your Linux distribution's package manager system. It usually comes preinstalled with GCC. It can also be found at the OpenMP website.

Screen

Screen is a windowing system found on the QNX operating system. Screen is usually found as part of the root filesystem.

X11

X11 is a windowing system commonly found on *-nix style operating systems. X11 can be installed using your Linux distribution's package manager, and comes preinstalled on Mac OS X systems.

EGL

EGL is an interface between Khronos rendering APIs (such as OpenGL, OpenGL ES or OpenVG) and the underlying native platform windowing system.

EGLOutput

EGLOutput is a set of EGL extensions which allow EGL to render directly to the display.

EGLSync

EGLSync is a set of EGL extensions which provides sync objects that are synchronization primitive, representing events whose completion can be tested or waited upon.

NVSCI

NvSci is a set of communication interface libraries out of which CUDA interops with NvSciBuf and NvSciSync. NvSciBuf allows applications to allocate and exchange buffers in memory. NvSciSync allows applications to manage synchronization objects which coordinate when sequences of operations begin and end.

NvMedia

NvMedia provides powerful processing of multimedia data for true hardware acceleration across NVIDIA Tegra devices. Applications leverage the NvMedia Application Programming Interface (API) to process the image and video data.

CUDA Features

These CUDA features are needed by some CUDA samples. They are provided by either the CUDA Toolkit or CUDA Driver. Some features may not be available on your system.

CUFFT Callback Routines

CUFFT Callback Routines are user-supplied kernel routines that CUFFT will call when loading or storing data. These callback routines are only available on Linux x86_64 and ppc64le systems.

CUDA Dynamic Paralellism

CDP (CUDA Dynamic Paralellism) allows kernels to be launched from threads running on the GPU. CDP is only available on GPUs with SM architecture of 3.5 or above.

Multi-block Cooperative Groups

Multi Block Cooperative Groups(MBCG) extends Cooperative Groups and the CUDA programming model to express inter-thread-block synchronization. MBCG is available on GPUs with Pascal and higher architecture.

Multi-Device Cooperative Groups

Multi Device Cooperative Groups extends Cooperative Groups and the CUDA programming model enabling thread blocks executing on multiple GPUs to cooperate and synchronize as they execute. This feature is available on GPUs with Pascal and higher architecture.

CUBLAS

CUBLAS (CUDA Basic Linear Algebra Subroutines) is a GPU-accelerated version of the BLAS library.

CUDA Interprocess Communication

IPC (Interprocess Communication) allows processes to share device pointers.

CUFFT

CUFFT (CUDA Fast Fourier Transform) is a GPU-accelerated FFT library.

CURAND

CURAND (CUDA Random Number Generation) is a GPU-accelerated RNG library.

CUSPARSE

CUSPARSE (CUDA Sparse Matrix) provides linear algebra subroutines used for sparse matrix calculations.

CUSOLVER

CUSOLVER library is a high-level package based on the CUBLAS and CUSPARSE libraries. It combines three separate libraries under a single umbrella, each of which can be used independently or in concert with other toolkit libraries. The intent ofCUSOLVER is to provide useful LAPACK-like features, such as common matrix factorization and triangular solve routines for dense matrices, a sparse least-squares solver and an eigenvalue solver. In addition cuSolver provides a new refactorization library useful for solving sequences of matrices with a shared sparsity pattern.

NPP

NPP (NVIDIA Performance Primitives) provides GPU-accelerated image, video, and signal processing functions.

NVJPEG

NVJPEG library provides high-performance, GPU accelerated JPEG decoding functionality for image formats commonly used in deep learning and hyperscale multimedia applications.

NVRTC

NVRTC (CUDA RunTime Compilation) is a runtime compilation library for CUDA C++.

Stream Priorities

Stream Priorities allows the creation of streams with specified priorities. Stream Priorities is only available on GPUs with SM architecture of 3.5 or above.

Unified Virtual Memory

UVM (Unified Virtual Memory) enables memory that can be accessed by both the CPU and GPU without explicit copying between the two. UVM is only available on Linux and Windows systems.

16-bit Floating Point

FP16 is a 16-bit floating-point format. One bit is used for the sign, five bits for the exponent, and ten bits for the mantissa. FP16 is only available on specific mobile platforms.

C++11 CUDA

NVCC Support of C++11 features.

7. Frequently Asked Questions

Answers to frequently asked questions about CUDA can be found at http://developer.nvidia.com/cuda-faq and in the CUDA Toolkit Release Notes.

Notices

Notice

This document is provided for information purposes only and shall not be regarded as a warranty of a certain functionality, condition, or quality of a product. NVIDIA Corporation (“NVIDIA”) makes no representations or warranties, expressed or implied, as to the accuracy or completeness of the information contained in this document and assumes no responsibility for any errors contained herein. NVIDIA shall have no liability for the consequences or use of such information or for any infringement of patents or other rights of third parties that may result from its use. This document is not a commitment to develop, release, or deliver any Material (defined below), code, or functionality.

NVIDIA reserves the right to make corrections, modifications, enhancements, improvements, and any other changes to this document, at any time without notice.

Customer should obtain the latest relevant information before placing orders and should verify that such information is current and complete.

NVIDIA products are sold subject to the NVIDIA standard terms and conditions of sale supplied at the time of order acknowledgement, unless otherwise agreed in an individual sales agreement signed by authorized representatives of NVIDIA and customer (“Terms of Sale”). NVIDIA hereby expressly objects to applying any customer general terms and conditions with regards to the purchase of the NVIDIA product referenced in this document. No contractual obligations are formed either directly or indirectly by this document.

NVIDIA products are not designed, authorized, or warranted to be suitable for use in medical, military, aircraft, space, or life support equipment, nor in applications where failure or malfunction of the NVIDIA product can reasonably be expected to result in personal injury, death, or property or environmental damage. NVIDIA accepts no liability for inclusion and/or use of NVIDIA products in such equipment or applications and therefore such inclusion and/or use is at customer’s own risk.

NVIDIA makes no representation or warranty that products based on this document will be suitable for any specified use. Testing of all parameters of each product is not necessarily performed by NVIDIA. It is customer’s sole responsibility to evaluate and determine the applicability of any information contained in this document, ensure the product is suitable and fit for the application planned by customer, and perform the necessary testing for the application in order to avoid a default of the application or the product. Weaknesses in customer’s product designs may affect the quality and reliability of the NVIDIA product and may result in additional or different conditions and/or requirements beyond those contained in this document. NVIDIA accepts no liability related to any default, damage, costs, or problem which may be based on or attributable to: (i) the use of the NVIDIA product in any manner that is contrary to this document or (ii) customer product designs.

No license, either expressed or implied, is granted under any NVIDIA patent right, copyright, or other NVIDIA intellectual property right under this document. Information published by NVIDIA regarding third-party products or services does not constitute a license from NVIDIA to use such products or services or a warranty or endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual property rights of the third party, or a license from NVIDIA under the patents or other intellectual property rights of NVIDIA.

Reproduction of information in this document is permissible only if approved in advance by NVIDIA in writing, reproduced without alteration and in full compliance with all applicable export laws and regulations, and accompanied by all associated conditions, limitations, and notices.

THIS DOCUMENT AND ALL NVIDIA DESIGN SPECIFICATIONS, REFERENCE BOARDS, FILES, DRAWINGS, DIAGNOSTICS, LISTS, AND OTHER DOCUMENTS (TOGETHER AND SEPARATELY, “MATERIALS”) ARE BEING PROVIDED “AS IS.” NVIDIA MAKES NO WARRANTIES, EXPRESSED, IMPLIED, STATUTORY, OR OTHERWISE WITH RESPECT TO THE MATERIALS, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE. TO THE EXTENT NOT PROHIBITED BY LAW, IN NO EVENT WILL NVIDIA BE LIABLE FOR ANY DAMAGES, INCLUDING WITHOUT LIMITATION ANY DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE, OR CONSEQUENTIAL DAMAGES, HOWEVER CAUSED AND REGARDLESS OF THE THEORY OF LIABILITY, ARISING OUT OF ANY USE OF THIS DOCUMENT, EVEN IF NVIDIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Notwithstanding any damages that customer might incur for any reason whatsoever, NVIDIA’s aggregate and cumulative liability towards customer for the products described herein shall be limited in accordance with the Terms of Sale for the product.

OpenCL

OpenCL is a trademark of Apple Inc. used under license to the Khronos Group Inc.

Trademarks

NVIDIA and the NVIDIA logo are trademarks or registered trademarks of NVIDIA Corporation in the U.S. and other countries. Other company and product names may be trademarks of the respective companies with which they are associated.