CUDA

CUDA is a parallel computing platform and application programming interface (API) model created by Nvidia.^[1] It allows software developers and software engineers to use a CUDA-enabled graphics processing unit (GPU) for general purpose processing – an approach termed GPGPU (General-Purpose computing on Graphics Processing Units). The CUDA platform is a software layer that gives direct access to the GPU's virtual instruction set and parallel computational elements, for the execution of compute kernels.^[2]

The CUDA platform is designed to work with programming languages such as C, C++, and Fortran. This accessibility makes it easier for specialists in parallel programming to use GPU resources, in contrast to prior APIs like Direct3D and OpenGL, which required advanced skills in graphics programming. Also, CUDA supports programming frameworks such as OpenACC and OpenCL.^[2] When it was first introduced by Nvidia, the name CUDA was an acronym for Compute Unified Device Architecture,^[3] but Nvidia subsequently dropped the use of the acronym.

Background

The graphics processing unit (GPU), as a specialized computer processor, addresses the demands of real-time high-resolution 3D graphics compute-intensive tasks. By 2012, GPUs had evolved into highly parallel multi-core systems allowing very efficient manipulation of large blocks of data. This design is more effective than general-purpose central processing unit (CPUs) for algorithms in situations where processing large blocks of data is done in parallel, such as:

• push-relabel maximum flow algorithm
• fast sort algorithms of large lists
• two-dimensional fast wavelet transform
• molecular dynamics simulations

Programming abilities

Example of CUDA processing flow
1. Copy data from main mem to GPU mem
2. CPU instructs the process to GPU
3. GPU execute parallel in each core
4. Copy the result from GPU mem to main mem

The CUDA platform is accessible to software developers through CUDA-accelerated libraries, compiler directives such as OpenACC, and extensions to industry-standard programming languages including C, C++ and Fortran. C/C++ programmers use 'CUDA C/C++', compiled with nvcc, Nvidia's LLVM-based C/C++ compiler.^[4] Fortran programmers can use 'CUDA Fortran', compiled with the PGI CUDA Fortran compiler from The Portland Group.

In addition to libraries, compiler directives, CUDA C/C++ and CUDA Fortran, the CUDA platform supports other computational interfaces, including the Khronos Group's OpenCL,^[5] Microsoft's DirectCompute, OpenGL Compute Shaders and C++ AMP.^[6] Third party wrappers are also available for Python, Perl, Fortran, Java, Ruby, Lua, Haskell, R, MATLAB, IDL, and native support in Mathematica.

In the computer game industry, GPUs are used for graphics rendering, and for game physics calculations (physical effects such as debris, smoke, fire, fluids); examples include PhysX and Bullet. CUDA has also been used to accelerate non-graphical applications in computational biology, cryptography and other fields by an order of magnitude or more.^{[7][8][9][10][11]}

CUDA provides both a low level API and a higher level API. The initial CUDA SDK was made public on 15 February 2007, for Microsoft Windows and Linux. Mac OS X support was later added in version 2.0,^[12] which supersedes the beta released February 14, 2008.^[13] CUDA works with all Nvidia GPUs from the G8x series onwards, including GeForce, Quadro and the Tesla line. CUDA is compatible with most standard operating systems. Nvidia states that programs developed for the G8x series will also work without modification on all future Nvidia video cards, due to binary compatibility.

CUDA 8.0 comes with the following libraries (for compilation & runtime, in alphabetical order):

• CUBLAS - CUDA Basic Linear Algebra Subroutines library, see main and docs
• CUDART - CUDA RunTime library, see docs
• CUFFT - CUDA Fast Fourier Transform library, see main and docs
• CURAND - CUDA Random Number Generation library, see main and docs
• CUSOLVER - CUDA based collection of dense and sparse direct solvers, see main and docs
• CUSPARSE - CUDA Sparse Matrix library, see main and docs
• NPP - NVIDIA Performance Primitives library, see main and docs
• NVGRAPH - NVIDIA Graph Analytics library, see main and docs
• NVML - NVIDIA Management Library, see main and docs
• NVRTC - NVRTC RunTime Compilation library for CUDA C++, see docs

CUDA 8.0 comes with these other software components:

• nView - NVIDIA nView Desktop Management Software, see main and docs (pdf)
• NVWMI - NVIDIA Enterprise Management Toolkit, see main and docs (chm)
• PhysX - GameWorks PhysX is a scalable multi-platform game physics solution, see main and docs

Advantages

CUDA has several advantages over traditional general-purpose computation on GPUs (GPGPU) using graphics APIs:

• Scattered reads – code can read from arbitrary addresses in memory
• Unified virtual memory (CUDA 4.0 and above)
• Unified memory (CUDA 6.0 and above)
• Shared memory – CUDA exposes a fast shared memory region that can be shared among threads. This can be used as a user-managed cache, enabling higher bandwidth than is possible using texture lookups.^[14]
• Faster downloads and readbacks to and from the GPU
• Full support for integer and bitwise operations, including integer texture lookups

Limitations

• CUDA does not support the full C standard, as it runs host code through a C++ compiler, which makes some valid C (but invalid C++) code fail to compile.^[15][16]
• Interoperability with rendering languages such as OpenGL is one-way, with OpenGL having access to registered CUDA memory but CUDA not having access to OpenGL memory.
• Copying between host and device memory may incur a performance hit due to system bus bandwidth and latency (this can be partly alleviated with asynchronous memory transfers, handled by the GPU's DMA engine)
• Threads should be running in groups of at least 32 for best performance, with total number of threads numbering in the thousands. Branches in the program code do not affect performance significantly, provided that each of 32 threads takes the same execution path; the SIMD execution model becomes a significant limitation for any inherently divergent task (e.g. traversing a space partitioning data structure during ray tracing).
• Unlike OpenCL, CUDA-enabled GPUs are only available from Nvidia.^[17]
• No emulator or fallback functionality is available for modern revisions.
• Valid C/C++ may sometimes be flagged and prevent compilation due to optimization techniques the compiler is required to employ to use limited resources.
• A single process must run spread across multiple disjoint memory spaces, unlike other C language runtime environments.
• C++ run-time type information (RTTI) is unsupported in CUDA code, due to lack of support in the underlying hardware.
• Exception handling is unsupported in CUDA code due to performance overhead that would be incurred with many thousands of parallel threads running.
• CUDA (with compute ability 2.x) allows a subset of C++ class functionality, for example member functions may not be virtual (this restriction will be removed in some future release). [See CUDA C Programming Guide 3.1 – Appendix D.6]
• In single precision on first generation CUDA compute ability 1.x devices, denormal numbers are unsupported and are instead flushed to zero, and the precisions of the division and square root operations are slightly lower than IEEE 754-compliant single precision math. Devices that support compute ability 2.0 and above support denormal numbers, and the division and square root operations are IEEE 754 compliant by default. However, users can obtain the prior faster gaming-grade math of compute ability 1.x devices if desired by setting compiler flags to disable accurate divisions and accurate square roots, and enable flushing denormal numbers to zero.^[18]

GPUs supported

Supported CUDA Level of GPU and Card. See direct also Nvidia:

• CUDA SDK 6.5: Last Version with support for Tesla with Compute Capability 1.x
• CUDA SDK 7.5 support for Compute Capability 2.0 – 5.x (Fermi, Kepler, Maxwell)
• CUDA SDK 8.0 support for Compute Capability 2.0 – 6.x (Fermi, Kepler, Maxwell, Pascal)
• Next CUDA SDK Version no support for Fermi (2.x)

'*' – OEM-only products

Version features and specifications

Note: Any missing lines or empty entries do reflect some lack of information on that exact item.

For more information please visit this site: http://www.geeks3d.com/20100606/gpu-computing-nvidia-cuda-compute-capability-comparative-table/ and read Nvidia CUDA programming guide.^[22]

Example

This example code in C++ loads a texture from an image into an array on the GPU:

texture<float, 2, cudaReadModeElementType> tex;

void foo()
{
  cudaArray* cu_array;

  // Allocate array
  cudaChannelFormatDesc description = cudaCreateChannelDesc<float>();
  cudaMallocArray(&cu_array, &description, width, height);

  // Copy image data to array
  cudaMemcpyToArray(cu_array, image, width*height*sizeof(float), cudaMemcpyHostToDevice);

  // Set texture parameters (default)
  tex.addressMode[0] = cudaAddressModeClamp;
  tex.addressMode[1] = cudaAddressModeClamp;
  tex.filterMode = cudaFilterModePoint;
  tex.normalized = false; // do not normalize coordinates

  // Bind the array to the texture
  cudaBindTextureToArray(tex, cu_array);

  // Run kernel
  dim3 blockDim(16, 16, 1);
  dim3 gridDim((width + blockDim.x - 1)/ blockDim.x, (height + blockDim.y - 1) / blockDim.y, 1);
  kernel<<< gridDim, blockDim, 0 >>>(d_data, height, width);

  // Unbind the array from the texture
  cudaUnbindTexture(tex);
} //end foo()

__global__ void kernel(float* odata, int height, int width)
{
   unsigned int x = blockIdx.x*blockDim.x + threadIdx.x;
   unsigned int y = blockIdx.y*blockDim.y + threadIdx.y;
   if (x < width && y < height) {
      float c = tex2D(tex, x, y);
      odata[y*width+x] = c;
   }
}

Below is an example given in Python that computes the product of two arrays on the GPU. The unofficial Python language bindings can be obtained from PyCUDA.^[23]

import pycuda.compiler as comp
import pycuda.driver as drv
import numpy
import pycuda.autoinit

mod = comp.SourceModule("""
__global__ void multiply_them(float *dest, float *a, float *b)
{
  const int i = threadIdx.x;
  dest[i] = a[i] * b[i];
}
""")

multiply_them = mod.get_function("multiply_them")

a = numpy.random.randn(400).astype(numpy.float32)
b = numpy.random.randn(400).astype(numpy.float32)

dest = numpy.zeros_like(a)
multiply_them(
        drv.Out(dest), drv.In(a), drv.In(b),
        block=(400,1,1))

print dest-a*b

Additional Python bindings to simplify matrix multiplication operations can be found in the program pycublas.^[24]

import numpy
from pycublas import CUBLASMatrix
A = CUBLASMatrix( numpy.mat([[1,2,3]],[[4,5,6]],numpy.float32) )
B = CUBLASMatrix( numpy.mat([[2,3]],[4,5],[[6,7]],numpy.float32) )
C = A*B
print C.np_mat()

Language bindings

• Common Lisp – cl-cuda
• Fortran – FORTRAN CUDA, PGI CUDA Fortran Compiler
• F# – Alea.CUDA
• Haskell – Data.Array.Accelerate
• IDL – GPULib
• Java – jCUDA, JCuda, JCublas, JCufft, CUDA4J
• Lua – KappaCUDA
• Mathematica – CUDALink
• MATLAB – Parallel Computing Toolbox, MATLAB Distributed Computing Server,^[25] and 3rd party packages like Jacket.
• .NET – CUDA.NET, Managed CUDA, CUDAfy.NET .NET kernel and host code, CURAND, CUBLAS, CUFFT
• Perl – KappaCUDA, CUDA::Minimal
• Python – Numba, NumbaPro, PyCUDA, KappaCUDA, Theano
• Ruby – KappaCUDA
• R – gputools

Current and future usages of CUDA architecture

• Accelerated rendering of 3D graphics
• Accelerated interconversion of video file formats
• Accelerated encryption, decryption and compression
• Bioinformatics, e.g. NGS DNA sequencing BarraCUDA
• Distributed calculations, such as predicting the native conformation of proteins
• Medical analysis simulations, for example virtual reality based on CT and MRI scan images.
• Physical simulations, in particular in fluid dynamics.
• Neural network training in machine learning problems
• Face recognition
• Distributed computing
• Molecular dynamics
• Mining cryptocurrencies

See also

• OpenCL – A standard for programming a variety of platforms, including GPUs
• BrookGPU – the Stanford University graphics group's compiler
• Array programming
• Parallel computing
• Stream processing
• rCUDA – An API for computing on remote computers
• Molecular modeling on GPU
• Vulkan

References

External links

• Official website
• CUDA Community on Google+
• A little tool to adjust the VRAM size