Software Development Tools

• Easy, self-paced video and audio tutorials and webinars
• Full complement of CUDA documentation including Fermi tuning guides
• "Programming Massively Parallel Processors" by David Kirk, NVIDIA and Dr. Wen-mei Hwu, University of Illinois

• Start with the CUDA Forums
• Register as a CUDA Developer
• File bugs, get early access to software releases

• Drivers for Tesla GPUs
• CUDA C and C++ Compiler for GPUs
• Includes CUDA GDB Debugger and CUDA Visual Profiler
• OpenCL Drivers, Visual Profiler, SDK Code Samples and more
• PGI CUDA Fortran Compiler
• Parallel NSight Visual Studio IDE for GPU code development (formerly called "Nexus")
• ArrayFire GPU library for rapid development in C, C++, Fortran, and Python

• NVIDIA provided
• CUDA SDK Code samples
• CUDA BLAS library (part of toolkit)
• CUDA FFT library (part of toolkit)
• Sparse Matrix Multiply Vector from NVIDIA: Code, Paper 1, Paper 2
• NPP: NVIDIA Performance Primitives early access: Focuses on image and video processing
• Note: CUDA 4.1 RC is now available to CUDA Registered Developers.
• 3rd party provided
• CUDA data parallel primitives library (cuDPP)
• CULA Tools: LAPACK on CUDA GPUs from EM Photonics
• MAGMA: LAPACK on CUDA GPUs and Multi-core CPUs from Dongarra's Group
• Jacobi-preconditioned Conjugate Gradient
• GPULib: Library of Mathematical functions for IDL and MATLAB®
• GPU VSIPL signal processing library
• Computer Vision and imaging library
• OpenCurrent: Open-Source library of CUDA-Accelerated PDE (partial differential equation) Solvers Over Regular Grids
• libSVM on CUDA / GPUs
• Multisvm: Multiclass SVM on CUDA
• cuSVM: A CUDA Implementation of Support Vector Classification and Regression
• ArrayFire: Comprehensive GPU library with functions for math, matrix indexing, image processing, signal processing, statistics, and many more

• PGI Accelerator: Fortran and C to CUDA C Auto-parallelizing Compiler
• CAPS HMPP C and Fortran to CUDA C Auto-parallelizing Compiler
• Goose: C to CUDA C Auto-parallelizing Compiler
• NOAA F2C: Fortran to CUDA C Translator

• MATLAB
• MathWorks: MATLAB GPU Computing with NVIDIA GPUs
• ArrayFire GPU function library for C, C++, FORTRAN
• GPULib: Library of mathematical functions with MATLAB and IDL bindings
• Mathematica
• CUDA Programming with Wolfram Mathematica
• Mathematica Plugin for CUDA
• Enabling GPU Computing in the R Statistical Environment
• CUDA GPU Library for LabVIEW from National Instruments
• User Generated Way of using CUDA from LabVIEW CVI

• Fortran
• PGI CUDA Fortran compiler
• FLAGON: Fortran 95 library for GPU Numerics
• Python wrapper for CUDA: PyCUDA
• Java wrapper
• jCUDA: Java for CUDA
• Bindings for CUDA BLAS and FFT
• JaCUDA
• .NET integration for CUDA
• Thrust: C++ template Library for CUDA
• CuPP : C++ framework for CUDA
• Libra: C/C++ abstraction layer for CUDA
• F# for CUDA
• ArrayFire GPU library for accelerating C++, Fortran, and Python

• Eclipse support for CUDA
• CUDA Occupancy Calculator
• Rocks Rolls for CUDA from ClusterCorp
• Platform Computing Cluster Manager for Tesla GPUs
• PBS Works for Tesla GPUs
• Penguin Computing Scyld

MATLAB is a registered trademark of The MathWorks, Inc.
ArrayFire is a trademark of AccelerEyes.