Download NVIDIA CUDA Toolkit
Version 7.0
What’s new in CUDA Toolkit 7.0?
Support for the latest GPUs, including GeForce GTX TITAN X (Maxwell), GeForce GTX TITAN X (Pascal), Tesla P100, and Quadro GP100
Unified Memory programming with C++11 atomics and CUDA streams APIs
Support for new features in the Pascal architecture
Linux for Tegra R23.2 support
CUDA Fortran support with new capabilities, including coarrays, Pointer Attribution and additional derived type support
Support for application profiling on 64-bit Android devices using adb shell commands and ARM CPU top. See instructions here.
NVIDIA CUDA Toolkit is a C/C++ development environment for CUDA-enabled GPUs, library and driver for building GPU-accelerated applications.
The NVIDIA CUDA Toolkit includes:
CUDA-GDB: The NVIDIA Visual Profiler is a cross-platform performance profiling tool that delivers developers vital feedback for optimizing CUDA C/C++ applications. The Visual Profiler supports all 350 million+ CUDA capable NVIDIA GPUs shipped since 2006 on Linux, Mac OS X, and Windows.
The NVIDIA CUDA Toolkit provides a development environment for creating high performance GPU-accelerated applications. With the CUDA Toolkit, you can develop, optimize and deploy your applications on GPU-accelerated embedded systems, desktop workstations, enterprise data centers, cloud-based platforms and HPC supercomputers.
The toolkit includes GPU-accelerated libraries, debugging and optimization tools, a C/C++ compiler and a runtime library to deploy your application.
What’s New in This Release
The following are some of the new features introduced in CUDA Toolkit 10.1:
Support for the latest GPUs including Volta and Turing architectures**
Inclusion of cuDNN 7.6**
Improvements to cuBLAS performance on Volta GPUs**
Improvements to NCCL performance for all supported networks**
Addition of ncclReduceScatter support for large amounts of data**
New algorithms for nvGRAPH**
NVIDIA Container Toolkit support (beta)**
See the CUDA Toolkit 10.1 Release Notes for a complete list of features.
The CUDA Toolkit from NVIDIA provides everything you need to develop GPU-accelerated applications. The toolkit includes GPU-accelerated libraries, debugging and optimization tools, a C/C++ compiler, and a runtime library to deploy your application.
CUDA 9 is compatible with Ubuntu 17.04 and 16.04, CentOS 6 and 7, Red Hat Enterprise Linux 6 and 7, Windows 10 (x86_64), Windows Server 2008 R2 and later (x86_64), Windows 7 (x86_64) SP1 and later, RHEL 5.11 (x86_64), macOS 10.10 and later, Solaris 11 SPARC and x86_64
The NVIDIA CUDA Toolkit provides a development environment for creating high performance GPU-accelerated applications. With the CUDA Toolkit, you can develop, optimize and deploy your applications on GPU-accelerated embedded systems, desktop workstations, enterprise data centers, cloud-based platforms and HPC supercomputers.
The toolkit includes GPU-accelerated libraries, debugging and optimization tools, a C/C++ compiler and a runtime library to deploy your application.
The NVIDIA® CUDA® Toolkit provides a development environment for creating high performance GPU-accelerated applications. With the CUDA Toolkit, you can develop, optimize and deploy your applications on GPU-accelerated embedded systems, desktop workstations, enterprise data centers, cloud-based platforms and HPC supercomputers.
The toolkit includes GPU-accelerated libraries, debugging and optimization tools, a C/C++ compiler and a runtime library to deploy your application. In addition to supporting our latest flagship GPUs, including the TITAN Xp**, the NVIDIA GPU Cloud provides simple access to a comprehensive software stack that includes deep learning software such as Caffe2 & TensorFlow.
Read this article on the NVIDIA Developer Blog to learn more about the new releases of these exciting deep learning frameworks from Facebook AI Research (FAIR) and NVIDIA.