CUDA 6.5 Production Release

Read about 10 ways CUDA 6.5 improves performance and productivity in this blog by Mark Harris.

Review the latest CUDA 6.5 performance report to learn how much you could accelerate your code.

NOTE: GeForce GTX980 or GTX970 developers, please download CUDA 6.5 Toolkits with support for your GPU here.
NVIDIA Nsight™ Visual Studio Edition 4.2 is now available for download in the NVIDIA GameWorks Developer Program.
Version 64-bit 32-bit
Windows 8.1
Windows 7
Notebook EXE
Desktop EXE
Win Server 2012 R2 Desktop EXE  
Win Server 2008 R2 Desktop EXE
Windows XP Desktop   EXE
Windows Getting Started Guide  


If you are developing with CUDA 6.5 on a GeForce GTX980 or GTX970 download CUDA 6.5 Toolkits with support for your GPU here.
Distribution x86 64-bit x86 32-bit
Fedora 20 RPM     RUN  
OpenSUSE 13.1 RPM     RUN  
CentOS 6
RPM     RUN  
CentOS 5
SLES 11 (SP3) RPM     RUN  
SteamOS 1.0-beta**   RUN  
Ubuntu 14.04** DEB*    RUN   RUN
Ubuntu 12.04** DEB*    RUN   RUN
Linux Getting Started Guide  

* Includes ARMv7 cross-compilation tools.
** Includes x86 32-bit cross-compilation and profiling tools.

  ARMv8 64-bit*** ARMv7 32-bit
Generic Display Driver RUN  
Generic CUDA Toolkit RUN  
Generic CUDA Samples RUN  
Ubuntu 14.04   DEB*
Ubuntu 14.04 Cross-Compile (x86 64-bit)   DEB   
Ubuntu 12.04 Cross-Compile (x86 64-bit)   DEB   
L4T 21.2   DEB**
Linux Getting Started Guide  

* NOT to be used for L4T (Jetson TK1).
** ONLY to be used for Jetson TK1.
*** ARM-64 support has been tested on systems from Cirrascale and E4 based on the AppliedMicro X-Gene-C1 processor.

If you are developing with CUDA 6.5 on a GeForce GTX980 or GTX970 download CUDA 6.5 Toolkits with support for your GPU here.
OSX Release Package
10.8 PKG

Mac Getting Started Guide

Q: Can I get a version of the cuSOLVER library that works with the CUDA 6.5 toolkit?
A: We have made an unsupported patch available which provides a version of the cuSOLVER library that will work with the CUDA 6.5 Toolkit. The patch is only available for x86_64 systems running Linux. Also, cuSOLVER is only officially supported as of the CUDA 7.0 Toolkit, so the patch available here is provided without any additional documentation, and is meant as an aid for advanced developers that don’t have access to CUDA 7.0 systems yet.

Q: Are the latest NVIDIA drivers included in the CUDA Toolkit installers?
A: For convenience, the installer packages on this page include NVIDIA drivers which support application development for all CUDA-capable GPUs supported by this release of the CUDA Toolkit. If you are deploying applications on NVIDIA Tesla products in a server or cluster environment, please use the latest recommended Tesla driver that has been qualified for use with this version of the CUDA Toolkit. If a recommended Tesla driver is not yet available, please check back in a few weeks.

Q: How does this release differ from the current CUDA 6.5 Release?
A: These toolkits contain support for the GeForce GTX980 and GTX970, including compatible drivers. In addition, driver support for older generation GPUs with SM1.x has been deprecated.

Q: What's in the installer packages?
A: The installers include the CUDA Toolkit, CUDA samples, CUDA Tools including Nsight Visual Studio Edition (for Windows) and Nsight Eclipse Edition (for Linux / Mac OS X), and developer drivers.

Q: Will the installer replace the driver currently installed on my system?
A: The installer will provide an option to install the included driver, and if selected, it will replace the driver currently on your system.

Q: Why doesn't the cuda-repo package install the CUDA Toolkit and Drivers?
A: When using RPM or Deb, the downloaded package is a repository package. Such a package only informs the package manager where to find the actual installation packages, but will not install them.

Q: How do I report issues or bugs?
A: Members of the CUDA Registered Developer Program can report issues and file bugs using links on the Report Issues section of the CUDA Registered Developer Home Page.

Q: Where can I find old versions of the CUDA Toolkit?
A: Older versions of the toolkit can be found on the Legacy CUDA Toolkits page.