Powerful and reliable programming model and computing toolkit

NVIDIA CUDA Toolkit

NVIDIA CUDA Toolkit 11.5.0 (for Windows 10)

  -  2.4 GB  -  Freeware

Sometimes latest versions of the software can cause issues when installed on older devices or devices running an older version of the operating system.

Software makers usually fix these issues but it can take them some time. What you can do in the meantime is to download and install an older version of NVIDIA CUDA Toolkit 11.5.0 (for Windows 10).


For those interested in downloading the most recent release of NVIDIA CUDA Toolkit or reading our review, simply click here.


All old versions distributed on our website are completely virus-free and available for download at no cost.


We would love to hear from you

If you have any questions or ideas that you want to share with us - head over to our Contact page and let us know. We value your feedback!

What's new in this version:

- Device-side caching behavior is now configurable with annotated pointers
- Prefix sums (scans) for cooperative groups: Added four new functions for inclusive and exclusive scans
- Support for NvSciBufGeneralAttrKey_EnableGpuCompression in CUDA will be available in r470TRD2
- Preview release of a new data type, __int128, usable with compatible host compilers. As it is a preview, there is no broad support for math operations, library support, dev tools, and so on
- Added native support for signed and unsigned normalized 8- and 16-bit types
- Improved interoperability with graphics frameworks: Added support for normalized integer and block-compressed data types
- For multi-process sharing of GPUs, CUDA now supports per-process memory access policies
- Linking is supported with cubins larger than 2 GB
- GSP-RM is enabled as opt-in for Turing+ Tesla GPUs
- Floating point division is optimized when the divisor is known at compile time. This is disabled by default; enable with nvcc -Xcicc -opt-fdiv=1
- GA release of CUDA Python