Release Notes :: CUDA Toolkit Documentation

NVIDIA CUDA Toolkit Release Notes

The Release Notes for the CUDA Toolkit.

1. CUDA 11.2 Release Notes

The release notes for the CUDA Toolkit can be found online at http://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html.

1.1. CUDA Toolkit Major Component Versions

CUDA Components

Starting with CUDA 11, the various components in the toolkit are versioned independently.

For CUDA 11.2, the table below indicates the versions:

Table 1. CUDA 11.2 Component Versions
Component Name	Version Information	Supported Architectures
CUDA Runtime (cudart)	11.2.72	x86_64, POWER, Arm64
cuobjdump	11.2.67	x86_64, POWER, Arm64
CUPTI	11.2.67	x86_64, POWER, Arm64
CUDA cuxxfilt (demangler)	11.2.67	x86_64, POWER, Arm64
CUDA Demo Suite	11.2.67	x86_64
CUDA GDB	11.2.67	x86_64, POWER, Arm64
CUDA Memcheck	11.2.67	x86_64, POWER
CUDA NVCC	11.2.67	x86_64, POWER, Arm64
CUDA nvdisasm	11.2.67	x86_64, POWER, Arm64
CUDA NVML Headers	11.2.67	x86_64, POWER, Arm64
CUDA nvprof	11.2.67	x86_64, POWER, Arm64
CUDA nvprune	11.2.67	x86_64, POWER, Arm64
CUDA NVRTC	11.2.67	x86_64, POWER, Arm64
CUDA NVTX	11.2.67	x86_64, POWER, Arm64
CUDA NVVP	11.2.67	x86_64, POWER
CUDA Samples	11.2.67	x86_64, POWER, Arm64
CUDA Compute Sanitizer API	11.2.67	x86_64, POWER, Arm64
CUDA cuBLAS	11.3.1.68	x86_64, POWER, Arm64
CUDA cuFFT	10.4.0.72	x86_64, POWER, Arm64
CUDA cuRAND	10.2.3.68	x86_64, POWER, Arm64
CUDA cuSOLVER	11.0.2.68	x86_64, POWER, Arm64
CUDA cuSPARSE	11.3.1.68	x86_64, POWER, Arm64
CUDA NPP	11.2.1.68	x86_64, POWER, Arm64
CUDA nvJPEG	11.3.1.68	x86_64, POWER, Arm64
Nsight Eclipse Plugins	11.2.67	x86_64, POWER
Nsight Compute	2020.3.0.18	x86_64, POWER, Arm64
Nsight Windows NVTX	1.21018621	x86_64, POWER, Arm64
Nsight Systems	2020.4.3.7	x86_64, POWER, Arm64
Nsight Visual Studio Edition (VSE)	2020.3.0.20315	x86_64 (Windows)
NVIDIA Linux Driver	460.27.04	x86_64, POWER, Arm64
NVIDIA Windows Driver	460.89	x86_64 (Windows)

CUDA Driver

Running a CUDA application requires the system with at least one CUDA capable GPU and a driver that is compatible with the CUDA Toolkit. See Table 2. For more information various GPU products that are CUDA capable, visit https://developer.nvidia.com/cuda-gpus.

Each release of the CUDA Toolkit requires a minimum version of the CUDA driver. The CUDA driver is backward compatible, meaning that applications compiled against a particular version of the CUDA will continue to work on subsequent (later) driver releases.

More information on compatibility can be found at https://docs.nvidia.com/cuda/cuda-c-best-practices-guide/index.html#cuda-runtime-and-driver-api-version.

Note: Starting with CUDA 11.0, the toolkit components are individually versioned, and the toolkit itself is versioned as shown in the table below.

Table 2. CUDA Toolkit and Compatible Driver Versions
CUDA Toolkit	Linux x86_64 Driver Version	Windows x86_64 Driver Version
CUDA 11.2.0 GA	>=460.27.04	>=460.89
CUDA 11.1.1 Update 1	>=455.32	>=456.81
CUDA 11.1 GA	>=455.23	>=456.38
CUDA 11.0.3 Update 1	>= 450.51.06	>= 451.82
CUDA 11.0.2 GA	>= 450.51.05	>= 451.48
CUDA 11.0.1 RC	>= 450.36.06	>= 451.22
CUDA 10.2.89	>= 440.33	>= 441.22
CUDA 10.1 (10.1.105 general release, and updates)	>= 418.39	>= 418.96
CUDA 10.0.130	>= 410.48	>= 411.31
CUDA 9.2 (9.2.148 Update 1)	>= 396.37	>= 398.26
CUDA 9.2 (9.2.88)	>= 396.26	>= 397.44
CUDA 9.1 (9.1.85)	>= 390.46	>= 391.29
CUDA 9.0 (9.0.76)	>= 384.81	>= 385.54
CUDA 8.0 (8.0.61 GA2)	>= 375.26	>= 376.51
CUDA 8.0 (8.0.44)	>= 367.48	>= 369.30
CUDA 7.5 (7.5.16)	>= 352.31	>= 353.66
CUDA 7.0 (7.0.28)	>= 346.46	>= 347.62

For convenience, the NVIDIA driver is installed as part of the CUDA Toolkit installation. Note that this driver is for development purposes and is not recommended for use in production with Tesla GPUs.

For running CUDA applications in production with Tesla GPUs, it is recommended to download the latest driver for Tesla GPUs from the NVIDIA driver downloads site at http://www.nvidia.com/drivers.

During the installation of the CUDA Toolkit, the installation of the NVIDIA driver may be skipped on Windows (when using the interactive or silent installation) or on Linux (by using meta packages).

For more information on customizing the install process on Windows, see http://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html#install-cuda-software.

For meta packages on Linux, see https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#package-manager-metas

1.2. General CUDA

Added support for RHEL 7.9, RHEL 8.3, Fedora 33 and Debian 10.6 Buster on x86_64 platforms.
Stream ordered memory allocator: Added new APIs cudaMallocAsync() and cudaFreeAsync() to enable applications to order memory allocation and deallocation with other work launched into a CUDA stream. Provides significant performance improvements compared to cudaMalloc()/cudaFree(). Introduces the concept of memory pools to provide the application with more control over memory management. Each device has a default memory pool and custom memory pools can be created as needed.
Added support for importing DirectX11/12 textures with format DXGI_FORMAT_NV12 via the CUDA external resource interoperability APIs. For more details about external resource interoperability API functions, see https://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__EXTRES__INTEROP.html#group__CUDART__EXTRES__INTEROP.
Added new Driver and Runtime API functions, cuArrayGetPlane and cudaArrayGetPlane respectively, to get individual format plane CUDA arrays from multi-planar formatted CUDA arrays.
Two new API functions have been added to get the list of architectures supported by the NVRTC library:
- nvrtcGetNumSupportedArchs
- nvrtcGetSupportedArchs
CUDA Graph enhancements:
- Two new graph node types: external semaphore signal and external semaphore wait. This allows new types of synchronization between graph workloads and non-CUDA workloads. Enables support for explicitly adding these nodes to the graph as well as adding nodes by capturing calls to cuSignalExternalSemaphoresAsync(), cudaSignalExternalSemaphoresAsync(), cuWaitExternalSemaphoresAsync(), and cudaWaitExternalSemaphoresAsync().
  
  New APIs:
  - cudaGraphAddExternalSemaphoresSignalNode()
  - cudaGraphExternalSemaphoresSignalNodeGetParams()
  - cudaGraphExternalSemaphoresSignalNodeSetParams()
  - cudaGraphAddExternalSemaphoresWaitNode()
  - cudaGraphExternalSemaphoresWaitNodeGetParams()
  - cudaGraphExternalSemaphoresWaitNodeSetParams()
  - cudaGraphExecExternalSemaphoresSignalNodeSetParams()
  - cudaGraphExecExternalSemaphoresWaitNodeSetParams()
- Support for updating the function of an instantiated kernel node via explicit node update using cudaGraphExecKernelNodeSetParams or whole graph update cudaGraphExecUpdate APIs.
- Improved device-side performance for the first launch of a graph following instantiation or updates.
Cooperative groups enhancement: cudaLaunchCooperativeKernel now enables simultaneous launch in multiple streams allowing multiple co-operative grids to run concurrently.
Extended the external semaphore interface to work with Vulkan Timeline Semaphores on Linux and Windows platforms. Enables support to import, wait, and signal on Vulkan Timeline Semaphores exported from Vulkan via cuImportExternalSemaphore(), cudaImportExternalSemaphore(), cuSignalExternalSemaphoresAsync(), cudaSignalExternalSemaphoresAsync(), cuWaitExternalSemaphoresAsync(), and cudaWaitExternalSemaphoresAsync(). On Linux, semaphores are imported from an FD exported from a Vulkan semaphore. On Windows, semaphores are imported via an NT handle exported from a Vulkan semaphore.
CUDA Runtime: Protoype for cudaGetFuncBySymbol changed to add cdecl qualifier.

1.3. CUDA Tools

1.3.1. CUDA Compilers

nvcc has added an option --optimization-info=inline. Using this option, diagnostics are emitted about inlining decisions made by the compiler for device code. For functions not inlined, additional information about the reason for not inlining are emitted. The compiler may invoke inlining pass multiple times and a callsite not inlined in an earlier pass may be inlined in a future pass.
Added nvcc options:
- --display-error-number to display error number for warnings
- --diag-error errNum,... to override the severity of errEnum,... to errors
- --diag-suppress errNum,... to suppress warnings errEnum,...
- --diag-warn errNum,... to override the severity of errEnum,... to warning
nvcc now provides support for the following builtin functions, for providing optimization hints to the compiler:
- __builtin_assume()
- __assume()
- __builtin_assume_aligned()
In addition, pre-existing support for __builtin_expect() is now documented. See Compiler Optimization Hint Functions for details.

1.3.2. CUDA Tools

A preview version of a new tool, cu++filt, is included in this release. NVCC produces mangled names, appearing in PTX files, which do not strictly follow the mangling conventions of the Itanium ABI--and are thus not properly demangled by standard tools such as binutils' c++filt. Specifically, this is true for PTX function parameters. The new cu++filt utility will demangle all of these correctly. As this is a preview version of the utility, feedback is invited. For more information, see cu++filt.

1.3.3. CUDA Developer Tools

For changes to nvprof and Visual Profiler, see the changelog.
For new features, improvements, and bug fixes in CUPTI, see the changelog.
For new features, improvements, and bug fixes in Nsight Compute, see the changelog.

1.4. CUDA Libraries

1.4.1. cuFFT Library

Multi-GPU plans can be associated with a stream using the cufftSetStream API function call.
Performance improvements for R2C/C2C/C2R transforms.
Performance improvements for multi-GPU systems.

1.5. Deprecated or Removed Features

The following features are deprecated or removed in the current release of the CUDA software. Deprecated features still work in the current release, but their documentation may have been removed, and they will become officially unsupported in a future release. We recommend that developers employ alternative solutions to these features in their software.

General CUDA

Support for Visual Studio versions <= 2015 has been dropped in CUDA 11.2, following its deprecation in CUDA 11.1.

1.6. Resolved Issues

1.6.1. cuFFT Library

cuFFT is no longer stuck in a bad state if previous plan creation fails with CUFFT_ALLOC_FAILED.
Previously, single dimensional multi-GPU FFT plans ignored user input on cufftXtSetGPUswhichGPUs argument and assumed that GPUs IDs are always numbered from 0 to N-1. This issue has been resolved.
Plans with primes larger than 127 in FFT size decomposition or FFT size being a prime number bigger than 4093 do not perform calculations on second and subsequent cufftExecute* calls. Regression was introduced in cuFFT 11.1

1.6.2. cuSOLVER Library

cusolverDnIRSXgels sometimes returned CUSOLVER_STATUS_INTERNAL_ERROR when the precision is ‘z’. This issue has been fixed in CUDA 11.2; now cusolverDnIRSXgels works for all precisions.
ZSYTRF sometimes returned CUSOLVER_STATUS_INTERNAL_ERROR due to insufficient resources to launch the kernel. This issue has been fixed in CUDA 11.2
GETRF returned early without finishing the whole factorization when the matrix was singular. This issue has been fixed in CUDA 11.2

1.7. Known Issues

1.7.1. General CUDA

Upgrading the driver on SUSE or SLES may result in a failure to load the NVIDIA kernel modules. This occurs specifically when the G04 packages are upgraded to the G05 packages. To remedy this, run the following to force a reinstall of the KMP (kernel module package):
```
$ sudo zypper in --force `rpm -qa "nvidia-gfx*kmp*"
```

1.7.2. CUDA Compiler

Windows, when using recent versions of VS 2019 host compiler, a call to pow(double, int) or pow(float, int) in host or device code may cause build failure. This is an NVCC issue. For example:
```
//--
int main() {
  double x = pow(1.0, 1);
  return (int)x;
}
//--
```
The above program may fail to build with a linker error in CUDA 11.2, when using recent versions of VS 2019 host compiler. A source workaround is to ensure that both arguments are of double type, e.g. pow(1.0, (double)1).

1.7.3. cuBLAS Library

cublas<s/d/c/z>Gemm() with very large n and m=k=1 may fail on Pascal devices.

1.7.4. cuFFT Library

Plans with strides, primes larger than 127 in FFT size decomposition and total size of transform including strides bigger than 32GB produce incorrect results.
Reduced performance of power-of-2 single precision FFTs on GPUs with sm_86 architecture.
cuFFT planning and plan estimation functions may not restore correct context affecting CUDA driver API applications.

1.7.5. cuSPARSE Library

cusparseXdense2csr provides incorrect results for some matrix sizes.

Notices

Notice

This document is provided for information purposes only and shall not be regarded as a warranty of a certain functionality, condition, or quality of a product. NVIDIA Corporation (“NVIDIA”) makes no representations or warranties, expressed or implied, as to the accuracy or completeness of the information contained in this document and assumes no responsibility for any errors contained herein. NVIDIA shall have no liability for the consequences or use of such information or for any infringement of patents or other rights of third parties that may result from its use. This document is not a commitment to develop, release, or deliver any Material (defined below), code, or functionality.

NVIDIA reserves the right to make corrections, modifications, enhancements, improvements, and any other changes to this document, at any time without notice.

Customer should obtain the latest relevant information before placing orders and should verify that such information is current and complete.

NVIDIA products are sold subject to the NVIDIA standard terms and conditions of sale supplied at the time of order acknowledgement, unless otherwise agreed in an individual sales agreement signed by authorized representatives of NVIDIA and customer (“Terms of Sale”). NVIDIA hereby expressly objects to applying any customer general terms and conditions with regards to the purchase of the NVIDIA product referenced in this document. No contractual obligations are formed either directly or indirectly by this document.

NVIDIA products are not designed, authorized, or warranted to be suitable for use in medical, military, aircraft, space, or life support equipment, nor in applications where failure or malfunction of the NVIDIA product can reasonably be expected to result in personal injury, death, or property or environmental damage. NVIDIA accepts no liability for inclusion and/or use of NVIDIA products in such equipment or applications and therefore such inclusion and/or use is at customer’s own risk.

NVIDIA makes no representation or warranty that products based on this document will be suitable for any specified use. Testing of all parameters of each product is not necessarily performed by NVIDIA. It is customer’s sole responsibility to evaluate and determine the applicability of any information contained in this document, ensure the product is suitable and fit for the application planned by customer, and perform the necessary testing for the application in order to avoid a default of the application or the product. Weaknesses in customer’s product designs may affect the quality and reliability of the NVIDIA product and may result in additional or different conditions and/or requirements beyond those contained in this document. NVIDIA accepts no liability related to any default, damage, costs, or problem which may be based on or attributable to: (i) the use of the NVIDIA product in any manner that is contrary to this document or (ii) customer product designs.

No license, either expressed or implied, is granted under any NVIDIA patent right, copyright, or other NVIDIA intellectual property right under this document. Information published by NVIDIA regarding third-party products or services does not constitute a license from NVIDIA to use such products or services or a warranty or endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual property rights of the third party, or a license from NVIDIA under the patents or other intellectual property rights of NVIDIA.

Reproduction of information in this document is permissible only if approved in advance by NVIDIA in writing, reproduced without alteration and in full compliance with all applicable export laws and regulations, and accompanied by all associated conditions, limitations, and notices.

THIS DOCUMENT AND ALL NVIDIA DESIGN SPECIFICATIONS, REFERENCE BOARDS, FILES, DRAWINGS, DIAGNOSTICS, LISTS, AND OTHER DOCUMENTS (TOGETHER AND SEPARATELY, “MATERIALS”) ARE BEING PROVIDED “AS IS.” NVIDIA MAKES NO WARRANTIES, EXPRESSED, IMPLIED, STATUTORY, OR OTHERWISE WITH RESPECT TO THE MATERIALS, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE. TO THE EXTENT NOT PROHIBITED BY LAW, IN NO EVENT WILL NVIDIA BE LIABLE FOR ANY DAMAGES, INCLUDING WITHOUT LIMITATION ANY DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE, OR CONSEQUENTIAL DAMAGES, HOWEVER CAUSED AND REGARDLESS OF THE THEORY OF LIABILITY, ARISING OUT OF ANY USE OF THIS DOCUMENT, EVEN IF NVIDIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Notwithstanding any damages that customer might incur for any reason whatsoever, NVIDIA’s aggregate and cumulative liability towards customer for the products described herein shall be limited in accordance with the Terms of Sale for the product.

VESA DisplayPort

DisplayPort and DisplayPort Compliance Logo, DisplayPort Compliance Logo for Dual-mode Sources, and DisplayPort Compliance Logo for Active Cables are trademarks owned by the Video Electronics Standards Association in the United States and other countries.

HDMI

HDMI, the HDMI logo, and High-Definition Multimedia Interface are trademarks or registered trademarks of HDMI Licensing LLC.

OpenCL

OpenCL is a trademark of Apple Inc. used under license to the Khronos Group Inc.

Trademarks

NVIDIA and the NVIDIA logo are trademarks or registered trademarks of NVIDIA Corporation in the U.S. and other countries. Other company and product names may be trademarks of the respective companies with which they are associated.

Copyright

This product includes software developed by the Syncro Soft SRL (http://www.sync.ro/).