CUDA Compatibility

v   |    
 Table of Contents
Chapter 1. Why CUDA Compatibility.................................................................................... 1
Chapter 2. Minor Version Compatibility...............................................................................4
2.1. CUDA 11 and Later Defaults to Minor Version Compatibility Compatibility...........................4
2.2. Application Considerations for Minor Version Compatibility.................................................. 6
2.3. Deployment Considerations for Minor Version Compatibility.................................................6
Chapter 3. Forward Compatibility........................................................................................ 8
3.1. Forward Compatibility Support Across Major Toolkit Versions..............................................8
3.2. Installing the Forward Compatibility Package........................................................................ 9
3.2.1. From Network Repositories or Local Installers...............................................................9
3.2.2. Manually Installing from Runfile..................................................................................... 11
3.3. Deployment Considerations for Forward Compatibility........................................................ 11
3.3.1. Use the Right Compat Package...................................................................................... 11
3.3.2. Feature Exceptions...........................................................................................................12
3.3.3. Check for Compatibility Support..................................................................................... 13
3.4. Deployment Model for Forward Compatibility.......................................................................13
Chapter 4. Conclusion........................................................................................................ 15
Chapter 5. Frequently Asked Questions............................................................................ 16

CUDA Compatibility v   |   ii

 List of Figures
Figure 1. Components of CUDA ...................................................................................................... 2

Figure 2. CUDA Upgrade Path ........................................................................................................ 3

Figure 3. NVRTC supports minor version compatibility from CUDA 11.3 onwards ........................ 7

Figure 4. Forward Compatibility Upgrade Path ................................................................................ 9

CUDA Compatibility v   |   iii

 List of Tables
Table 1. Example CUDA Toolkit 11.x Minimum Required Driver Versions (Refer to CUDA
Release Notes)................................................................................................................................. 4

Table 2. CUDA Toolkit 10.x Minimum Required Driver Versions ..................................................... 5

Table 3. CUDA Application Compatibility Support Matrix ...............................................................12

Table 4. Forward-Compatible Feature-Driver Support Matrix .......................................................12

CUDA Compatibility v   |   iv

Chapter 1. Why CUDA Compatibility

®
The CUDA Toolkit enables developers to build NVIDIA GPU accelerated compute applications
for Desktop computers, Enterprise and Data centers to Hyperscalers. It consists of the CUDA
compiler toolchain including the CUDA runtime (cudart) and various CUDA libraries and
tools. To build an application, a developer has to install only the CUDA Toolkit and necessary
libraries required for linking.
In order to run a CUDA application, the system should have a CUDA enabled GPU and an
NVIDIA display driver that is compatible with the CUDA Toolkit that was used to build the
application itself. If the application relies on dynamic linking for libraries then the system
should have the right version of such libraries as well.

CUDA Compatibility v   |   1

 Why CUDA Compatibility

Figure 1. Components of CUDA

Every CUDA toolkit also ships with an NVIDIA display driver package for convenience. This
driver supports all the features introduced in that version of the CUDA Toolkit. Please check
the toolkit and driver version mapping in the release notes. The driver package includes both
the user mode CUDA driver (libcuda.so) and kernel mode components necessary to run the
application.
Typically, upgrading a CUDA Toolkit involves upgrading both the toolkit and the driver to get
the bleeding edge toolkit and driver capabilities.

CUDA Compatibility v   |   2

 Why CUDA Compatibility

Figure 2. CUDA Upgrade Path

But this is not always required. CUDA Compatibility guarantees allow for upgrading only
certain components and that will be the focus of the rest of this document. We will see how
the upgrade to a new CUDA Toolkit can be simplified to not always require a full system
upgrade.

CUDA Compatibility v   |   3

Chapter 2. Minor Version Compatibility

2.1.  CUDA 11 and Later Defaults to Minor

Version Compatibility Compatibility
From CUDA 11 onwards, applications compiled with a CUDA Toolkit release from within a
CUDA major release family can run, with limited feature-set, on systems having at least the
minimum required driver version as indicated below. This minimum required driver can be
different from the driver packaged with the CUDA Toolkit but should belong to the same major
release.
Refer to the CUDA Toolkit Release Notes for the complete table.

Table 1. Example CUDA Toolkit 11.x Minimum Required Driver Versions

(Refer to CUDA Release Notes)
CUDA Toolkit Linux x86_64 Minimum Windows Minimum Required
Required Driver Version Driver Version
CUDA 11.x >= 450.80.02* >=452.39*

* CUDA 11.0 was released with an earlier driver version, but by upgrading to Tesla Recommended
Drivers 450.80.02 (Linux) / 452.39 (Windows) as indicated, minor version compatibility is possible
across the CUDA 11.x family of toolkits.
While applications built against any of the older CUDA Toolkits always continued to function
on newer drivers due to binary backward compatibility, prior to CUDA 11, applications built
against newer CUDA Toolkit releases were not supported on older drivers without forward
compatibility package (see Forward Compatibility Support Across Major Toolkit Versions).
If you are using a new CUDA 10.x minor release, then the minimum required driver version
is the same as the driver that’s packaged as part of that toolkit release. Consequently, the
minimum required driver version changed for every new CUDA Toolkit minor release until
CUDA 11.1. Therefore, system administrators always have to upgrade drivers in order to
support applications built against CUDA Toolkits from 10.x releases.

CUDA Compatibility v   |   4

 Minor Version Compatibility

Table 2. CUDA Toolkit 10.x Minimum Required Driver Versions

CUDA Toolkit Linux x86_64 Minimum Windows MinimumRequired
Required Driver Version Driver Version
CUDA 10.2 >= 440.33 >=441.22
CUDA 10.1 >= 418.39 >=418.96
CUDA 10.0 >= 410.48 >=411.31

With minor version compatibility, upgrading to CUDA 11.1 is now possible on older drivers
from within the same major release family such as 450.80.02 that was shipped with CUDA
11.0, as shown below:
$ nvidia-smi

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 450.80.02 Driver Version: 450.80.02 CUDA Version:
11.0 |
|-------------------------------+----------------------
+----------------------+

...<snip>...

$ samples/bin/x86_64/linux/release/deviceQuery
samples/bin/x86_64/linux/release/deviceQuery Starting...

CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: "Tesla T4"

CUDA Driver Version / Runtime Version 11.0 / 11.1
CUDA Capability Major/Minor version number: 7.5

...<snip>...

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 11.0, CUDA

Runtime Version = 11.1, NumDevs = 1
Result = PASS

Minimum required driver version guidance can be found in the release notes. Note that if the
minimum required driver version is not installed in the system, applications will return an
error as shown below.
$ samples/bin/x86_64/linux/release/deviceQuery
samples/bin/x86_64/linux/release/deviceQuery Starting...

CUDA Device Query (Runtime API) version (CUDART static linking)

cudaGetDeviceCount returned 3
-> initialization error
Result = FAIL

CUDA Compatibility v   |   5

 Minor Version Compatibility

2.2.  Application Considerations for Minor

Version Compatibility
Developers and system admins should note two important caveats when relying on minor
version compatibility. If either of these caveats are limiting, then a new CUDA driver from the
same minor version of the toolkit that the application was built with or later is required.

‣ Limited feature set

Sometimes features introduced in a CUDA Toolkit version may actually span both the
toolkit and the driver. In such cases an application that relies on features introduced in
a newer version of the toolkit and driver may return the following error on older drivers:
cudaErrorCallRequiresNewerDriver. As mentioned earlier, admins should then
upgrade the installed driver also.
Application developers can avoid running into this problem by having the application
explicitly check for the availability of features. Refer to the CUDA Compatibility Developers
Guide for more details.
‣ Applications using PTX will see runtime issues
Applications that compile device code to PTX will not work on older drivers. If the
application requires PTX then admins have to upgrade the installed driver.
PTX Developers should refer to the CUDA Compatibility Developers Guide and PTX
programming guide in the CUDA C++ Programming Guide for details on this limitation.

2.3.  Deployment Considerations for Minor

Version Compatibility
As described, applications that directly rely only on the CUDA runtime can be deployed in the
following two scenarios:
1. CUDA driver that’s installed on the system is newer than the runtime.
2. CUDA runtime is newer than the CUDA driver on the system but they are from the same
major release of CUDA Toolkit.
In scenario 2, system admins should be aware of the aforementioned limitations and should be
able to tell why an application may be failing if they run into any issues.
Minor version compatibility has another benefit that offers flexibility in the use and deployment
of libraries. Applications that use libraries that support minor version compatibility can be
deployed on systems with a different version of the toolkit and libraries without recompiling
the application for the difference in the library version. This holds true for both older and
newer versions of the libraries provided they are all from the same major release family. Note
that libraries themselves have interdependencies that should be considered. For example,
each cuDNN version requires a certain version of cuBLAS.

CUDA Compatibility v   |   6

 Minor Version Compatibility

Figure 3. NVRTC supports minor version compatibility from CUDA 11.3

onwards

However, if an application is unable to leverage the minor version compatibility due to any
of the aforementioned reasons, then the Forward Compatibility model can be used as an
alternative even though Forward Compatibility is mainly intended for compatibility across
major toolkit versions.

CUDA Compatibility v   |   7

Chapter 3. Forward Compatibility

3.1.  Forward Compatibility Support

Across Major Toolkit Versions
Increasingly, data centers and enterprises may not want to update the NVIDIA GPU Driver
across major release versions due to the rigorous testing and validation that happens before
any system level driver installations are done.
To support such scenarios, CUDA introduced a Forward Compatibility Upgrade path in CUDA
10.0.

CUDA Compatibility v   |   8

 Forward Compatibility

Figure 4. Forward Compatibility Upgrade Path

Forward Compatibility is applicable only for systems with NVIDIA Data Center GPUs or select
NGC Server Ready SKUs of RTX cards. It’s mainly intended to support applications built on
newer CUDA Toolkits to run on systems installed with an older NVIDIA Linux GPU driver from
different major release families. This new forward-compatible upgrade path requires the use
of a special package called “CUDA compat package”.

3.2.  Installing the Forward Compatibility

Package
3.2.1.  From Network Repositories or Local
Installers
The CUDA compat package is available in the local installers or the CUDA network
repositories provided by NVIDIA as cuda-compat-11.5.
Install the package on the system using the package installer.

CUDA Compatibility v   |   9

 Forward Compatibility

On Ubuntu, for example:

$ sudo apt-get install -y cuda-compat-11-5

The compat package will then be installed to the versioned toolkit location typically found in
the toolkit directory. For example, for 11.5 it will be found in /usr/local/cuda-11.5/.
The cuda-compat package consists of the following files:

‣ libcuda.so.* - the CUDA Driver

‣ libnvidia-nvvm.so.* - JIT LTO ( CUDA 11.5 and later only)

‣ libnvidia-ptxjitcompiler.so.* - the JIT (just-in-time) compiler for PTX files

These files should be kept together as the CUDA driver is dependent on the libnvidia-
ptxjitcompiler.so.* of the same version.

Note: This package only provides the files, and does not configure the system.

Example
CUDA Compatibility is installed and the application can now run successfully as shown below.
In this example, the user sets LD_LIBRARY_PATH to include the files installed by the cuda-
compat-11-5 package.
$ sudo apt-get install -y cuda-compat-11-5
Selecting previously unselected package cuda-compat-11-5.
(Reading database ... 220718 files and directories currently installed.)
Preparing to unpack .../cuda-compat-11-5_495.29.05-1_amd64.deb ...
Unpacking cuda-compat-11-5 (495.29.05-1) ...
Setting up cuda-compat-11-5 (495.29.05-1) ...
Processing triggers for libc-bin (2.31-0ubuntu9.2) ...

Check the files installed under /usr/local/cuda/compat:

$ ls -l /usr/local/cuda/compat
total 59456
lrwxrwxrwx 1 root root 12 Oct 1 11:07 libcuda.so -> libcuda.so.1
lrwxrwxrwx 1 root root 20 Oct 1 11:07 libcuda.so.1 -> libcuda.so.495.29.05
-rw-r--r-- 1 root root 22708136 Oct 1 00:16 libcuda.so.495.29.05
lrwxrwxrwx 1 root root 19 Oct 8 10:59 libnvidia-nvvm.so -> libnvidia-
nvvm.so.4
lrwxrwxrwx 1 root root 23 Oct 8 10:15 libnvidia-nvvm.so.4 -> libnvidia-
nvvm.so.4.0.0
-rw-r--r-- 1 root root 26807480 Oct 1 00:29 libnvidia-nvvm.so.4.0.0
lrwxrwxrwx 1 root root 37 Oct 1 11:07 libnvidia-ptxjitcompiler.so.1 ->
libnvidia-ptxjitcompiler.so.495.29.05
-rw-r--r-- 1 root root 11365720 Sep 30 23:55 libnvidia-ptxjitcompiler.so.495.29.05

The user can set LD_LIBRARY_PATH to include the files installed before running the CUDA 11.5
application:
$ LD_LIBRARY_PATH=/usr/local/cuda/compat:$LD_LIBRARY_PATH samples/bin/x86_64/linux/
release/deviceQuery
samples/bin/x86_64/linux/release/deviceQuery Starting...

CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: "Tesla V100-PCIE-32GB"

CUDA Driver Version / Runtime Version 11.5 / 11.5
CUDA Capability Major/Minor version number: 7.5
...<snip>...

CUDA Compatibility v   |   10

 Forward Compatibility

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 11.5, CUDA Runtime Version
= 11.5, NumDevs = 1
Result = PASS

3.2.2.  Manually Installing from Runfile

The cuda-compat package files can also be extracted from the appropriate datacenter driver
‘runfile’ installers (.run) available in NVIDIA driver downloads. To do this:
1. Download the latest NVIDIA Data Center GPU driver , and extract the .run file using option
-x.
2. Copy the three CUDA compatibility upgrade files, listed at the start of this section, into a
user- or root-created directory.
3. Follow your system’s guidelines for making sure that the system linker picks up the new
libraries.

Note: Symlinks under /usr/local/cuda/compat need to be created manually when using the
runfile installer.

3.3.  Deployment Considerations for

Forward Compatibility
3.3.1.  Use the Right Compat Package
CUDA forward compat packages should be used only in the following situations when forward
compatibility is required across major releases.
The CUDA compat package is named after the highest toolkit that it can support. If you are on
the R418 driver but require 11.2 application support, please install the cuda-compat package
for 11.2. But when performing a full system upgrade, when choosing to install both the toolkit
and the driver, remove any forward compatible packages present in the system.
For example, if you are upgrading the driver to 450.80.02 which is the minimum required driver
version for the 11.2 toolkit, then the cuda-compat package is not required in most cases.
10.x and 11.0 applications will be supported due to backward compatibility and future 11.x
applications will be supported due to minor-version compatibility.
But there are feature restrictions that may make this option less desirable for your scenario
- for example: Applications requiring PTX JIT compilation support. Unlike the minor-version
compatibility that is defined between CUDA runtime and CUDA driver, forward compatibility
is defined between the kernel driver and the CUDA driver, and hence such restrictions do not
apply. In order to circumvent the limitation, a forward compatibility package may be used in
such scenarios as well.

CUDA Compatibility v   |   11

 Forward Compatibility

Table 3. CUDA Application Compatibility Support Matrix

NVIDIA Kernel Mode Driver - Production Branch
CUDA 418.40.04+ 450.36.06+ 455.23.05+ 460.27.04+ 465.19.01+ 470.57.02+ 495.29.05+ 510.39.01+
Forward (CUDA (CUDA (CUDA (CUDA (CUDA (CUDA (CUDA (CUDA
Compatible 10.1) 11.0) 11.1) 11.2) 11.3) 11.4) 11.5) 11.6)
Upgrade
11-6 C C X* C X* C X* Not
required
11-5 C C C C X
11-4 C C C Not X
required
11-3 C C C X X
11-2 C C Not X X
required
11-1 C C X X X
11-0 C Not X X X
required
10-2 C X X X X
10-1 Not X X X X
required
10-0 X X X X X

C - Compatible
X – Not compatible
* – New Feature Branches (such as 465.xx) are not supported targets for CUDA Forward
Compatibility.

Note: 440.33.01+ (CUDA 10.2) is End of Life.

3.3.2.  Feature Exceptions

There are specific features in the CUDA driver that require kernel-mode support and will
only work with a newer kernel mode driver. A few features depend on other user-mode
components and are therefore also unsupported.

Table 4. Forward-Compatible Feature-Driver Support Matrix

CUDA Forward CUDA – OpenGL/Vulkan POWER9 ATS cuMemMap* set of
Compatible Upgrade Interop functionalities
System Base Installation: 450 (>=.80.02) Driver
11-x No Yes [2] Yes [1]
System Base Installation: 418 (>=.XX) Driver

CUDA Compatibility v   |   12

 Forward Compatibility

11-x No Yes [2] Yes [1]

10-2 No No Yes [1]

[1] This relies on

CU_DEVICE_ATTRIBUTE_HANDLE_TYPE_POSIX_FILE_DESCRIPTOR_SUPPORTED and
CU_DEVICE_ATTRIBUTE_VIRTUAL_ADDRESS_MANAGEMENT_SUPPORTED, which should be
queried if you intend to use the full range of this functionality.
[2] Supported on Red Hat Enterprise Linux operating system version 8.1 or higher.

3.3.3.  Check for Compatibility Support

In addition to the CUDA driver and certain compiler components, there are other drivers in the
system installation stack (for example, OpenCL) that remain on the old version. The forward-
compatible upgrade path is for CUDA only.
A well-written application should use following error codes to determine if CUDA Forward
Compatible Upgrade is supported. System administrators should be aware of these error
codes to determine if there are errors in the deployment.
1. CUDA_ERROR_SYSTEM_DRIVER_MISMATCH = 803. This error indicates that there is a
mismatch between the versions of the display driver and the CUDA driver.
2. CUDA_ERROR_COMPAT_NOT_SUPPORTED_ON_DEVICE = 804. This error indicates that the
system was upgraded to run with forward compatibility but the visible hardware detected
by CUDA does not support this configuration.

3.4.  Deployment Model for Forward

Compatibility
There are two models of deployment for the CUDA compat package. We recommend the use
of the ‘shared’ deployment mode.

‣ Shared deployment: Allows sharing the same compat package across installed toolkits
in the system. Download and extract the latest forward compatibility package with the
highest toolkit version in its name. Using RPATH, or through LD_LIBRARY_PATH or
through an automatic loader (for example. ld.so.conf), point to that package. This is the
most recommended choice.
The user can set LD_LIBRARY_PATH to include the files installed before running the CUDA
11.1 application:

$ LD_LIBRARY_PATH=/usr/local/cuda/compat:$LD_LIBRARY_PATH
‣ Per-application deployment: Individual applications can choose a package of their choice
and place it as part of the Modules system tied to the toolkit and the libraries. Using the
Modules system, the admin, or the user, can set up ‘module’ scripts for each version of
each toolkit package, and then load the module script for the toolkit as needed.

$ module load cuda/11.0

CUDA Compatibility v   |   13

 Forward Compatibility

There is an important consideration to the per-application deployment approach. Older

forward compatibility packages are not supported on new driver versions. Therefore the
module load scripts should proactively query the system for whether the compatibility
package can be used before loading the files. This is especially critical if there was a
full system upgrade. In the cases where the module script cannot use CUDA compatible
upgrade, a fallback path to the default system’s installed CUDA driver can provide a more
consistent experience and this can be achieved using RPATH.

CUDA Compatibility v   |   14

Chapter 4. Conclusion

The CUDA driver maintains backward compatibility to continue support of applications built
on older toolkits. Using a compatible minor driver version, applications build on CUDA Toolkit
11 and newer are supported on any driver from within the corresponding major release.
Using the CUDA Forward Compatibility package, system administrators can run applications
built using a newer toolkit even when an older driver that does not satisfy the minimum
required driver version is installed on the system. This forward compatibility allows the CUDA
deployments in data centers and enterprises to benefit from the faster release cadence and
the latest features and performance of CUDA Toolkit.
CUDA compatibility helps users by:

‣ Faster upgrades to the latest CUDA releases: Enterprises or data centers with NVIDIA
GPUs have complex workflows and require advance planning for NVIDIA driver rollouts.
Not having to update the driver for newer CUDA releases can mean that new versions of
the software can be made available faster to users without any delays.
‣ Faster upgrades of the CUDA libraries: Users can upgrade to the latest software libraries
and applications built on top of CUDA (for example, math libraries or deep learning
frameworks) without an upgrade to the entire CUDA Toolkit or driver. This is possible as
these libraries and frameworks do not have a direct dependency on the CUDA runtime,
compiler or driver.

CUDA Compatibility v   |   15

Chapter 5. Frequently Asked Questions

This section includes some FAQs related to CUDA compatibility.

‣ What is the difference between CUDA forward compatible upgrade and CUDA minor
version compatibility? When should users use these features?

CUDA Forward Compatible CUDA Minor Version

Area Upgrade Compatibility
Compatibility Across older drivers from Across minor release versions
different major release of CUDA only.
versions of CUDA.
Between kernel driver and user Between libraries or runtimes
mode CUDA driver. that link to the CUDA driver.
When to use If you cannot upgrade the If you want to support newer
kernel driver but need to use applications on older drivers
the latest CUDA Toolkit. within the same major release
family.
GPUs supported NVIDIA Data Center GPUs only All GPU products supported
OS distributions supported Linux only Windows, Linux
Mobile platforms supported No Yes
Features Supported Some features such as (CUDA- All existing CUDA features
GL interop, Power 9 ATS, (from older minor releases)
cuMemMap APIs) are not work. Users may have to
supported. These features incorporate checks in their
depend on a new kernel application when using new
mode driver and thus are not features in the minor release
supported. (that require a new driver) to
ensure graceful errors. More
These are explicitly called out information can be found in
in the documentation. the “Handling New CUDA
Features” section.
CUDA Releases Supported All CUDA releases supported Only works within a ‘major’
through the lifetime of the release family (e.g. 11.0
datacenter driver branch. through 11.x).
For example, R418 (CUDA
10.1) EOLs in March 2022 - so Compatibility is not supported
all CUDA versions released across major CUDA releases.
(including major releases)

CUDA Compatibility v   |   16

 Frequently Asked Questions

CUDA Forward Compatible CUDA Minor Version

Area Upgrade Compatibility
during this timeframe are
supported.
Application includes PTX or No additional workflow Users should use the new
uses NVRTC required. PTX static library to rebuild
binaries. Refer to the workflow
section for more details.
Requires administrator Depends on the deployment. Not required.
involvement Users can also set up
LD_LIBRARY_PATH with the
new libraries from the cuda-
compat-* package.

‣ Does CUDA forward compatible upgrades work intra-branch?

Users can upgrade the kernel mode driver within the same branch. Sometimes this may
require updating the cuda-compat* package. This use-case is supported only for drivers
on LLB and LTS branches of driver for select GPUs.
‣ Which GPUs are supported by the driver ?
The CUDA compatible upgrade is meant to ease the management of large production
systems for enterprise customers. 11.4 UMD (User Mode Driver) and later will extend
forward compatibility support to select NGC Ready NVIDIA RTX boards. Prior to that
forward compatibility will be supported only on NVIDIA Data Center cards.
It’s important to note that HW support is defined by the kernel mode driver and as such,
newer CUDA drivers on their own will not enable new HW support. Refer to the following
table to determine which hardware is supported by your system.

Hardware Compute Driver

GenerationCapability 418.40.04+440.33.01+450.36.06+460.27.0+ 470.57.02+495.29.05+510.39.01+
NVIDIA 8.x No No Yes Yes Yes Yes Yes
Ampere
GPU
Arch.
Turing 7.5 Yes Yes Yes Yes Yes Yes Yes
Volta 7.x Yes Yes Yes Yes Yes Yes Yes
Pascal 6.x Yes Yes Yes Yes Yes Yes Yes
Maxwell 5.x Yes Yes Yes Yes Yes Yes Yes
Kepler 3.x Yes Yes Yes Yes Yes No No

‣ What’s the minimum required driver version of a toolkit?

Refer to the Release notes.
‣ The developer is using PTX code in the application and seeing some errors or issues.
What should we do?
PTX and application compatibility information can be found in the CUDA Compatibility
Developer’s Guide CUDA C++ Programming Guide.

CUDA Compatibility v   |   17

 Frequently Asked Questions

‣ If we build our CUDA application using CUDA 11.0, can it continue to be used with newer
NVIDIA drivers (e.g. CUDA 11.1/R455, 11.x etc.)? Or is it only the other way around?
Drivers have always been backwards compatible with CUDA. This means that a CUDA 11.0
application will be compatible with R450 (11.0), R455 (11.1) and beyond. CUDA applications
typically statically include all the libraries (for example cudart, CUDA math libraries
such as cuBLAS, cuFFT) they need, so they should work on new drivers or CUDA Toolkit
installations.
In other words, since CUDA is backward compatible, existing CUDA applications can
continue to be used with newer CUDA versions.
‣ What is the minimum CUDA 11.x driver that supports the CUDA minor version
compatibility?
The minimum driver version required is 450.80.02.
‣ What about new features introduced in minor releases of CUDA? How does a developer
build an application using newer CUDA Toolkits (e.g. 11.x) work on a system with a CUDA
11.0 driver (R450)?
By using new CUDA versions, users can benefit from new CUDA programming model APIs,
compiler optimizations and math library features.

‣ A subset of CUDA APIs don’t need a new driver and they can all be used without any
driver dependencies. For example, async copy APIs introduced in 11.1 do not need a
new driver.
‣ To use other CUDA APIs introduced in a minor release (that require a new driver), one
would have to implement fallbacks or fail gracefully. This situation is not different from
what is available today where developers use macros to compile out features based on
CUDA versions. Users should refer to the CUDA headers and documentation for new
CUDA APIs introduced in a release.
There are some issues that admins can advise the application developers to
accommodate in their code. Please refer to the Best Practices Guide for further
information.
‣ Does CUDA compatibility work with containers?
Yes. CUDA minor version compatibility and CUDA forward compatible upgrade both work
when using either NGC Deep Learning Framework containers or using containers that
are based on the official CUDA base images. The images include the CUDA compatible
upgrade libraries and the NVIDIA Container Toolkit (nvidia-docker2) has logic to correctly
load the required libraries.

‣ I’m running an NGC container and see this error: “This container was built for NVIDIA
Driver Release 450.51 or later, but version 418.126.02 was detected and compatibility
mode is UNAVAILABLE.”. What could be wrong?
It is possible you are either running a wrong version of the NVIDIA driver on the system or
your system does not have an NVIDIA Data Center GPU.

CUDA Compatibility v   |   18

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

Copyright
© 2007-2022 NVIDIA Corporation and affiliates. All rights reserved.

NVIDIA Corporation  |  2788 San Tomas Expressway, Santa Clara, CA 95051
https://www.nvidia.com
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 and/or registered trademarks of NVIDIA Corporation in the Unites States and other countries. Other company and product names may be trademarks of the respective
companies with which they are associated.

Copyright
© 2007-2022 NVIDIA Corporation and affiliates. All rights reserved.

NVIDIA Corporation  |  2788 San Tomas Expressway, Santa Clara, CA 95051
https://www.nvidia.com