NVIDIA /open-gpu-kernel-modules
NVIDIA Linux open GPU kernel module source
★ 17,591⑂ 1,979👁 189⑃ 316● 878▤ 223
+6
日增长
17,591
stars
1,979
forks
189
watchers
316
prs
878
issues
223
releases
README
NVIDIA Linux Open GPU Kernel Module Source
This is the source release of the NVIDIA Linux open GPU kernel modules,
version 620.70.06.
How to Build
To build:
make modules -j$(nproc)
To install, first uninstall any existing NVIDIA kernel modules. Then,
as root:
make modules_install -j$(nproc)
Note that the kernel modules built here must be used with GSP
firmware and user-space NVIDIA GPU driver components from a corresponding
620.70.06 driver release. This can be achieved by installing
the NVIDIA GPU driver from the .run file using the --no-kernel-modules
option. E.g.,
sh ./NVIDIA-Linux-[...].run --no-kernel-modules
Supported Target CPU Architectures
Currently, the kernel modules can be built for x86_64 or aarch64.
If cross-compiling, set these variables on the make command line:
TARGET_ARCH=aarch64|x86_64
CC
LD
AR
CXX
OBJCOPY
E.g.,
compile on x86_64 for aarch64
make modules -j$(nproc) \
TARGET_ARCH=aarch64 \
ARCH=arm64 \
CC=aarch64-linux-gnu-gcc \
LD=aarch64-linux-gnu-ld \
AR=aarch64-linux-gnu-ar \
CXX=aarch64-linux-gnu-g++ \
OBJCOPY=aarch64-linux-gnu-objcopy
Other Build Knobs
NV_VERBOSE - Set this to "1" to print each complete command executed;
otherwise, a succinct "CC" line is printed.
DEBUG - Set this to "1" to build the kernel modules as debug. By default, the
build compiles without debugging information. This also enables
various debug log messages in the kernel modules.
These variables can be set on the make command line. E.g.,
make modules -j$(nproc) NV_VERBOSE=1
Supported Toolchains
Any reasonably modern version of GCC or Clang can be used to build the
kernel modules. Note that the kernel interface layers of the kernel
modules must be built with the toolchain that was used to build the
kernel.
Supported Linux Kernel Versions
The NVIDIA open kernel modules support the same range of Linux kernel
versions that are supported with the proprietary NVIDIA kernel modules.
This is currently Linux kernel 4.15 or newer.
How to Contribute
Contributions can be made by creating a pull request on
https://github.com/NVIDIA/open-gpu-kernel-modules
We'll respond via GitHub.
Note that when submitting a pull request, you will be prompted to accept
a Contributor License Agreement.
This code base is shared with NVIDIA's proprietary drivers, and various
processing is performed on the shared code to produce the source code that is
published here. This has several implications for the foreseeable future:
The GitHub repository will function mostly as a snapshot of each driver
release.
We do not expect to be able to provide revision history for individual
changes that were made to NVIDIA's shared code base. There will likely
only be one git commit per driver release.
We may not be able to reflect individual contributions as separate
git commits Top issues
#472Suspend doesn't work when PreserveVideoMemoryAllocations is enabled227 comments · 2023-03-11#483Unable to change power limit with nvidia-smi182 comments · 2023-04-01#880RTX 5090 often freezes presentation for 3-5 seconds in certain games152 comments · 2025-06-13#650 Low fps on external monitor connected to nvidia hdmi port122 comments · 2024-06-02#905GPU resumes to D0 immediately after ending transition to D3Cold (RTD3)95 comments · 2025-07-20#1151RTX 5080 (GB203): Random Xid 79 with zero precursor — any load level, instant atomic GPU death, open kernel module 595.71.0581 comments · 2026-05-18#446Timeout waiting for RPC from GSP!75 comments · 2023-01-26#1045Intermittent desktop lockup: Xid 62/45 -> Xid 119 (GSP timeout) -> Xid 154 on RTX 5080 (Arch Linux, nvidia-open 590.48.01)55 comments · 2026-03-01
Discussions
#1548RTX 4080 Laptop / 620.70.06: intermittent S3 resume DMA failure; request GSP NvLog interpretation0 comments · 2026-11-14#1543Does CUDA VMM teardown serialize across independent GPUs?1 comments · 2026-11-13#1529dododo0 comments · 2026-11-06#1526Using KCSAN to detect race conditions after executing the command "sh ./NVIDIA-Linux-[...].run --no-kernel-modules".0 comments · 2026-11-05#1524All GPU interrupts arrive on MSI-X vector 0 of 8, so one CPU core limits concurrent NVDEC decode per GPU0 comments · 2026-11-05#1519VRR only while fullscreen gives ~2 s black screen on RTX 3090 + DP 1.4 QD-OLED @ 360 Hz (open and closed module 615.78.08)0 comments · 2026-11-04#1496MX570 A: recurring Xid 79 without suspend on 610.57.04 and 615.78.08 — diagnostic guidance?1 comments · 2026-10-26#1494Native DirectX 12 on Linux 😁0 comments · 2026-10-26
Latest release
620.70.06 620.70.06 2026-11-05https://www.nvidia.com/en-us/drivers/details/283867/