How to compile FFMpeg with older NVEnc? - ffmpeg

I get this error when running FFMpeg with NVEnc.
[h264_nvenc # 0x2a27d00] Driver does not support the required nvenc API version. Required: 9.1 Found: 9.0
[h264_nvenc # 0x2a27d00] The minimum required Nvidia driver for nvenc is 435.21 or newer
I can't upgrade the driver to latest version because I'm limited by what's supported in GKE.

Resolved. I was compiling ffmpeg with the latest code in https://git.videolan.org/git/ffmpeg/nv-codec-headers.git
Switching to the sdk/9.0 branch worked for the nvidia driver on GKE.

Which drivers are supported largely depends on the node OS (COS or Ubuntu) and the GKE version the node is running. When you add a GPU node to your cluster, GKE provisions a DaemonSet that automatically installs the drivers for you. A list of supported NVIDIA driver versions per GKE version can be found in the document [1] that Google has at present.
The Nvidia drivers on your cluster get updated when the node itself gets an update to the newer version of GKE. You can follow the newest feature release of GKE in the document [2].
Now, as we do not have the required NVENC driver version, an alternative approach might be to pull required older version of FFmpeg.
[1]--https://cloud.google.com/kubernetes-engine/docs/how-to/gpus#installing_drivers
[2]--https://cloud.google.com/kubernetes-engine/docs/release-notes

Related

QLI Support for Greenplum version 6.22

As gpperfmon is no longer supported for version 6.22, How can we able to perform ingesting QLI.

Upgrading QT5 package in buildroot-2019.02.8

Currently using Buildroot-2019.02 with support for Qt5.6.3 and Qt5.11.x. BSP is built with customized Linux kernel.
Please suggest if it is possible to upgrade Qt5.15.x version in the existing Buildroot 2019.02.x version
If it is not feasible, the only option is to upgrade to a higher version of Buildroot-2022* where it is Qt5.15.8 and also upgrade the Linux kernel and customize it as older GCC wouldn't be supported(which could take some time).

Ubuntu 20.4 Asus Tuf Dash F15 Can't Install Nvidia Driver

I have installed a fresh version of Ubuntu 20.4 on my band new Asus Tuf Dash F15 laptop. Then I went into the software update settings and switched the driver to nvidia-460 proprietary using the GUI. I apply the changes and restart and I get stuck at a boot loading screen. The error is /dev/nvm1e0 clean: xxxblocks/xxx.
Specs: Nvidia RTX 36060 mobile Intel i7 40gb ram 1tb storage
I have tried installing the nvidia drivers 4 different ways, according to the nvidia official installation guide and multiple medium forums. Every time I get the same problem and have to alt+shift+f2 my way into the terminal to sudo apt-get purge nvidia* to be able to boot back in.
I've installed nvidia diver on ubuntu many other times according to the nvidia official docs and never ran into this error. I am concerned it may be hardware incompatibility at this point?
I've been trying for five days now and re-installed ubuntu multiple times, varying the partitions and installation methods. Everything is working fine until I try and switch to nvidia driver. I even tried some older and new driver versions and the nvidia cuda toolkit 10 and 11. Please help, thank you.

The minimum required Cuda capability is 3.5

After installing TensorFlow and its dependencies on a g2.2xlarge EC2 instance I tried to run an MNIST example from the getting started page:
python tensorflow/models/image/mnist/convolutional.py
But I get the following warning:
I tensorflow/core/common_runtime/gpu/gpu_device.cc:611] Ignoring gpu device
(device: 0, name: GRID K520, pci bus id: 0000:00:03.0) with Cuda compute
capability 3.0. The minimum required Cuda capability is 3.5.
Is this a hard requirement? Any chance I could comment that check out in a fork of TensorFlow? It would be super nice to be able to train models in AWS.
There is a section in the official installation page that guides you to enable Cuda 3, but you need to build Tensorflow from source.
$ TF_UNOFFICIAL_SETTING=1 ./configure
# Same as the official settings above
WARNING: You are configuring unofficial settings in TensorFlow. Because some
external libraries are not backward compatible, these settings are largely
untested and unsupported.
Please specify a list of comma-separated Cuda compute capabilities you want to
build with. You can find the compute capability of your device at:
https://developer.nvidia.com/cuda-gpus.
Please note that each additional compute capability significantly increases
your build time and binary size. [Default is: "3.5,5.2"]: 3.0
Setting up Cuda include
Setting up Cuda lib64
Setting up Cuda bin
Setting up Cuda nvvm
Configuration finished
Currently only GPUs with compute capability >= 3.5 are officially supported. However, GitHub user #infojunkie has offered a patch that makes it possible to use TensorFlow with a GPU with compute capability 3.0.
The official fix is in development. Meanwhile, check out the discussion on the GitHub issue for adding this support.
There is a simple trick. You don't even have to build TF from sources.
In the file tensorflow\python\_pywrap_tensorflow.pyd there are two occurences of regex 3\.5.*5\.2. Just replace both 3.5 with 3.0.
Tested on Windows 10, Anaconda 4.2.13, Python 3.5.2, TensorFlow 0.12, CUDA 8, NVidia GTX 660m (CUDA cap. 3.0).

Firefox on Red hat 5.6

What is the maximum version of Firefox that can run on Red hat 5.6?
The default installation is 3.2.6... I want to run some WebDriver tests but having issues which are due to the Firefox version.
When attempting to update Firefox to 35.0 a error arises with libgio-2.0.so.0()(64bit) not found.
I believe this library is in GLib package and internal to the OS so can't be updated?
Firefox puts out what they call ESRs (Extended Service Releases) which typically have a wider target audience of distributions that they support. ESR 17 was the last to work on CentOS 5.x so I wouldn't expect to go much further than this one.
What is Mozilla Firefox ESR?
Mozilla will offer an Extended Support Release (ESR) based on an official release of Firefox for desktop for use by organizations including schools, universities, businesses and others who need extended support for mass deployments. You can read more about the plan here.
Version 24 of Firefox will become ESR 24, and this version definitely will not run on CentOS 5.x (according to the testing I've done on 5.9). So you're essentially stuck on this older version until Mozilla releases a newer version built with the older libraries support.

Resources