Replies - Activity - Karun - DeepTalk - Deep Learning Community

What drivers/patches are necessary for the razer tensorbook?

@eureka
I am not a LambdaLab person.
I noticed that your github link above is not valid as it has space and hence it is starting to land at b.com/eureka…
Good luck.

Lambda Stack seems to be wiped out after Ubuntu update

Problem is solved.
Since via
sudo apt-get remove lambda-stack-cuda && sudo apt-get install lambda-stack-cuda
This failed in the middle due to interruption and I have to do
sudo dpkg --configure -a and reissue the above install of lambda-stack-cuda.
This went through, I rebooted the system and a…

Lambda Stack CUDA and Nvidia drivers

Yes I also want to upgrade Cuda to 12.1 and compatible torch version.
As per this Nvidia Official Drivers | NVIDIA
We can upgrade the driver via the above link and the torch via the pytorch.org page

Is there a lambda-stack way to upgrade them?

Lambda Stack seems to be wiped out after Ubuntu update

Hi Cody,
When I rebooted again, the system came back. I am able to connect.
How to check whether the
sudo apt-get install lambda-stack-cuda is complete?
When I do
sudo apt-cache policy lambda-stack-cuda I get

Also when I checked tensorflow and torch versions I see the following:

Thanks for reply.

On my LambdaBlade with Ubuntu 20.04 and 4X A6000, I wanted to upgrade the cuda stack.
So I issued the this command:

$ sudo apt-get remove lambda-stack-cuda && sudo apt-get install lambda-stack-cuda  

After it successfully fetched 115 packages, extracted templates, read database c…

Just want to clarify.

System LambdaBlade with Ubuntu 20.04:
Is there a difference between the command here Getting-started Removing and Installing Lambda-stack-in-ubuntu
Uninstall (purge) the existing Lambda Stack by running:

sudo rm -f /etc/apt/sources.list.d/{graphics,nvidia,cuda}* &&  
dpkg ...  

Lambda Quad - How to Unbond NVLINK from P2P/SLI mode

Any update?
I have a blade system with 4 A6000.
What does the below output mean?

nvidia-smi nvlink -gt d -i 0,1,2,3  

GPU 0: NVIDIA RTX A6000 (UUID: GPU-938d731b-…)
Link 0: Data Tx: 63,654,019,510 KiB
Link 0: Data Rx: 7,631,152,229 KiB
Link 1: Data Tx: 63,650,865,559 KiB
Link 1: Data Rx: 7,6…

Unable to determine the device handle for GPU, GPU is lost. Reboot the system to recover this GPU

As per the tech support, reboot did the trick.
Hope this is not a persistent vibration related issue.
-Karun

BERT Multi-GPU TensorFlow and Horovod for prediction - no improvement

We have noticed a similar issue with PyTorch version of Transformer for NLP.
This is the system config:
It is a Lambda blade server with 8 RTX 3090 GPUs, NVIDIA Driver Version: 460.56, CUDA Version: 11.2.
Ubuntu 20.04 LTS, Dual AMD EPYC 7302 16-Core CPUs, 512GB of RAM.
We did a quick profiling o…

How to install libcudnn.8 Ubuntu 20.04 NVIDIA-SMI 460.56 Driver Version: 460.56 CUDA Version: 11.2

Got it:
Followed this python - "Could not load dynamic library 'libcudnn.so.8'" when running tensorflow on ubuntu 20.04 - Stack Overflow
Steps are:

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/cuda-ubuntu2004.pin  
sudo mv cuda-ubuntu2004.pin /etc/apt/preferences…  

Dist-upgrade does not upgrade to Cuda 11.1

Just a hunch, but I’m thinking there still might be some python*-*dev packages installed that are blocking this.
Can you send me the output of
dpkg -l | grep python

$ dpkg -l | grep python | grep minimal  
ii libpython2.7-minimal:amd64 2.7.18-1~20.04.1 amd64 …  

Dist-upgrade does not upgrade to Cuda 11.1

After I upgraded (with several hoops and tries) to Ubuntu 20.04 Desktop (from 18.04) via

sudo sed -i ‘/^AutomaticLoginEnable/ s/true/false/’ /etc/gdm3/custom.conf &&  
sudo apt -y update && sudo apt -y dist-upgrade && sudo do-release-upgrade -d -f DistUpgradeViewNonInteractive  

And then rebooted and…

How to replace the SSD in LambdaQuad?

I meant smartctl --all /dev/nvme0n1p2 says the
SMART overall-health self-assessment test result: PASSED