conda error while installation

[KHBHH]

Hi all,

I was having an issue for the installation to start, but looking at one of the previous posted issues I was enable to start the installation using:

./install_gui.sh -f ClearMapUi39.yml

But now I am getting error as such:

Updating CUDA dependencies for ClearMap
Creating temporary environment
usage: conda [-h] [-v] [--no-plugins] [-V] COMMAND ...
conda: error: unrecognized arguments: --experimental-solver=libmamba

What might the issue be for the conda error?

[KHBHH]

Hi, I was trying to install cuda following this link:

Now instead of the above error, I am getting the following: https://developer.nvidia.com/cuda-11-6-0-download-archive?target_os=Linux&target_arch=x86_64&Distribution=Ubuntu&target_version=20.04&target_type=deb_local

Do you wish to use this feature ([y]/n)?y
Checking nVIDIA drivers and system CUDA installation
Error, a working installation of NVidia CUDA could not be detected on your computer. This software is required to run ClearMap.

How should I solve this issue? (my Ubuntu version is: Ubuntu 22.04.3 LTS)

To check for the installation of NVIDIA Cuda, I ran the following:

(base) karni@wu-lct-analytics:~/ClearMap2$ nvcc --version

The output was:
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Wed_Jul_22_19:09:09_PDT_2020
Cuda compilation tools, release 11.0, V11.0.221
Build cuda_11.0_bu.TC445_37.28845127_0

This means I have NVIDIA Cuda installed right? The version of the cudatoolkit here is 11.0 and in the ClearMAp39ui.yml file it is 11.6, do you think that might be the issue?

[crousseau]

@KHBHH

It seems there are some recent changes in conda making this incompatible. You can try replacing
--experimental-solver=libmamba
in the install_gui.sh script
with
--solver=libmamba

Please let me know if this works

[KHBHH]

Hi, I guess the error now is no longer that but with nvidia cuda toolkit, I have an error as such:

ClearMap uses neural networks to perform vasculature analysis.
The implementation of these networks relies on proprietary technology
from nVIDIA called CUDA. To perform vasculature analysis, you will
need a compatible graphics card and drivers.
Do you wish to use this feature ([y]/n)?y
Checking nVIDIA drivers and system CUDA installation
/bin/sh: 1: nvidia-smi: not found
Error, a working installation of NVidia CUDA could not be detected on your computer. This software is required to run ClearMap.

When I check the (base) karni@wu-lct-analytics:~/ClearMap2$ nvcc --version

I have an error as such:

(base) karni@wu-lct-analytics:~/ClearMap2$ nvcc --version

It shows I have NVIDIA Cuda compiler driver

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2021 NVIDIA Corporation
Built on Thu_Nov_18_09:45:30_PST_2021
Cuda compilation tools, release 11.5, V11.5.119
Build cuda_11.5.r11.5/compiler.30672275_0

What do you think might be the issue?

[KHBHH]

Hi, so when I check this:

(base) karni@wu-lct-analytics:~/ClearMap2$ nvidia-smi

NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

It means I have the driver installed but can not communicate, how to solve?

[crousseau]

If you are using ubuntu, the simplest is to use the additional drivers graphical interface
sudo update-manager -> settings -> additional drivers. Clearmap has been tested with drivers up to 525 (no server and no open-source)
Then reboot your machine for the driver to be reloaded and try the install procedure again