1

OS: Linux 4.18.0-193.28.1.el8_2.x86_64 anaconda: anaconda3/2022.10

Trying to install RAPIDS, I get:

$ conda install -c rapidsai rapids
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: | 
Found conflicts! Looking for incompatible packages.
This can take several minutes.  Press CTRL-C to abort.
failed                                                                                                                                                                     

UnsatisfiableError: The following specifications were found to be incompatible with each other:

Output in format: Requested package -> Available versionsThe following specifications were found to be incompatible with your system:

  - feature:/linux-64::__glibc==2.28=0
  - feature:|@/linux-64::__glibc==2.28=0
  - rapids -> cucim=22.12 -> __glibc[version='>=2.17|>=2.17,<3.0.a0']

Your installed version is: 2.28


$

As has been asked by others (but, as far as I can tell, not answered), why is "__glibc" version 2.28 not between 2.17 & 3.0?

talonmies
  • 70,661
  • 34
  • 192
  • 269
Mark Bower
  • 569
  • 2
  • 16

2 Answers2

2

i've got the same problem.

My solution was to create a new conda enviroment with python 3.8/3.9 and use the pip installation recommended by RAPIDS (just create the env and install rapids)

pip install cudf-cu11 dask-cudf-cu11 --extra-index-url=https://pypi.ngc.nvidia.com

pip install cuml-cu11 --extra-index-url=https://pypi.ngc.nvidia.com

pip install cugraph-cu11 --extra-index-url=https://pypi.ngc.nvidia.com

Marcelo Oyaneder
  • 36
  • 2
  • This fix got me most the way there, I had to also install CUDA 11.5 toolkit fully, the cuda packages downloaded via pip weren't enough. libcudart.so was still reporting as missing when importing. Thanks though!! – jamboNum5 Jan 30 '23 at 10:46
0

Please try using the full install command, as show in the getting started guide, which pins the rapids, python, and cuda tool kit versions, as well as some of the channels to retrieve any supporting packages:

conda install -c rapidsai -c conda-forge -c nvidia rapids=22.12 python=3.9 cudatoolkit=11.5

TaureanDyerNV
  • 1,208
  • 8
  • 9
  • This is a decent start, but they probably need a new environment, so `conda create` instead of `conda install` – merv Jan 01 '23 at 15:31