question-mark
Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

docker_image: cos7-cuda:10.2 chosen over cos7-x86_64 for os_version: { linux_64: cos7 }

See original GitHub issue

Re-rendered with conda-build 3.21.4, conda-smithy 3.9.0, and conda-forge-pinning 2021.03.02.12.06.47

If

os_version:
  linux_64: cos7

is used in conda-forge.yml, re-rendering causes the docker_image key to get value quay.io/condaforge/linux-anvil-cos7-cuda:10.2 rather than quay.io/condaforge/linux-anvil-cos7-x86_64 even for non-CUDA recipes.

AFAICT, https://github.com/conda-forge/conda-smithy/blob/v3.9.0/conda_smithy/configure_feedstock.py#L197 sorts the configuration values which puts ...-cos7-cuda:10-2 before ..-cos7-x86_64 (since "c" < "x", lexicographically). But https://github.com/conda-forge/conda-smithy/blob/v3.9.0/conda_smithy/configure_feedstock.py#L386 chooses the first value, which is likely supposed to be the first one from the original (conda-forge-pinnings’) order, not from the sort-order. Just omitting the sort would not work either since conda-build’s output is not deterministic and can either have *-cos7-x86_64 or *-cos7-cuda:10.2 as the first entry.

It’s quite possible my assessment above isn’t 100 % correct, but it should be enough to help fix it by someone who’s more versed in the code base.

Issue Analytics

  • State:open
  • Created 3 years ago
  • Comments:5 (5 by maintainers)

github_iconTop GitHub Comments

2reactions
isurufcommented, Mar 2, 2021

conda-build’s output is not deterministic, but we can sort the variants using the original order that they came from where *-cos7-x86_64 comes first.

0reactions
beckermrcommented, Mar 2, 2021

Ah ok. So what I said above then.

Read more comments on GitHub >

github_iconTop Results From Across the Web

nvidia/cuda - Docker Image
CUDA is a parallel computing platform and programming model developed by NVIDIA for general computing on graphical processing units (GPUs).
Read more >
pytorch/pytorch:1.8.1-cuda10.2-cudnn7-devel | Docker Hub
OS /ARCH. linux/amd64. Compressed Size. 4.27 GB. Last pushed. 2 years ago by seemethere. TYPE. Image. Image Layers. 1. ADD file ... in...
Read more >
Explore Docker's Container Image Repository - Docker Hub
A minimal Docker image based on Alpine Linux with a complete package index and only 5 MB in size! Linux. riscv64. x86-64. ARM....
Read more >
docker - Official Image
Docker is an open-source project that automates the deployment of applications inside software containers, by providing an additional layer of abstraction and ...
Read more >
docker version - Docker Documentation
docker version : By default, this will render all version information in an easy to read layout. If a format is specified, the...
Read more >

github_iconTop Related Medium Post

No results found

github_iconTop Related StackOverflow Question

No results found

github_iconTroubleshoot Live Code

Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start Free

github_iconTop Related Reddit Thread

No results found

github_iconTop Related Hackernoon Post

No results found

github_iconTop Related Tweet

No results found

github_iconTop Related Dev.to Post

No results found

github_iconTop Related Hashnode Post

No results found