#
backendai
Here are 13 public repositories matching this topic...
adrysn
commented
Mar 3, 2022
GPU value is broken when creating a cluster session (N=3) with 0.3 fGPU.
Steps to reproduce
- Create a cluster compute session with a size of 3. It doesn't matter the node type: single-node or multi-node.
- In the last page of the session launch dialog, you will notice the broken GPU unit: 0.999999999....
<img width="430" alt="image" src="https://user-images.githubusercontent.com/753935
Jupyter kernel integration for Backend.AI
-
Updated
Nov 9, 2018 - Python
Backend.AI Client Library for Python
-
Updated
Apr 18, 2022 - Python
A programmable security sandbox for Backend.AI kernels
-
Updated
May 17, 2021 - Go
Visual Studio Code extension to run code snippets using Backend.AI
-
Updated
Jul 27, 2018 - TypeScript
ATOM package to run code snippets using Sorna API server
-
Updated
Jul 27, 2018 - JavaScript
Improve this page
Add a description, image, and links to the backendai topic page so that developers can more easily learn about it.
Add this topic to your repo
To associate your repository with the backendai topic, visit your repo's landing page and select "manage topics."
To guarantee liveness of distributed locks, we need to ensure release of the locks even when the lock holder gets hang or abruptly killed.
We have implemented auto-release of locks with
EtcdLockwith explict leases, butFileLockcurrently does not have such mechanism and just relies on the kernel's automatic file descriptor cleanup upon (abrupt) shutdown.This issue is separated from the ori