lablup / lablup/backend.ai

Explicitly visible absolute resource amounts for fractionally scaled resources

Open
#97 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
670
Forks
183
Avg merge
15h 13m
Merged PRs (30d)
368

Description

For example, depending on configuration, 0.1 vGPU may have different amounts GPU memory and SMP counts in different agents/scaling-groups/clusters.There are customer feedbacks that they want to see the actual amount of such resources.

Since technially the manager does not interact with accelerator plugins because they are installed in agents only and the manager only considers the amount of heterogeneous resource slots, there is no simple way to provide an API to query this, unless we **_break_ the abstraction layer** (e.g., directly referencing the etcd's plugin configuration in a dedicated API handler).

Let's make a way to calculate this generally for any fractional-resource-providing plugins.

┆Issue is synchronized with this [Asana task](https://app.asana.com/0/1159751085623729/1159756855908144) by [Unito](https://www.unito.io/learn-more)

JIRA Issue: BA-313

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.