aws / aws/res

Filter instance types by software stack attributes (CPU/GPU)

Open
#58 1 comment 0 reactions 0 assignees View on GitHub
enhancement
Dominant language
Python
Stars
115
Forks
35
PR merge metrics
No merged PRs in 30d

Description

**Is your feature request related to a problem? Please describe.**

GPU instances shouldn't be selectable when software stack doesn't provide relevant support (AMD/NVIDIA).

**Describe the solution you'd like**

Instance list should filter based on "GPU Manufacturer":
* N/A - CPU instance types only
* NVIDIA - Appropriate GPU instances
* AMD - Appropriate GPU instances

The same consideration could be used on CPU types ARM64 vs x86_64, this is not captured in the Software Stack table at present.

**Describe alternatives you've considered**

Unaware of alternatives to filter list shows under "Virtual Desktop Size" other than CPU/Memory scaling.

**Additional context**

Impacts user experience for self-service due to increased barrier to entry.

The current option allows for easy filtering to a feature of an instance type for example "CPUs: 4" will filter the list quite well. Unfortunately this does not remove GPU instance types from the remaining list. Use of a GPU instance type requires additional understanding of the instance types to select the correct chipset for the stack.

As noted the same could occur for CPU.

Contributor guide

Open the contributing guide

Research direction

Start at the Virtual Desktop Size instance list and the Software Stack table described in the issue; trace how CPU/GPU attributes are represented and how list filters are applied. Done means CPU-only, NVIDIA, and AMD selections show appropriate instance types, with ARM64/x86_64 considered if supported by the same data.

Written by the indexing model from the issue text.

Assessment

Tech stack
aws
Domain
cloud
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.