mlcommons / mlcommons/power-dev

Integrate client into loadgen

Open
#226 1 comment 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
Python
Stars
28
Forks
30
PR merge metrics
No merged PRs in 30d

Description

Instead of having standalone executable client.py (or C++ version, see #112) that runs arbitrary workload, we could integrate the client into loadgen itself. This would simplify running the power measurement on the SUT side, and allow us to measure in more environments, e.g. mobile.

We considered this option in the past (2020-12-11), but decided to postpone it after round 1.0. Probably should be done in the inference repo.

Estimate time of implementation: 3-4 weeks.

  • Analyze loadgen codebase to understand what is the best way of integration
  • Create minimal working prototype that can connect to the server
  • Integrate all necessary features

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by analyzing the loadgen codebase and the existing standalone client.py; use the C++ version referenced in #112 for comparison. A minimal prototype should connect to the server, followed by integration of all necessary features so power measurement can run on the SUT and in environments such as mobile.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp, python
Domain
performance, tooling
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.