Refactor common result metadata processing into a shared helper

Open
#2,105 1 comment 0 reactions 1 assignee View on GitHub

@cquil11 is already working on this.

Since Jul 6, 2026.

Assessment

This issue has not been assessed yet.

Description

Problem

utils/process_result.py and utils/agentic/aggregation/process_agentic_result.py independently parse and validate common benchmark metadata. PR #2100 adds heterogeneous prefill/decode hardware metadata to both paths, which highlights the maintenance cost and risk of the two implementations drifting.

Goal

Move shared environment parsing, topology validation, and common result fields behind one reusable implementation consumed by both regular and agentic result processors.

Acceptance criteria

  • Regular and agentic result processors use the same code for common metadata and multinode topology fields.
  • Optional PREFILL_HARDWARE / DECODE_HARDWARE handling is implemented once, with both-or-neither validation.
  • Existing homogeneous, heterogeneous, single-node, and multinode result schemas remain compatible.
  • Shared table-driven tests cover the common behavior; processor-specific tests focus on integration and metric output.
Dominant language
Python
Stars
1.7k
Forks
303
Avg merge
1d 13h
Merged PRs (30d)
284

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.

More from SemiAnalysisAI/InferenceX

All issues in SemiAnalysisAI/InferenceX

Similar issues

More Python issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.