Skip to content

[0.3] Add E.M.B.E.R. admission and guest-aware benchmark planning - #17

Merged
pschildgen87-code merged 23 commits into
mainfrom
feat/0.3.0-alpha.4-ember-admission
Sep 5, 2026
Merged

pschildgen87-code merged 23 commits into
mainfrom
feat/0.3.0-alpha.4-ember-admission

Conversation

@pschildgen87-code

Copy link
Copy Markdown
Contributor

Summary

Builds the next Runtime 0.3 layer on top of alpha.3 host discovery/placement.

  • adds a machine-readable application deployment-profile contract and builtin ember-production profile
  • keeps the E.M.B.E.R. production requirements separate from generic Runtime minimums: 64 GiB memory, ECC required, DDR5 preferred, 8 physical cores, GPU optional, MoE mandatory; the historical ~2.6 GHz figure remains advisory only
  • adds explicit admission results (admitted, rejected, needs_operator_evidence) with per-requirement evidence instead of guessing ECC or silently substituting dense models
  • adds guest/process CPU-limit discovery from affinity, cgroup cpusets and CPU quota plus virtualization evidence; CPU-model names cannot imply that all host-physical cores are allocated to a KVM guest
  • adds operator core budgets so a 20-core guest can reserve part of the allocation for the studio/game/system workload
  • adds a real-model benchmark-plan contract that generates CPU worker-count candidates without pretending those candidates are measured throughput
  • the current 20-core / 4-core reserve reference yields the intended 8/12/16 production sweep plus 20-core full-host comparison when requested
  • adds /v1/deployment-profiles, /v1/admission, /v1/benchmark-plan and exposes host limits through /v1/host

The exact smallest historical Netcup root-server hardware is intentionally not guessed. Its observed ability to run the E.M.B.E.R. construct with small models/RAG-Lite/LTM is treated as functional-compatibility evidence, not as the current supported production intelligence profile.

Refs #8, #11, #12, #13.

@chatgpt-codex-connector

Copy link
Copy Markdown

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Sep 5, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
🔒 Security Review ✅ Completed 2026-09-05T23:16:38.249115Z 9771c4f PR opened
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant