Conversation
Signed-off-by: Dmitry Shmulevich <dshmulevich@nvidia.com>
Greptile SummaryThis PR adds a "Motivation and Problem Statement" section to the README, explaining why topology-aware scheduling matters for GPU workloads at scale. It also renames the old "Components" section to "Architecture," updates the project's introductory sentence, and rewrites the API Server component description for clarity. All changes are documentation-only with no code impact. Confidence Score: 5/5Documentation-only change, safe to merge with no code impact. All changes are to README.md prose. The new Motivation section is technically accurate, well-written, and consistent with the rest of the document. No code, logic, or configuration is modified. No files require special attention. Important Files Changed
Flowchart%%{init: {'theme': 'neutral'}}%%
flowchart TD
Sources["Topology Sources\n(Cloud APIs, Fabric Mgmt, System Tools)"]
Provider["Provider\n(discovers & normalizes topology)"]
Internal["Internal Topology Representation"]
Engine["Engine\n(translates to scheduler format)"]
Schedulers["Schedulers\n(Slurm / Kubernetes / Slinky)"]
APIServer["API Server\n(aggregates requests, async results)"]
NodeObserver["Node Observer\n(Kubernetes node watch)"]
NodeDataBroker["Node Data Broker\n(Kubernetes node annotations)"]
Sources --> Provider
NodeObserver -->|"node change event"| APIServer
NodeDataBroker -->|"node attributes"| Provider
APIServer -->|"trigger discovery"| Provider
Provider --> Internal
Internal --> Engine
Engine --> Schedulers
Reviews (1): Last reviewed commit: "docs(readme): add 'Motivation and Proble..." | Re-trigger Greptile |
No description provided.