Analysis orchestrator
Coordinate system-level analysis (CPU/idle, kernel fusion, multi-kernel / comm / memcpy) and compute-kernel analysis (GEMM, SDPA, elementwise, etc.): one trace load, shared prep, parallel subagents, then aggregation into analysis.md.
Full procedure
Follow reference.md for every step (user prompts, <prefix> / {CMD} usage, CLI commands, subagent launch text, validation, report tee order, plot embedding, and trace diagnostics).
Workflow index
0. Query User Inputs (Platform, Trace Path(s), Analysis Mode, Environment Setup)
1. Generate Performance Report (branches on analysis mode: training vs inference then, comparison scope)
2. Trace-Quality Gate → Deterministic fallback analysis
3-6. Prepare Category Data (GPU Util, Top Ops, Tree Data, Multi-Kernel Data, Category Filtering)
7. System-Level Analysis (PARALLEL) → system_findings/
8. Compute Kernel Subagents (PARALLEL) → category_findings/
8.5. Aggregate → priority_data.json::findings[]
9. Validate Subagent Outputs
10. load_findings + Model Identification (subagent) → metadata/model_info.json
11. Render performance PNG if agent_extension.py is absent
12. Generate analysis.md (orchestrator writes via <prefix> tee), optional extension, embed PNG
Rules
- Subagents: Use the Task tool only where reference.md says “subagent” (Steps 7, 8, 10). The orchestrator runs everything else, including Step 8.5, using the command prefix from
<output_dir>/cache/cmd_prefix.txt ({CMD} substitution).
- Language: Prefer vendor-agnostic terms (GPU kernels, collective communication, vendor GEMM library, DNN primitives, GPU graph). When quoting trace data, real kernel names are fine.
- Subagent prompts: Point each subagent at the checked-in agent file under
TraceLens/Agent/Analysis/skills/analysis-orchestrator/agents/<name>.md (see reference.md for exact paths and prompt shells).
Primary outputs
- Deliverable:
<output_dir>/analysis.md
- Internals:
system_findings/, category_findings/, category_data/, metadata/, perf_report*.xlsx, CSV folders — see package README for layout.
Agent layout
Project subagents ship with this skill: TraceLens/Agent/Analysis/skills/analysis-orchestrator/agents/*.md.
1---2name: tracelens-analysis-orchestrator3description: Orchestrates modular PyTorch profiler trace analysis with TraceLens: generates perf reports, prepares category data, runs system-level and compute-kernel subagents in parallel, validates outputs, and writes a prioritized stakeholder report (analysis.md). Use when the user asks to follow the analysis orchestrator, run the agentic analysis workflow, analyze a trace, compare two traces, or mentions standalone or comparative TraceLens analysis.4license: MIT5---67<!--8Copyright (c) 2026 Advanced Micro Devices, Inc. All rights reserved.910See LICENSE for license information.11-->1213# Analysis orchestrator1415Coordinate **system-level** analysis (CPU/idle, kernel fusion, multi-kernel / comm / memcpy) and **compute-kernel** analysis (GEMM, SDPA, elementwise, etc.): one trace load, shared prep, parallel subagents, then aggregation into `analysis.md`.1617## Full procedure1819Follow **[reference.md](reference.md)** for every step (user prompts, `<prefix>` / `{CMD}` usage, CLI commands, subagent launch text, validation, report `tee` order, plot embedding, and trace diagnostics).2021## Workflow index2223```240. Query User Inputs (Platform, Trace Path(s), Analysis Mode, Environment Setup)251. Generate Performance Report (branches on analysis mode: training vs inference then, comparison scope)262. Trace-Quality Gate → Deterministic fallback analysis273-6. Prepare Category Data (GPU Util, Top Ops, Tree Data, Multi-Kernel Data, Category Filtering)287. System-Level Analysis (PARALLEL) → system_findings/298. Compute Kernel Subagents (PARALLEL) → category_findings/30 8.5. Aggregate → priority_data.json::findings[]319. Validate Subagent Outputs3210. load_findings + Model Identification (subagent) → metadata/model_info.json3311. Render performance PNG if agent_extension.py is absent3412. Generate analysis.md (orchestrator writes via <prefix> tee), optional extension, embed PNG35```3637## Rules3839- **Subagents:** Use the Task tool **only** where reference.md says “subagent” (Steps **7**, **8**, **10**). The orchestrator runs everything else, including Step 8.5, using the command prefix from `<output_dir>/cache/cmd_prefix.txt` (`{CMD}` substitution).40- **Language:** Prefer vendor-agnostic terms (GPU kernels, collective communication, vendor GEMM library, DNN primitives, GPU graph). When quoting trace data, real kernel names are fine.41- **Subagent prompts:** Point each subagent at the checked-in agent file under `TraceLens/Agent/Analysis/skills/analysis-orchestrator/agents/<name>.md` (see reference.md for exact paths and prompt shells).4243## Primary outputs4445- **Deliverable:** `<output_dir>/analysis.md`46- **Internals:** `system_findings/`, `category_findings/`, `category_data/`, `metadata/`, `perf_report*.xlsx`, CSV folders — see package README for layout.4748## Agent layout4950Project subagents ship with this skill: `TraceLens/Agent/Analysis/skills/analysis-orchestrator/agents/*.md`.
amd