to

Energy & Carbon Trend

Energy by Pipeline Stage

Insights

AMD Runners (x86)

Mac Runners (ARM)

Mac vs AMD Energy by Stage

Carbon Intensity by Platform

Carbon Intensity Over Time

Carbon per Scenario

SCI Formula (ISO/IEC 21031)

SCI = (E × I) + M
  • E: Energy consumed (kWh)
  • I: Grid intensity (gCO₂eq/kWh)
  • M: Embodied carbon (ignored)
Showing...

Energy by Span Type

Verification Proof (Methodology)

Phase 1 (1M items): sort_operation ~1.76 J

Phase 2 (5M items): sort_operation ~14.85 J

O(n log n) scaling proven


Infrastructure Tax remains flat.

1

Pick a scenario

Each CI/CD stage is profiled with the same baseline/workload/cooldown protocol. Five scenarios, two architectures.

2

Profiling timeline (animated)

10 s baseline → workload → 10 s cooldown. The profiler runs end-to-end at 100 ms sampling. Press play to see how data fills the window.

t = 0.0 s
Samples: 0
Pre-baseline 10 s
Workload
Cooldown 10 s
START_TIME
END_TIME
0 s    
Profiler running, system idle — first 100 samples become Pidle Unmodified BookStack command runs inside Docker Profiler captures thermal decay — bounds the timestamp arithmetic
3

Baseline subtraction — see it happen

The raw power trace mixes the workload's own energy with the runner's idle background. Phase 3 strips the background out. Step through the four operations the analyst performs on every CSV.

P_idle

Click 1. Raw trace to begin.

Samples
Pidle
Peak active
Esoftware

Synthetic trace for illustration — the analyst applies the same four operations to every real timechart.csv. Real runs at ResearchData/cpudata/ use 100 ms sampling and the live runner's idle level.

4

The 5 pipeline gates — click any gate to expand

Every run passes through these five gates in order. Each gate has a fail-fast safeguard that blocks bad data from entering the research database.

5

What we keep — click an artefact to see a real sample

Per scenario we persist 4–5 files. Together they let anyone reconstruct, audit, and re-process the run from scratch.

6

W3C trace context — the seam between power data and span data

Hover any part of the traceparent below to see its role.

00 - 168962b8fd5b4ab9b459644 3eb33ae61 - 1ed8bf1bc786403c - 01
Hover any segment above …

Phase 1: Hardware Acquisition

AMD µProf: 100 ms polling, socket0-package-power (W)

Apple powermetrics: 100 ms polling, CPU+GPU+ANE combined

Nyquist constraint limits resolution of spikes < 50 ms — explicitly disclosed.

Phase 3: Signal Deconvolution

P_idle = (1/100) × Σ P_i
E_software = Σ max(0, P_sample − P_idle) × 0.1 s

Phase 4: Temporal Correlation

overlap_ratio = span_duration_in_interval / 100 ms
E_span += P_sample × overlap_ratio × 0.1 s

Conservation of energy: E_Total = E_Spans + E_Infrastructure

Validation Packet

Full methodology audit, validity threats, and reviewer checklist:

Database Schema & Relations

Loading schema…

Relationship Explorer

Pick a pipeline run to see how one energy_metrics row links to its child span_metrics rows (joined on project_name · run_id · scenario).

Table Data (all rows)

Download Full Dataset (CSV)

Download = one unified dataset (energy_metrics joined to span_metrics), not separate files.

Loading...
Project Run ID Scenario Prefix (bucket area) Size Last Modified Download
Loading...

Drag & drop a CSV file here

or click to browse

Awaiting connection...

Display Settings

  • Timezone:

Database Stats

  • Loading...

Research Team

  • Project: A Multi-Agent Framework for Request-Level Energy Attribution
  • Member 1: M.M.N.D. Seneviratne (TG/2021/1017)
  • Member 2: B.D.D. Devendra (TG/2021/1038)
  • Supervisor: Dr. Chandana Pushpakumara