Understanding Metrics¶

This guide explains all the metrics collected by A-LEMS and what they mean for your research.

📊 Core Energy Metrics¶

Energy Measurements¶

Metric	Unit	Description
`pkg_energy_uj`	µJ	Total package energy (raw)
`core_energy_uj`	µJ	Core energy (raw)
`uncore_energy_uj`	µJ	Uncore energy (cache, memory controller, I/O)
`dram_energy_uj`	µJ	DRAM energy (if available)
`total_energy_uj`	µJ	Raw package energy
`dynamic_energy_uj`	µJ	Workload energy (raw - idle)
`baseline_energy_uj`	µJ	Idle energy for same duration

Derived Energy Metrics¶

Metric	Formula	Meaning
`workload_energy`	`package - idle`	Energy actually used by your workload
`reasoning_energy`	`core - idle_core`	Energy for actual computation
`orchestration_tax`	`workload - reasoning`	Overhead of agentic orchestration
`energy_per_token`	`workload / tokens`	Energy efficiency per token
`energy_per_instruction`	`workload / instructions`	Energy per CPU instruction

🎯 Orchestration Tax¶

The orchestration tax is A-LEMS's core metric:

tax = agentic_energy / linear_energy
tax_percent = (agentic - linear) / agentic * 100

Interpretation:

Tax Value	Meaning
1.0x	No overhead (rare)
1.5x	50% more energy
2.0x	2× more energy
5.0x+	High orchestration overhead

Example from real data:

Linear: 1.2 J
Agentic: 2.6 J
Tax: 2.2x (120% more energy)

⚡ Power Metrics¶

Metric	Unit	Description
`avg_power_watts`	W	Average power during run
`package_power`	W	Instantaneous package power
`core_power`	W	Instantaneous core power
`dram_power`	W	DRAM power (if available)

Power curves from energy_samples show how power changes over time:

Power (W)
25 ├─────────────────────•────
20 │ •───•
15 │ •──•
10 │ •──•
5  │ •──•
0  └────•────•────•────•────•────•
    0   2   4   6   8   10 Time (s)

💻 Performance Counters¶

Metric	Description	Good Value
`ipc`	Instructions Per Cycle	> 2.0
`cache_miss_rate`	LLC cache miss rate	< 5%
`instructions`	Total instructions executed	N/A
`cycles`	Total CPU cycles	N/A
`page_faults`	Memory page faults	Low

IPC (Instructions Per Cycle) indicates how efficiently the CPU is used:

< 1.0: Memory-bound or stalled
1.0 - 2.0: Mixed workload
> 2.0: Compute-bound, efficient

⏱️ Timing Metrics¶

Metric	Unit	Description
`duration_ns`	ns	Total run duration
`planning_time_ms`	ms	Planning phase (agentic only)
`execution_time_ms`	ms	Tool execution phase
`synthesis_time_ms`	ms	Response synthesis phase
`api_latency_ms`	ms	Time waiting for API
`compute_time_ms`	ms	Actual computation time
`waiting_time_ms`	ms	Time between LLM calls

Phase ratios for agentic workflows:

Planning: 2.3s (30%)
Execution: 4.1s (54%)
Synthesis: 1.2s (16%)
Total: 7.6s

🌡️ Thermal Metrics¶

Metric	Unit	Description
`package_temp_celsius`	°C	CPU package temperature
`start_temp_c`	°C	Temperature at run start
`max_temp_c`	°C	Peak temperature
`thermal_delta_c`	°C	Temperature rise (max - start)
`thermal_gradient`	°C/s	Rate of temperature change

Thermal thresholds:

< 60°C: Normal operation
60-80°C: Warm, still efficient
80-95°C: Hot, possible throttling
> 95°C: Thermal throttling active

Thermal Profile Example¶

🔄 C-State Metrics¶

Metric	Description	Power Savings
`c2_time_seconds`	Time in C2 (light sleep)	Moderate
`c3_time_seconds`	Time in C3 (deeper sleep)	High
`c6_time_seconds`	Time in C6 (very deep)	Very high
`c7_time_seconds`	Time in C7 (package sleep)	Maximum

C-state residency shows how efficiently the CPU enters low-power states during idle periods.

📊 Scheduler Metrics¶

Metric	Description	High Value Indicates
`context_switches_voluntary`	Thread yielding	Normal operation
`context_switches_involuntary`	Forced preemption	Contention
`thread_migrations`	CPU hopping	Poor cache locality
`run_queue_length`	Runnable processes	System load
`interrupt_rate`	Interrupts per second	I/O activity

🧠 Agentic Metrics¶

Metric	Description	Typical Range
`llm_calls`	Number of LLM invocations	1-10
`tool_calls`	Number of tool executions	0-5
`steps`	Total workflow steps	1-15
`complexity_level`	1-3 scale	Task difficulty
`complexity_score`	1-10 scale	Normalized complexity

🌍 Sustainability Metrics¶

Metric	Unit	Description
`carbon_g`	g CO₂	Carbon footprint
`water_ml`	ml	Water consumption
`methane_mg`	mg	Methane emissions

Country-specific factors are applied based on grid intensity:

Country	Carbon (g/kWh)	Water (ml/kWh)
US	0.389	2.1
IN	0.708	3.4
FR	0.055	1.2
CN	0.555	2.8

📈 Efficiency Metrics¶

Metric	Formula	Good Value
Energy per token	`workload / tokens`	< 0.01 J/token
Energy per instruction	`workload / instructions`	< 1e-9 J/inst
Instructions per token	`instructions / tokens`	> 1000
Interrupts per second	`interrupt_rate`	< 5000

🔍 Sample Queries¶

Get All Metrics for a Run¶

SELECT * FROM runs WHERE run_id = 977;

Compare Linear vs Agentic¶

SELECT 
    r.workflow_type,
    AVG(r.dynamic_energy_uj/1e6) as avg_energy_j,
    AVG(r.duration_ns/1e9) as avg_duration_s,
    AVG(r.ipc) as avg_ipc
FROM runs r
WHERE r.exp_id = 185
GROUP BY r.workflow_type;

Find High Tax Experiments¶

SELECT 
    e.exp_id,
    e.task_name,
    ots.tax_percent
FROM orchestration_tax_summary ots
JOIN runs r ON ots.linear_run_id = r.run_id
JOIN experiments e ON r.exp_id = e.exp_id
WHERE ots.tax_percent > 200
ORDER BY ots.tax_percent DESC;

📊 Metric Categories Summary¶

Category	Key Metrics	Use For
Energy	workload, reasoning, tax	Core research
Performance	ipc, cache_miss_rate	CPU efficiency
Timing	phase times, latency	Bottleneck analysis
Thermal	temperature, delta	Cooling analysis
C-State	residency times	Power management
Scheduler	context switches	OS overhead
Agentic	llm_calls, steps	Workflow complexity
Sustainability	carbon, water	Environmental impact