Energy consumption
Energy consumption is the amount of energy used by a system or process over a period of operation; power consumption is the rate at which energy is used.
Introduction to Generative AI, Second Edition
Energy consumption is the amount of electrical or other energy consumed by a computational process, system, or activity.
- How it is used
- Frames energy consumption as an environmental cost of training and deploying large language models.
- How deep it goes
- Explains electricity use, emissions uncertainty, cooling, and measurement tools at a high level, without a detailed measurement procedure.
- Examples
- Cites an estimated 500 metric tons of carbon dioxide from GPT-3 training and describes CodeCarbon as a measurement tool.
- What you can do after
- Evaluate environmental costs and uncertainty when considering large language model development and deployment.
Read how Introduction to Generative AI, Second Edition explains energy consumption →
The Art of Code
Energy consumption is the amount of electrical or other energy required to operate software and its supporting computing systems.
- How it is used
- Uses energy consumption to assess software’s sustainability alongside responsiveness and infrastructure use.
- How deep it goes
- Distinguishes static and dynamic analysis and names profiling tools, but does not detail an energy-measurement procedure.
- Examples
- Uses a PostgreSQL customer-table exercise and compares server electricity use with lifecycle emissions.
- What you can do after
- Evaluate energy consumption alongside performance, storage needs, and hardware lifecycle impacts.
Digital Twins in Action
Electricity consumption is the amount of electrical energy consumed by a system or facility.
- How it is used
- Connects home electricity measurements to decisions about solar self-consumption, load scheduling, and heating and cooling.
- How deep it goes
- Develops decision-linked data requirements, sensor selection, and an AutoML forecasting workflow.
- Examples
- Uses a home with whole-house and appliance sensors, solar generation, and grid imports and exports.
- What you can do after
- Apply it via AutoGluon to forecast grid imports using environmental conditions and inform electricity-use decisions.
Read how Digital Twins in Action explains energy consumption →
Machines that Think
The amount of energy required by a system or process while it operates.
- How it is used
- Frames energy consumption as a constraint on large-model training, deployment, and the viability of applications.
- How deep it goes
- Explains energy drivers and design trade-offs, using broad scale estimates rather than precise model-specific measurements.
- Examples
- Compares brain power draw with large-model training and estimates training-run energy in gigawatt-hours.
- What you can do after
- Evaluate energy trade-offs in training and deployment without mistaking per-operation efficiency for lower total consumption.
Rearchitecting LLMs
Energy consumption is the quantity of electrical energy used by a computation or hardware system.
- How it is used
- Evaluates the energy cost of inference after width pruning, alongside generation speed.
- How deep it goes
- Reports per-token energy measurements and comparative results, while deferring evaluation-system details to appendices.
- Examples
- Compares base and pruned models at 1.322 versus 1.236–1.239 joules per token.
- What you can do after
- Evaluate pruning variants by comparing measured joules per generated token as comparative baselines.
Build a Text-to-Image Generator (from Scratch)
The amount of electrical energy consumed by computing processes such as model training and inference.
- How it is used
- Frames energy consumption as an environmental cost and scalability constraint in text-to-image model development and deployment.
- How deep it goes
- Explains resource drivers and mitigation techniques but stops short of implementation details or quantitative energy measurement.
- Examples
- Uses large datasets, powerful hardware, efficiency techniques, and a projected 2040 global-energy scenario as concrete reference points.
- What you can do after
- Evaluate computational-footprint trade-offs involving model quality, scalability, and environmental impact.
Read how Build a Text-to-Image Generator (from Scratch) explains energy consumption →
The AI Pocket Book
The amount of electrical energy used by AI systems during training or inference.
- How it is used
- Frames electricity consumption as a practical environmental and resource-cost consideration for AI training and use.
- How deep it goes
- Defines the issue, quantifies model-use examples, and flags inconsistent measurement without deriving energy estimates.
- Examples
- Uses text generation, image generation, coding autocompletions, search queries, household consumption, and petrol-car emissions.
- What you can do after
- Evaluate AI energy claims by checking usage scale, units, model differences, and whether estimates are direct or indirect.
Also covered in
- Fast by DefaultIn early access (MEAP)
- DuckDB in Action
- Inside AI
- Rust in Action: Systems programming concepts and techniques
What to know first
- language modelA language model is a computational model that represents patterns in linguistic sequences to assign probabilities, predict, generate, or evaluate language.
- computeCompute is the processing capacity, resources, or computational work involved in executing software and other computational workloads.
- hardwareHardware is the physical equipment and components of a computing system, including processors, memory, storage, and peripherals.
- optimizationOptimization is the systematic selection or modification of values, configurations, or implementations to improve a defined objective, often subject to constraints.
Related concepts
- CodeCarbonmeasured by · CodeCarbon is an open-source Python library that estimates the energy consumption and carbon-dioxide-equivalent emissions of computational workloads.
- language modelmeasured by · A language model is a computational model that represents patterns in linguistic sequences to assign probabilities, predict, generate, or evaluate language.
- heat mapused by · A heat map is a two-dimensional visualization that encodes numeric values, frequencies, or activity using color intensity or hue.
- computeuses · Compute is the processing capacity, resources, or computational work involved in executing software and other computational workloads.
Questions about energy consumption
What is energy consumption?
Energy consumption is the amount of energy used by a system or process over a period of operation; power consumption is the rate at which energy is used.
Which Manning books explain energy consumption?
Introduction to Generative AI, Second Edition, The Art of Code, Digital Twins in Action, Machines that Think, Rearchitecting LLMs, Build a Text-to-Image Generator (from Scratch) and The AI Pocket Book.
What should I know before learning energy consumption?
Manning books that teach energy consumption usually introduce language model, compute, hardware and optimization first.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.






