# Definitions | Cloud methodology
Source: https://greenpixie.com/docs/methodology/cloud/definitions

Key terms used throughout the methodology.

**Chiplet.** One of several small dies that together make up a processor package. Modern server CPUs pair compute chiplets, which hold the cores, with an IO die that holds the memory and PCIe controllers.

**Cloud Provider.** A company that offers cloud computing services and solutions to businesses and individuals. In this methodology, the term refers to AWS, Azure, and GCP.

**CO₂e.** Carbon dioxide equivalent. A metric that expresses the global warming potential of various greenhouse gases as the equivalent amount of carbon dioxide.

**Data Centre.** A facility composed of networked computers and storage used to process, store, and distribute data.

**Die.** In semiconductor manufacturing, a small block of semiconducting material on which a functional circuit is fabricated.

**Electricity Generation Water Consumption.** The water consumed during the generation of the electricity that a data centre draws, reflecting the regional generation mix. Classified as Scope 2.

**Electricity Grid.** The interconnected network for delivering electricity from producers to consumers, consisting of generating stations, transmission lines, and distribution lines.

**Embodied Emissions.** The greenhouse gas emissions associated with the manufacturing and transport of physical server components, together with estimated allowances for end-of-life processing and for the data centre shell.

**Emissions.** The discharge of greenhouse gases into the atmosphere, expressed in CO₂e.

**Expansion Card.** A small printed circuit board plugged into a server's motherboard rather than built onto it. The network interface card, the PCIe riser cards and the RAID controller card are counted as a fixed value per server.

**GHG.** Greenhouse gases. Gases that trap heat in the atmosphere and contribute to the greenhouse effect and global warming.

**GPU.** Graphics processing unit. In this methodology, any accelerator card counted separately from the CPU, including NVIDIA and AMD data centre GPUs, FPGAs and cloud providers' own inference and training chips.

**Grid Carbon Intensity.** The average emissions produced per unit of electricity consumed from a given electricity grid, varying by region and time. Abbreviated as GCI.

**HBM.** High bandwidth memory. Stacked memory placed beside a GPU die on the same package. Its capacity is the main driver of a GPU's embodied emissions after the die itself.

**Instance.** In cloud computing, a virtual server used for running applications on the cloud.

**IO Die.** The die in a chiplet processor that carries the memory controllers, PCIe lanes and other input and output logic. Counted in the CPU die area alongside the compute chiplets.

**IT Cooling Water Consumption.** The water consumed on site in cooling the IT equipment of a data centre, for example through cooling loops, evaporative cooling, and humidification. Classified as Scope 1.

**IT Electricity Usage.** The sum of electricity consumed by the four component categories within the IT equipment of a data centre, being computation, memory, storage, and networking. Abbreviated as ITEU.

**kWh.** Kilowatt-hour. The unit of energy consumption used to calculate associated GHG emissions based on the carbon intensity of the energy source.

**LCA.** Life Cycle Assessment. A method for assessing the environmental impacts associated with all stages of a product's life, from raw material extraction through manufacturing, use, and end of life.

**Location-Based Method.** A method for calculating carbon emissions based on the average emissions intensity of the electricity grid on which energy consumption occurs.

**NAND Generation.** The layer count of the 3D NAND flash in an SSD, such as 64, 128, 176 or 232 layers. A taller stack holds more data per cm² of die and costs more to make per cm², so both the die density and the die coefficient used for SSDs depend on it.

**Power Usage Effectiveness.** A ratio expressing the energy efficiency of a data centre, defined as Total Facility Energy divided by IT Electricity Usage. A PUE of 1.0 would indicate no energy overhead beyond the IT equipment itself. Abbreviated as PUE.

**Process Node.** The feature size of a chip's manufacturing process in nanometres, such as 5 nm or 14 nm. A newer node takes more lithography energy and more process gas per cm² of wafer, so the die coefficient used for CPUs and GPUs depends on it.

**Public Cloud.** A model of delivering computing services over the internet through third-party providers, typically on a pay-as-you-go basis, without the user owning or maintaining the underlying physical infrastructure.

**Recycled Content Method.** The GHG Protocol Product Standard's method for end-of-life accounting in which no credit is taken for material recycled at end of life. Used here where cloud hardware recycling is neither under Greenpixie's control nor visible to it.

**Semiconductor.** A material with controllable electrical conductivity used in creating electronic components such as transistors and integrated circuits in computer hardware.

**Server.** A computer or system that provides resources, data, services, or programs to other computers over a network.

**SKU.** Stock Keeping Unit. In cloud billing, a distinct combination of service, resource type, region, and pricing tier that appears as a line item on a cloud usage report.

**Thermal Design Power.** The maximum heat a chip is designed to produce under sustained load, in watts. Abbreviated as TDP. The cooling solution term of the GPU calculation scales with it.

**Total Facility Energy.** The total energy consumed by a data centre, comprising both IT Electricity Usage and facility overhead such as cooling, lighting, and power distribution. Abbreviated as TFE.

**Usage Emissions.** The greenhouse gas emissions arising from the electricity consumed during cloud operations, calculated by multiplying Total Facility Energy by Grid Carbon Intensity. Abbreviated as UE.

**vCPU.** Virtual central processing unit. The standard unit of compute capacity allocated to a cloud instance, representing a share of a physical processor core.

**Water Usage Effectiveness.** The ratio between the water consumed by data centre systems and the energy consumed by the IT equipment, expressed in litres per kilowatt-hour. Abbreviated as WUE.