GPUServer

Params

name

A human readable description of the object.

server_type

Server type of on premise gpu server.

For example, on-premise.

gpu_power

On premise gpu server gpu power from estimating the carbon footprint of bloom in watt / gpu.

gpu_idle_power

On premise gpu server gpu idle power from estimating the carbon footprint of bloom in watt / gpu.

ram_per_gpu

On premise gpu server ram per gpu from estimating the carbon footprint of bloom in gigabyte_ram / gpu.

carbon_footprint_fabrication_per_gpu

On premise gpu server carbon footprint one gpu from estimating the carbon footprint of bloom in kilogram / gpu.

average_carbon_intensity

Average carbon intensity of on premise gpu server electricity in gram / kilowatt_hour.

compute

Nb gpus of on premise gpu server in gpu.

carbon_footprint_fabrication_without_gpu

On premise gpu server carbon footprint without gpu from estimating the carbon footprint of bloom in kilogram.

lifespan

Lifespan of on premise gpu server in year.

power_usage_effectiveness

Pue of on premise gpu server in dimensionless.

utilization_rate

On premise gpu server utilization rate in dimensionless.

base_compute_consumption

Base gpu consumption of on premise gpu server in gpu.

base_ram_consumption

Base ram consumption of on premise gpu server in gigabyte_ram.

storage

An instance of Storage.

fixed_nb_of_instances

User defined number of on premise gpu server instances. Can be an EmptyExplainableObject in which case the optimum number of instances will be computed, or an ExplainableQuantity with a dimensionless value, in which case e-footprint will raise an error if the object needs more instances than available.

Backwards links

GPUJob

Calculated attributes

carbon_footprint_fabrication

ExplainableQuantity in kilogram, representing the On premise gpu server carbon footprint fabrication.

Example value: 3.1 t

Depends directly on:

through the following calculations:

You can also visit the link to on premise GPU server carbon footprint fabrication’s full calculation graph.

power

ExplainableQuantity in watt, representing the On premise gpu server power.

Example value: 1.6 kW

Depends directly on:

through the following calculations:

You can also visit the link to on premise GPU server power’s full calculation graph.

idle_power

ExplainableQuantity in watt, representing the On premise gpu server idle power.

Example value: 200 W

Depends directly on:

through the following calculations:

You can also visit the link to on premise GPU server idle power’s full calculation graph.

ram

ExplainableQuantity in gigabyte_ram, representing the On premise gpu server ram.

Example value: 2560 B

Depends directly on:

through the following calculations:

You can also visit the link to on premise GPU server RAM’s full calculation graph.

hour_by_hour_ram_need

On premise gpu server hour by hour ram need in gigabyte_ram.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in k:
first 10 vals [222, 889, 1000, 889, 333, 111, 778, 667, 1000, 222],
last 10 vals [333, 889, 1000, 222, 889, 667, 556, 222, 667, 333]

Depends directly on:

through the following calculations:

You can also visit the link to on premise GPU server hour by hour ram need’s full calculation graph.

hour_by_hour_compute_need

On premise gpu server hour by hour compute need in gpu.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in gpu:
first 10 vals [0.000556, 0.00222, 0.0025, 0.00222, 0.000833, 0.000278, 0.00194, 0.00167, 0.0025, 0.000556],
last 10 vals [0.000833, 0.00222, 0.0025, 0.000556, 0.00222, 0.00167, 0.00139, 0.000556, 0.00167, 0.000833]

Depends directly on:

through the following calculations:

You can also visit the link to on premise GPU server hour by hour compute need’s full calculation graph.

occupied_ram_per_instance

ExplainableQuantity in gigabyte_ram, representing the Occupied ram per on premise gpu server instance including services.

Example value: 0 GB_ram

Depends directly on:

Base RAM consumption of on premise GPU server from e-footprint hypothesis

through the following calculations:

You can also visit the link to Occupied RAM per on premise GPU server instance including services’s full calculation graph.

occupied_compute_per_instance

ExplainableQuantity in gpu, representing the Occupied cpu per on premise gpu server instance including services.

Example value: 0 gpu

Depends directly on:

Base gpu consumption of on premise GPU server from e-footprint hypothesis

through the following calculations:

You can also visit the link to Occupied CPU per on premise GPU server instance including services’s full calculation graph.

available_ram_per_instance

ExplainableQuantity in gigabyte_ram, representing the Available ram per on premise gpu server instance.

Example value: 2300 B

Depends directly on:

through the following calculations:

You can also visit the link to Available RAM per on premise GPU server instance’s full calculation graph.

available_compute_per_instance

ExplainableQuantity in gpu, representing the Available cpu per on premise gpu server instance.

Example value: 3.6 gpu

Depends directly on:

through the following calculations:

You can also visit the link to Available CPU per on premise GPU server instance’s full calculation graph.

raw_nb_of_instances

Hourly raw number of on premise gpu server instances in concurrent.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in :
first 10 vals [0.000154, 0.000617, 0.000694, 0.000617, 0.000231, 0.0000772, 0.00054, 0.000463, 0.000694, 0.000154],
last 10 vals [0.000231, 0.000617, 0.000694, 0.000154, 0.000617, 0.000463, 0.000386, 0.000154, 0.000463, 0.000231]

Depends directly on:

through the following calculations:

You can also visit the link to Hourly raw number of on premise GPU server instances’s full calculation graph.

nb_of_instances

Hourly number of on premise gpu server instances in concurrent.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in :
first 10 vals [1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
last 10 vals [1, 1, 1, 1, 1, 1, 1, 1, 1, 1]

Depends directly on:

through the following calculations:

You can also visit the link to Hourly number of on premise GPU server instances’s full calculation graph.

instances_fabrication_footprint

Hourly on premise gpu server instances fabrication footprint in kilogram.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in g:
first 10 vals [58.9, 58.9, 58.9, 58.9, 58.9, 58.9, 58.9, 58.9, 58.9, 58.9],
last 10 vals [58.9, 58.9, 58.9, 58.9, 58.9, 58.9, 58.9, 58.9, 58.9, 58.9]

Depends directly on:

through the following calculations:

You can also visit the link to Hourly on premise GPU server instances fabrication footprint’s full calculation graph.

instances_energy

Hourly energy consumed by on premise gpu server instances in kilowatt_hour.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in Wh:
first 10 vals [240, 241, 241, 241, 240, 240, 241, 241, 241, 240],
last 10 vals [240, 241, 241, 240, 241, 241, 241, 240, 241, 240]

Depends directly on:

through the following calculations:

You can also visit the link to Hourly energy consumed by on premise GPU server instances’s full calculation graph.

energy_footprint

Hourly on premise gpu server energy footprint in kilogram.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in g:
first 10 vals [24, 24.1, 24.1, 24.1, 24, 24, 24.1, 24.1, 24.1, 24],
last 10 vals [24, 24.1, 24.1, 24, 24.1, 24.1, 24.1, 24, 24.1, 24]

Depends directly on:

through the following calculations:

You can also visit the link to Hourly on premise GPU server energy footprint’s full calculation graph.

job_repartition_weights

Dictionary with GPUJob as keys and Manually defined gpu job weight in on premise gpu server impact repartition as values, in concurrent.

Example value: {
GPUJob Manually defined GPU job (ca3955): 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in :
first 10 vals [0.000139, 0.000556, 0.000625, 0.000556, 0.000208, 0.0000695, 0.000486, 0.000417, 0.000625, 0.000139],
last 10 vals [0.000208, 0.000556, 0.000625, 0.000139, 0.000556, 0.000417, 0.000347, 0.000139, 0.000417, 0.000208],
}

Depends directly on:

through the following calculations:

You can also visit the link to Manually defined GPU job weight in on premise GPU server impact repartition’s full calculation graph.

fabrication_impact_repartition_weight_sum

Sum of on premise gpu server fabrication impact repartition weights in concurrent.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in :
first 10 vals [0.000139, 0.000556, 0.000625, 0.000556, 0.000208, 0.0000695, 0.000486, 0.000417, 0.000625, 0.000139],
last 10 vals [0.000208, 0.000556, 0.000625, 0.000139, 0.000556, 0.000417, 0.000347, 0.000139, 0.000417, 0.000208]

Depends directly on:

Manually defined GPU job weight in on premise GPU server impact repartition

through the following calculations:

You can also visit the link to Sum of on premise GPU server fabrication impact repartition weights’s full calculation graph.

fabrication_impact_repartition

Dictionary with GPUJob as keys and On premise gpu server fabrication impact attribution to manually defined gpu job as values, in concurrent.

Example value: {
GPUJob Manually defined GPU job (ca3955): 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in :
first 10 vals [1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
last 10 vals [1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
}

Depends directly on:

through the following calculations:

You can also visit the link to on premise GPU server fabrication impact attribution to Manually defined GPU job’s full calculation graph.

usage_impact_repartition_weight_sum

Sum of on premise gpu server usage impact repartition weights in concurrent.

Example value: 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in :
first 10 vals [0.000139, 0.000556, 0.000625, 0.000556, 0.000208, 0.0000695, 0.000486, 0.000417, 0.000625, 0.000139],
last 10 vals [0.000208, 0.000556, 0.000625, 0.000139, 0.000556, 0.000417, 0.000347, 0.000139, 0.000417, 0.000208]

Depends directly on:

Manually defined GPU job weight in on premise GPU server impact repartition

through the following calculations:

You can also visit the link to Sum of on premise GPU server usage impact repartition weights’s full calculation graph.

usage_impact_repartition

Dictionary with GPUJob as keys and On premise gpu server usage impact attribution to manually defined gpu job as values, in concurrent.

Example value: {
GPUJob Manually defined GPU job (ca3955): 26298 values from 2024-12-31 23:00:00+00:00 to 2028-01-01 17:00:00+00:00 in :
first 10 vals [1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
last 10 vals [1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
}

Depends directly on:

through the following calculations:

You can also visit the link to on premise GPU server usage impact attribution to Manually defined GPU job’s full calculation graph.