Smart complete CPU based on CPU name

[da-ekchajzer]

Problem

CPU family are used has a proxy to get the die_size_per_core of the the CPU. CPU model will be used has a proxy to get the consumption profile of the CPU. CPU family are not trivial to get for all users. We should provide a way to smart complete CPU family and model from a CPU name.

Solution

Cloud carbon footprint provides a dataset of CPU name for each CPU family. We could add :

• A name attribute to CPU objects
• A model_range attribute to CPU objects
• A model_id attribute to CPU objects
• A step in CPU smart_complete function to translate (when available) a CPU name into a CPU family and to extract CPU model from the name.
• We should provide a way to find the associate number of core

'Translation' shouldn't be strict since CPU name format given by a user can be a little different from our data format

Example

Name : Intel Xeon Gold 6142
Manufacturer : Intel
model_range : Xeon Gold
Family : skylake // from /intel-skylake.csv
nb_cores : TODO ?

Name : Intel Xeon W-2133
Manufacturer : Intel
model_range : Xeon W
Family : skylake // from /intel-skylake.csv
nb_cores : TODO ?

Name : Intel Core i9-9820X
model_range : Intel
Model : Core i9
Family : skylake // from /intel-skylake.csv
nb_cores : TODO ?

Name : AMD RYZEN 3 1200
Manufacturer : AMD
model_range : RYZEN 3
Family : TODO ?
nb_cores : TODO ?

Name : AMD RYZEN 3 1200
Manufacturer : AMD
model_range : RYZEN 3
Family : TODO ?
nb_cores : TODO ?

Additional context or elements

Intel

• Skylake : https://github.com/cloud-carbon-footprint/cloud-carbon-coefficients/blob/main/data/intel-skylake.csv
• Sandybridge : https://github.com/cloud-carbon-footprint/cloud-carbon-coefficients/blob/main/data/intel-sandybridge.csv
• Ivybridge : https://github.com/cloud-carbon-footprint/cloud-carbon-coefficients/blob/main/data/intel-ivybridge.csv
• haswell : https://github.com/cloud-carbon-footprint/cloud-carbon-coefficients/blob/main/data/intel-haswell.csv
• Coffeelake : https://github.com/cloud-carbon-footprint/cloud-carbon-coefficients/blob/main/data/intel-coffeelake.csv
• Cascadelake : https://github.com/cloud-carbon-footprint/cloud-carbon-coefficients/blob/main/data/intel-cascadelake.csv
• Broadwell : https://github.com/cloud-carbon-footprint/cloud-carbon-coefficients/blob/main/data/intel-broadwell.csv

AMD

• Naple : https://github.com/cloud-carbon-footprint/cloud-carbon-coefficients/blob/main/data/amd-epyc-gen1.csv
• Rome : https://github.com/cloud-carbon-footprint/cloud-carbon-coefficients/blob/main/data/amd-epyc-gen2.csv
• Milan : https://github.com/cloud-carbon-footprint/cloud-carbon-coefficients/blob/main/data/amd-epyc-gen3.csv

See https://doc.api.boavizta.org/Explanations/smart-complete/

[da-ekchajzer]

https://docs.google.com/spreadsheets/d/1Ev6BIDQ0sWkMIfA5qLuO_AD81VVWvl2r8O2ohQYqUKs/edit?usp=sharing

[da-ekchajzer]

Implemented in #113 thank you, @benjaminlebigot and @samuelrince