Comments (2)
da-ekchajzer
commented on September 24, 2024
Methodology
This methodology is a bottom-up approach to evaluate the manufacture impact of a server from the impact of its components.
The impact of CPU, RAM and SSD are evaluated depending on their die density (which as been found as the most influent variable). The impact of the other components are evaluated directly with emission factors.
Equations
For CPU
GWP_Manuf_CPU (kgCO2eq) = CPUNumber (Unit) x ((CPUCoreNumber (unit) x DieSizePerCore (cm2) + 0.491) x CPUDieGWPImpact (kg CO2eq) + CPUGWPImpact (kgCO2eq) )
ADP_Manuf_CPU (kgSbeq)= CPUNumber (Unit) x ((CPUCoreNumber (unit) x DieSizePerCore (cm2) + 0.491) x CPUDieADPImpact (kg Sbeq) + CPUADPImpact (kgSbeq) )
PE_Manuf_CPU (MJ)= CPUNumber (Unit) x ((CPUCoreNumber (unit) x DieSizePerCore (cm2) + 0.491) x CPUDiePEImpact (MJ) + CPUPEImpact (MJ) )
with:
CPUNumber : server data
CPUCoreNumber : component data (cpu)
DieSizePerCore : component data (cpu) or 0.245 cm2 (default)
CPUDie[*]PImpact : impact factor
CPU[*]Impact : impact factor
For RAM
GWP_Manuf_RAM (kgCO2eq) = RAMModuleNumber (Unit) x ((RAMSize (Go) / RAMStorageDensity (Go/cm2)) x RAMDieGWPImpact (kg CO2eq) + RAMModuleGWPImpact (kgCO2eq) )
ADP_Manuf_RAM (kg Sb eq) = RAMModuleNumber (Unit) x ((RAMSize (Go) / RAMStorageDensity (Go/cm2)) x RAMDieADPImpact (kg Sb eq) + RAMModuleADPImpact (kg Sb eq) )
PE_Manuf_RAM (MJ) = RAMModuleNumber (Unit) x ((RAMSize (Go) / RAMStorageDensity (Go/cm2)) x RAMDiePEImpact (MJ) + RAMModulePEImpact (MJ) )
with :
RAMModuleNumber : server data
RAMSize : component data (ram)
RAMStorageDensity : component data (ram) or 1.79Go/cm2 (default)
RAMDie[*]Impact : impact factor
RAMModule[*]Impact : impact factor
For SSD
GWP_Manuf_SSD (kgCO2eq) = SSDDiskNumber (Unit) x ((SSDDiskSize (Go) / SSDStorageDensity (Go/cm2)) x SSDDieGWPImpact (kg CO2eq) + SSDDiskGWPImpact (kgCO2eq) )
ADP_Manuf_SSD (kg Sb eq) = SSDDiskNumber (Unit) x ((SSDDiskSize (Go) / SSDStorageDensity (Go/cm2)) x SSDDieADPImpact (kg Sb eq) + SSDDiskGWPImpact (kg Sb eq) )
PE_Manuf_SSD (MJ) = SSDDiskNumber (Unit) x ((SSDDiskSize (Go) / SSDStorageDensity (Go/cm2)) x SSDDieEPImpact (MJ) + SSDDiskEPImpact (MJ) )
with:
SSDDiskNumber : server data
SSDDiskSize : component data (ssd)
SSDStorageDensity : component data (ssd) or 50.6 Go/cm2 (default)
SSDDie[*]Impact : impact factor
SSDDisk[*]Impact : impact factor
For the all server
GWP_Manuf_Server (kgCO2eq) = GWP_Manuf_CPU + GWP_Manuf_RAM + GWP_Manuf_SSD + HDDDriveNumber*HDDGWPImpact + MotherBoardGWPImpact + PowerSupplyUnitNumber*PowerSupplyUnitWeight*PSUGWPImpact + ServerAssemblyGWPImpact + RackGWPImpact (si serveur rack) + BladeGWPImpact + 1/16 (BladeSlotsGWPImpact) (if blade server)
ADP_Manuf_Server (kgCO2eq) = ADP_Manuf_CPU + ADP_Manuf_RAM + ADP_Manuf_SSD + HDDDriveNumber*HDDADPImpact + MotherBoardADPImpact + PowerSupplyUnitNumber*PowerSupplyUnitWeight*PSUADPImpact + ServerAssemblyADPImpact + RackADPImpact (if server rack) + BladeADPImpact + 1/16 (BladeSlotsADPImpact) (if blade server)
EP_Manuf_Server (kgCO2eq) = EP_Manuf_CPU + EP_Manuf_RAM + EP_Manuf_SSD + HDDDriveNumber*HDDEPImpact + MotherBoardEPImpact + PowerSupplyUnitNumber*PowerSupplyUnitWeight*PSUEPImpact + ServerAssemblyEPImpact + RackEPImpact (si serveur rack) + BladeEPImpact + 1/16 (BladeSlotsEPImpact) (if blade server)
with:
[*]_Manuf_CPU : equation
[*]_Manuf_RAM : equation
[*]_Manuf_SSD : equation
HDD[*]Impact : impact factor
MotherBoard[*]Impact : impact factor
PSU[*]Impact : impact factor
ServerAssembly[*]Impact : impact factor
Rack[*]Impact : impact factor
Blade[*]Impact : impact factor
BladeSlots[*]Impact : impact factor
HDDDriveNumber : server data
PowerSupplyUnitNumber : server data
PowerSupplyUnitWeight : component data
Data
Impact factor
| Composant | Variable (GWP) | Unité | ADP (kgSbeq) | GWP (kgCO2eq) | PE (MJ) |
|---|---|---|---|---|---|
| CPU | CPUGWPImpact | Unit | 2.04E-02 | 9.14 | 156.00 |
| CPU Die | CPUDieGWPImpact | 1 cm2 | 5.80E-07 | 1.97 | 26.50 |
| RAM Module | Unit | 1.69E-03 | 5.22 | 74.00 | |
| RAM Die | 1 cm2 | 6.30E-05 | 2.20 | 27.30 | |
| SSD (excluding die) Disk | Unit | 5.63E-04 | 6.34 | 76.90 | |
| SSD Die | 1 cm2 | 6.30E-05 | 2.20 | 27.30 | |
| HDD Disk | Unit | 2.50E-04 | 31.10 | 276.00 | |
| Motherboard | Unit | 3.69E-03 | 66.10 | 836.00 | |
| Rack Server | Unit | 2.02E-02 | 150.00 | 2 200.00 | |
| Blade 16 Slots | Unit | 4.32E-01 | 880.00 | 12 700.00 | |
| Blade Server | Unit | 6.72E-04 | 30.90 | 435.00 | |
| Server Assembly | Unit | 1.41E-06 | 6.68 | 68.60 | |
| Power Supply Unit | kg | 8.30E-03 | 24.30 | 352.00 |
Components
CPU
| CPU Family | Introduction Year | Process (nm) | Die size (mm2) | Core Number | Size/Core (mm2) |
|---|---|---|---|---|---|
| Skylake | 2017 | 14 | 694 | 28 | 24.8 |
| Skylake | 2017 | 14 | 485 | 18 | 26.9 |
| Skylake | 2017 | 14 | 325 | 10 | 32.5 |
| Coffee Lake | 2017 | 14 | 149 | 6 | 24.8 |
| Coffee Lake | 2017 | 14 | 174 | 8 | 21.8 |
| Broadwell | 2014 | 14 | 456 | 24 | 19.0 |
| Broadwell | 2014 | 14 | 306 | 14 | 21.9 |
| Broadwell | 2014 | 14 | 246 | 10 | 24.6 |
| Haswell | 2013 | 22 | 622 | 18 | 34.6 |
| Ivy Bridge | 2011 | 22 | 160 | 4 | 40.0 |
| Ivy Bridge | 2011 | 22 | 257 | 6 | 42.8 |
| Ivy Bridge | 2011 | 22 | 341 | 10 | 34.1 |
| Ivy Bridge | 2011 | 22 | 541 | 15 | 36.1 |
| Sandy Bridge | 2010 | 32 | 216 | 4 | 54.0 |
RAM
| Constructeur | Architecture | Go/cm2 |
|---|---|---|
| Samsung | 30nm | 0.625 |
| Samsung | 25nm | 1.25 |
| Samsung | 20nm | 1.75 |
| Samsung | 18nm | 2.38 |
| SK hynix | 30nm | 0.750 |
| SK hynix | 26nm | 1.00 |
| SK hynix | 21nm | 1.31 |
| SK hynix | 21nm | 1.88 |
| Micron | 30nm | 0.750 |
| Micron | 30nm | 0.875 |
| Micron | 20nm | 1.13 |
| Micron | 20nm | 1.13 |
SSD
| Constructeur | Densité de stockage (Go/cm2) |
|---|---|
| Micron | 49.6 |
| Toshiba | 48.5 |
| Samsung | 53.6 |
from boaviztapi.
da-ekchajzer
commented on September 24, 2024
Exemple of post request
{
"model":{
"brand": "Dell",
"name": "R740",
"type": "rack",
"year": 2020
},
"configuration": {
"cpu":{
"number": 2,
"die": 0.245,
"core_number": 24
},
"ram":{
"capacity": 32,
"quantity": 12,
"die": 1.79
},
"ssd":{
"capacity": 400,
"quantity": 1,
"die": 50.6
},
"hdd":{
"number": 0
},
"power_supply":{
"weight": 2.99,
"quantity": 2
}
}}
Any element can be removed and will be replaced server side by a default value
Response
{
"hypothesis": "not implemented",
"impacts": {
"adp": 0.1493534978742977,
"gwp": 963.6083516103959,
"pe": 12834.909589529003
}
}
from boaviztapi.
Related Issues (20)
- Internal server error when requesting https://api.boavizta.org/v1/server/ with archetypes dellR740 and mac2.metal HOT 5
- No CPU core units default for lots of archetypes HOT 3
- Remove 0.491 in CPU die calculation
- GWP use impact value is "not implemented" in last version for at least desktop and laptop HOT 2
- Integrate DC (technical environment en building) footprint estimation
- Extending AWS servers lifetime in servers.csv, to match new official AWS refresh policy ?
- Missing AWS platforms / servers for several instance references
- cloud/instance with is4gen.8xlarge leads to "ZeroDivisionError: float division by zero" HOT 3
- Instances referencing non existing "platform_aws_m1" platform, leads to 500 error / HOT 1
- The impact of RAM and CPU usage is counted twice
- Chore[CI]: update github actions that rely on Node16
- Compliance with ISO 21031/GSF Software Carbon Intensity
- Update electricity impact factors
- Verbose output of CPU and RAM use impact values are inconsistent with total use impact values
- RAM coefficient value - implementation vs. HotCarbon paper
- Add a API route that returns current version of the API
- Run python tests CI workflow on dependency update HOT 2
- Improve our security posture by allowing dependabot to open PR HOT 2
- Upgrade fastapi dependency HOT 3
- Upgrade to pydantic v2 HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from boaviztapi.