Mirror of forum.nim-lang.org

6972 :: What is the carbon footprint of the NIm?

• mrhdias (orginal) [2020-10-22T13:08:54+02:00] view original

  I saw this website that shows the carbon footprint of a website on the planet.

  It would be nice - if it were possible to do - to show the carbon footprint calculation (or an estimate) of the use of the program, at the end of compiling a Nim program.

  Message Example:

  
  ...
  Hint: 9814 lines; 0.838s; 144.633MiB peakmem; Debug build; proj: test.nim; out: test [SuccessX]
  Hint: test  [Exec]
  Nim Carbon Footprint: Only 0.09g of CO2 is produced every time someone use this program.
  Run

  The following aricle compares the different programming languages in terms of energy efficiency (table 4).

  Where could we find Nim programming language in table 4, if it were part of the study?

---

• xigoi (orginal) [2020-10-22T14:05:00+02:00] view original

  The electricity consumption is mostly related to performance, so I imagine Nim would be doing pretty well.

---

• tmsa04 (orginal) [2020-10-22T15:14:03+02:00] view original

  Beware of using "carbon footprint" as it's an inexact figure, more a political buzzword or general concept than a scientific measure. It's derived by taking various emissions and converting them to carbon dioxide equivalent using arbitrary estimates of their comparative effect. Measuring energy consumption would be more plausible, the electricity could be produced from any power source, which the software could not take into account.

---

• federico3 (orginal) [2020-10-22T16:24:19+02:00] view original

  +1 on measuring energy (Joules, Wh...) rather than CO2

  Also, the bulk of energy usage of an application is in its runtime rather than compilation - by far.

  Nim might be one of the most energy-efficient languages around, as it is very much correlated with CPU efficiency.

  As the ratio of battery-powered devices keeps increasing (tablets, domotics, "iot"... ), energy efficiency could be a nice selling point for Nim. It would be nice to see benchmarks on the website!

---

• doofenstein (orginal) [2020-10-22T16:51:04+02:00] view original

  wouldn't measuring how much energy a (compiled) program consumes require calculating first how long it executes? I'm not too much into theoretical computer science, but that sounds the halting problem.

---

• shirleyquirk (orginal) [2020-10-22T17:38:48+02:00] view original

  https://github.com/kostya/benchmarks

  includes energy use as a metric, which correlates 99.8% with running time (pearson coefficient) so I'm gonna go out on a limb and say go ahead and take runtime as a proxy for energy use.

---