Rendering an animation movie takes a lot of energy and so does heating buildings. What if we found a way to combine both and thus save one of the two expenses?

This is exactly what we do at Qarnot! Thanks to our computing heaters and our digital boilers, we heat people for free with their embedded computing power. On its part, Scalemax, our Joint Venture with the french retail group Casino, changes unused spots of huge warehouses into data centers that heat them thanks to green energy partially provided by solar panels on the warehouses' roofs.

This little study evaluates the impact of the rendering of an animation movie on two different scales of home heating:

- - an average french house;
- - a few french cities from different sizes and latitudes.

How long could they be heated if they used Qarnot's data centers? Let's find out!

# Rendering

## What is the energy needed to render an animation movie?

Let's take an animation movie of **24 frames per second**, lasting **90 minutes**:

$ N_{f} = 130hspace{0,1cm}000hspace{0,2cm}frames$

Knowing that to get a definitive shot, about **fifteen transitional renders** are necessary (working renders), the number of render jobs is put to :

$ N_{RJ} = 2,0times 10^{6} hspace{0,2cm} renders$

The mean rendering time of a single frame can be estimated at **3 hours** (taking in account the influence of transitional renders and choosing an animation movie of medium-quality).

A 90 minutes long animation movie production is equivalent to:

$ T_{R} = 6,0times 10^{6} hspace{0,2cm} hourshspace{0,1cm}ofhspace{0,1cm}rendering$

Making the hypothesis that the rendering is made with Qarnot, with a Threadripper 2950X (**270W** taking in account the motherboard and the RAM), without taking in account data storage and transfer, the electrical consumption of an animation movie rendering on Qarnot represents:

$ E_{Movie} = 1 575 hspace{0,2cm} MWh$

# Heating

## What is the energy needed to heat an average french home for a year?

According to the government numbers, the mean french house expense for energy consumption in 2018 is **1514 €**. In 2018, the french mean price for electricity on the residentiel market is about **171 €/MWh**.

This leads to an average yearly electrical consumption in one french house in 2018 of:

$ E_{House} = 8,8 hspace{0,2cm} MWh$

Let's make the *estimate* that heating represents **25%** of the electricity consumption (INSEE 2016).

**Note:** these numbers change a lot according to the source (EDF:62%; ADEME: 40%; INSEE: 25%), we're making the choice to use INSEE's numbers but this question can be deepened.

The electricity consumption dedicated to heating a french house is:

$ E_{H.House} = 2,2 hspace{0,2cm} MWh$

## What is the energy needed to heat different french cities during winter?

During the month of January 2020, the electricity consumption (given as an average power) was:

Hauts-de-France | $hspace{0,3cm} 6hspace{0,1cm}232 hspace{0,2cm} MW$ |

Ile de France | $hspace{0,3cm} 8hspace{0,1cm}849 hspace{0,2cm} MW$ |

Provence Alpes Côte d'Azur | $hspace{0,3cm} 5hspace{0,1cm}120 hspace{0,2cm} MW$ |

The following approximation is made to get an *order of magnitude*. Knowing that the size of Ile de France and Paris' populations (from INSEE's figures) is:

$ H_{IDF} = 12,2 times 10^{6} hspace{0,2cm} peoplehspace{0,3cm}$ and $hspace{0,3cm} H_{Paris} = 2,1 times 10^{6} hspace{0,2cm} people$

Paris represents **17%** of the whole Ile de France population.

By following the same reasoning for two other french cities of different size and latitude:

Calais (1,2%) | $hspace{0,3cm} 77 hspace{0,2cm} MW$ |

Paris (17%) | $hspace{0,3cm} 1hspace{0,1cm}523 hspace{0,2cm} MW$ |

Montpellier (5,7%) | $hspace{0,3cm} 292 hspace{0,2cm} MW$ |

As of before, we make the *estimate* that heating represents the following part of the electricity consumption according to the city's latitude (INSEE 2016):

Calais | $hspace{0,3cm} 30 %$ |

Paris | $hspace{0,3cm} 25 %$ |

Montpellier | $hspace{0,3cm} 20 %$ |

**Note:** these numbers change a lot according to the source (EDF: 62%; ADEME: 40%; INSEE: 25%), we're making the choice to use INSEE's numbers but this question can be deepened.

The electricity consumption dedicated to heating in January 2020 (31 days) is:

Calais | $hspace{0,3cm} 23 hspace{0,2cm} MW$ |

Paris | $hspace{0,3cm} 381 hspace{0,2cm} MW$ |

Montpellier | $hspace{0,3cm} 58 hspace{0,2cm} MW$ |

Which gives, in terms of energy:

- - $ E_{H.Calais} = 1,7 times 10^{4}hspace{0,2cm} MWh$
- - $ E_{H.Paris} = 2,8 times 10^{5}hspace{0,2cm} MWh$
- - $ E_{H.Montpellier} = 4,3 times 10^{4}hspace{0,2cm} MWh$

# Let's compare!

## On the scale of an average french house

$ E_{Movie} div E_{H.House} hspace{0,2cm}$ will give the number of years during which the render of the animation movie could heat an average french house.

This implies that the rendering of a medium-quality animation movie lasting 90 minutes could heat an average french house during a little more than 715 years, and that for free. |

## On the scale of a few french cities

$ E_{Movie} div E_{City} times 31times 24hspace{0,2cm}$ will give the number of hours of January 2020 the render of the animation movie could heat the city.

This implies that the rendering of an animation movie of 90 minutes on Qarnot could heat the whole city of Paris for 4 hours and 8 minutes during the coldest time of the year. |

And also

- - Calais city during 2 day 20 hours and 27 minutes.
- - Montpellier during 1 day 3 hours and 8 minutes.

# Contact us!

If you have any inquiries, may it be on the details of this study, or for more information about our work, feel free to contact us, we'll be very happy to answer!