Building Green Software

[Michael Krimgen]

Hi Anne, Sarah, and Sara,

Great book, I already love the table of contents. I was wondering if you also included some research if there is a rebound effect in software development and how big is the impact on the environmental impact of sofware development.

Another interesting topic might be the choice and appropriate configuration of your IDE (and other development tools).

Cheers,
Michael

[Anne Currie]

Hi Michael,

We don't give many figures in the book as we have tried to stick to principles because they have greater longevity. We do give some high level numbers. The impact of DCs is estimated at a few % of carbon emissions, which is not as big as something like agriculture or transport but we all need to get that down to zero.

Cheers,
Anne

[Campbell Ritchie]

I presume DC means data centre.
Even if they are a few percent of CO₂ emissions, that is still a lot of climate change.
Do you think that use of reversible hardware would reduce that energy consumption?

[Michael Krimgen]

Thanks Anne for your answer. I fully agree that we need to get emissions down. However, do you think that getting them down to zero is a realistic goal?

[Sara Bergman]

Campbell Ritchie wrote:I presume DC means data centre.
Even if they are a few percent of CO₂ emissions, that is still a lot of climate change.
Do you think that use of reversible hardware would reduce that energy consumption?

Hey Campbell Ritchie!
Interesting question, and it taught me something new, which is always nice.
So in general, I'd say it is a balancing act. Hardware that use less energy is GREAT, that means that we can do more with less, which is one of the things we talk about in the book about a lot. And to my, newfound, understanding of reversible hardware, it would use less energy by minimizing heat generation, which is (literally) cool.
The flip side of it is the embodied carbon of the actual equipment. Say we throw out all existing (perfectly working) chips, in favor of reversable ones, that is likely not worth the embodied carbon cost that we can potentially save in energy gain. Especially in a post-energy transition world.

[Campbell Ritchie]

Reversible computing works by not deleting any information, but as you say, it wouldn't be worth introducing to replace old data centres, not until they wear out.
There are two irreducible energy costs under thermodynamics: one is information deletion, and the other speed; reversible hardware would in theory use no power at all if it moved at an infinitely slow speed. That latter problem is likely to be unacceptable to anybody, even if removing information loss would be transparent to users.

[Michael Krimgen]

Campbell Ritchie wrote:Reversible computing works by not deleting any information, but as you say, it wouldn't be worth introducing to replace old data centres, not until they wear out.
There are two irreducible energy costs under thermodynamics: one is information deletion, and the other speed; reversible hardware would in theory use no power at all if it moved at an infinitely slow speed. That latter problem is likely to be unacceptable to anybody, even if removing information loss would be transparent to users.

Thanks for pointing this out. I just read a little about reversible computing and learnt quite some new things.

[Anne Currie]

reversible computing does sound super cool ;-) We do have to remember that what is also cool, available now, and revolutionary is renewable power generation. If we can be smart and align computation with cheap renewable power, that is a quicker win