Folding Forum

This might be an unpopular opinion, but as much as folding uses compute power for a good cause, the combined co2 emissions from folding are also immense!

Some suggestions on how to make folding more efficienct, to reduce carbon emissions, lower energy prices, and reduce foreign energy dependency:

Using AI to calculate an efficiency score, to compare performance per watt between devices, users, and teams.

Promoting and increasing ARM hardware support (Android, snapdragon laptop chips, apple silicon), to make people switch from x86 and discrete GPU's, which are more inefficiency in terms of performance per watt.

Ending support for the oldest and most inefficient hardware, to make people upgrade and switch to newer more energy efficienct hardware.

If CPU's and GPU's are doing the same tasks, only GPU's, especially iGPU's, should run those tasks instead of CPU's, since they are much faster and way more efficient per watt than CPU's doing the same tasks.

Just not seeing anybody talking about this, and I think the Folding community should contribute to reducing carbon emissions and saving the environment, like everyone else.

FAH will not influence energy price adjustments
FAH is not single country project, so dependency on foreign energy makes no sense. What is foreign country? FAH is worldwide project.
Android is not a compute platform, and I don't care what everyone's feelings about it are. Phones simply are the most terrible platform for compute tasks. And best fire hazard in your house, if you fold on your phone. Believe me, we tried
FAH will not influence better adoption of ARM. ARM should influence better adoption of ARM hardware.
Qualcomm is the next intel with their "innovations". Snapdragon can stay in its niche market. Folding on laptops in general is terrible idea.
Apple chips, while very impressive, are walled garden, no thank you. User will not just simply go for Mac, just because folding...
Ending support for old hardware makes you a consumerist. World DOES NOT need another project pushing people to needlessly upgrade to latest and greatest. We have enough of this sh*t with company car policies across the globe, pushing people buy into brand new cars every 6 years, just to comply with company policies.
Keep AI out of FAH (Yes, Yes, some FAH projects use AI generated starting positions, etc). The world is falling into oblivion with those two letters, no thank you. And, since when AI is synonymous with efficiency? Last I checked power stations are being built dedicated to AI datacenters. GPU chip manufacturers forgot what efficiency is, as long as it has AI in its name, and arrow goes up and to the right, everything is awesome. The more you buy, the more ... yeah, nah. Keep that boiling rot far away from FAH

If CPU's and GPU's are doing the same tasks, only GPU's, especially iGPU's, should run those tasks instead of CPU's, since they are much faster and way more efficient per watt than CPU's doing the same tasks.

That's not how FAH works. FAH does not create CPU projects and GPU projects just for diversity. CPU projects exit, because GPUs cannot do them, which means CPUs are the most efficient way of folding those projects. Do you think that researcher would create their project for CPUs only, if their simulations were compatible with the GPU? "Let's fold our project in 6 months with CPU, instead of 2 weeks on GPU" - said no one ever.

Just not seeing anybody talking about this, ...

My personal view is just that the benefits of Fah to the world ecosystem - not just the potential health benefits of the science, but the visible demonstration of the possibility of civil international collaboration by ordinary folk - well, OK, maybe slightly geekish folk - justifies the carbon footprint, and this is just as true for any collaborative investment that any participant makes, whether "efficient" by the notional (and basically unquantifiable) metric of "Science Increment per Tonne of CO2" or "Disability Adjusted Human Life Years gained per Science increment per Tonne of C02", let alone by the simplified notion "PPD per kWh per day" ignoring capital costs etc.

And although the "efficiency" debate is worthwhile, it is only one of many worthwhile things that folk talk about in these forums - see for example the very interesting discussion at viewtopic.php?p=359128#p359128.