Page 1 of 1

Ars Technica: Is distributed computing dying, or just fading into the background?

Posted: Tue Jul 11, 2023 9:02 pm
by bikeaddict
I am still undecided on whether I will continue with distributed computing long term. Some days I think about buying newer hardware to do more work and improve efficiency. But the cost of occasional hardware upgrades and the electricity bill combined with doubts about whether I'm really making a difference sometimes make me wonder whether I should walk away and sell off my hardware.

It's interesting that F@H points trended way up last September and then fell again in Feb. and April. Maybe former Ethereum miners repurposing GPUs for a while.


https://arstechnica.com/science/2023/07 ... -backdrop/

"With the outbreak of the COVID pandemic, there was a new darling in the distributed world: Folding@home..."

"But as the pandemic waned, so did interest in the project."

"Bowman thinks there is a combination of reasons for interest dropping off. “The pandemic gave huge motivation and a lot of time for new hobbies..."

“Inflation and energy prices soared,” Bowman said.

Re: Ars Technica: Is distributed computing dying, or just fading into the background?

Posted: Tue Jul 11, 2023 9:47 pm
by BobWilliams757
I see it much the same as any other charitable donation. It might help the end cause. And it is the same if you give someone down on their luck a bit of work, a few bucks, etc. Long term we might never know if it mattered. Many other charities are the same.... what did it really help or did it even help?

In this case, many of the things they are researching are very complex, and have cost way too many lives over the years. If it helps advance the science even slightly it's bound to help people in the end, even if that is not an immediate thing.

As such, I just contribute what I do, and that might change over time depending on my other wants, needs, financial situation, etc. I doubt I will even contribute less than I do now, as I have but a single modest GPU folding on a joint use system. And chances are I will at least upgrade sooner, or maybe build a second box to fold on more likely. At the same time, I can't see myself doing as some do and building huge farms of equipment either, even if money was no object... which it always is on some level.

But one thing for certain that I learned long ago.... not trying to help with something never helps that cause. All we can do it try.

Re: Ars Technica: Is distributed computing dying, or just fading into the background?

Posted: Wed Jul 12, 2023 8:30 pm
by [Ars] For Caitlin
I read the article. As my username suggests, that was my "home" technical news site. I'm pretty sure I joined to discuss Seti@Home. Anyway, Ars is well... I dunno what to call it. A marginally better version of Wired magazine? The article appears pretty low effort, although for the most part true. I've tried to keep up with distributed computing over the years, and its basically fallen out of any mainstream discussion. The DCA at ARS is dead, and even the HardOCP and Anandtech forums are sparse.

I think Dr. Anderson hit the nail on the head when he wrote the obituary in his 2022 "BOINC in retrospect" blog post. In many ways, technology moved on to the detriment of volunteer compute. Scoring systems disincented the common volunteer. Needing a very high end gaming system just to become an average contributor and multiple of them to rise above that kind of limits participation. On the researcher/project side, nobody is going to build infrastructure these days for volunteer donors that may or may not show up, except those with not much to lose in the first place.

At this point, incumbency counts for everything. F@H and a few of the BOINC based astrophysics projects appear to be the only projects that actually generate academic papers. I consider that the measuring stick of whether I'm wasting my time or not. And just in case anybody missed it, Seti@Home essentially fizzled out in the data analysis phase and won't even wind up publishing anything. Sad end to the project that mainstreamed the volunteer compute idea.

What I see left is a hardcore group of hobbyists for lack of a better term. I volunteer my resources because I'm interested in science, academic research, and computing in general and volunteer compute is sort of the nexus of those things. I do find it sad that there's not a whole lot of camaraderie about it any more.