ArsTechnica: Is distributed computing dying, or just fading into the background?
ArsTechnica: Is distributed computing dying, or just fading into the background?
There seems to be much less excitement about distributed computing these days.
An incalculable amount of PC cycles and electricity wasted for nothing.
I hate this sentence with a burning passion for two reasons.
Firstly its position on an article like this implies that SETI@home not finding anything is a failing of distributed computing, but they could have run that code on a supercomputer and still not found anything.
But more importantly it speaks to a big misunderstanding of science that is not only ubiquitous among the general public, but is also common among funders of science and even scientists themselves. Science isn't about discoveries, it's about investigation. Don't get me wrong, discovering something new is amazing and it's what we would all want even if our funding didn't hinge on it, but the fact that SETI haven't found anything is still an important result and not a failure. I genuinely think this focus on discoveries will be the downfall of modern science if we're not careful, it means that important things like the sustainability of scientific software or reproducibility of research are being left by the wayside because they don't lead directly to discoveries.
Sorry, went off on a massive tangent there...
This reminds me of the plastics and fossil fuel industries telling consumers to recycle to fix the pollution problem.
Distributed computing for science research is like 0.01% of the wasted computing electricity while the idiocy of cryptocurrency where people literally waste electricity to create proof that their fiat imaginary coin has value to the suckers in their bigger fool scam.
Last I heard the world has a second UKs worth of burned fuel to make electricity for that ongoing scam.
@seamsay
True, if you see nothing, this is still something to study and to acknowledge.Too many paper are not published today because they don't give result, which means a lot of experiments are done again and again because nobody though of publishing that it does'mt work.
@makeasnek
Worth noting that BOINC (the distributed computing platform behind SETI@Home) is alive and well with over a dozen projects. You can help scientists cure diseases, map the galaxy, and more. The Large Hadron Collider even has a BOINC project you can crunch for. See the Lemmy for BOINC https://sopuli.xyz/c/boinc
I was into DC 25 years ago when it was new and exciting, computers had limited power management and I got free electricity from the university.
Now it’s old news, power management means DC causes CO2 emissions, and I pay the electric bill.
I will not join any DC projects unless I generate surplus solar power. To do otherwise is ecologically irresponsible.
We never really get to see the results of our machines' work, so it fell pointless. Also didn't help the UX was all terrible.
Yeah, not seeing any results was the big downside. I remember being so excited to participate in SETI. Only for nothing to ever come from it. Admittedly, SETI may have been overly ambitious and that set it up for failure.
But I'm also a bit skeptical of how effective home computers can be for a lot of these projects, considering how unfathomably massive data centers are these days. Not saying they aren't impactful, but rather that any really compelling study is likely to get a grant or corporate sponsorship that can pay for a bonkers amount of computational power.
Consumer hardware is relatively inefficient by comparison and requires doing redundant extra work to prevent fake results (because trolls will troll anything). Plus it's not considered acceptable to run at 100% on people's home PCs. If I remember correctly, they usually throttled the work so that it wouldn't be so noticeable.
Much of science is broke/underfunded and home computers have collectively a shit load of idle computing power on traditional processors and GPUs which are harder to get and more expensive in DCs. The idea of distributed computing for science is sound.
I was mostly disillusioned by the lack of feeling of participation or accomplishment.
I used to run folding@home on my PS3 until Sony removed the ability. No idea how successful it was.
Uh… am I missing something here? (I don’t have a computer science degree) What about GPT from OpenAI? Amazon Web Services?
What about smartphones? They seem pretty popular. How much computing on them can you do without a network connection?