The untapped potential of volunteer computing | Hacker News Reader

The untapped potential of volunteer computing

inverse.com

39 pointsdodders13 comments

The article doesn't mention what I suspect is the biggest reason for the decline in popularity - improved power management. When these projects were new they really were using computing power that would otherwise have gone to waste. Idle power consumption was very similar to power under load, so it made sense to keep your CPU busy. But now idle power is much less wasteful, and everybody participating in these projects pays for it in higher electricity bills. This is unsurprisingly less popular than helping at no additional cost.

Not to mention the move to portable electronics. Having SETI@home running may make only a minor difference in performance on a desktop, but that's precious battery life for a laptop user.

I mostly run my laptop on the power grid and many other people, too. So your argument only makes sense for people that run their laptop mostly on battery

The much worse problem is that many laptops have non-decent cooling - so if you really put them under hard workload for a longer time, you will at least hear that by the loud annoying fan or the processor will reduce its clock frequency making the laptop slower.

I am not sure if the current boinc clients do this but to gain more users they could

* run only while connected to wall power

* run while battery is 80% charged or higher (don't compete with a 0% charged battery)

* don't spin the fan (meaning throttle back and don't make the user's machine sound like a jet engine)

Exactly. It's not a small effect either - my previous pc jumped from sub 100W idle to 400+ under load. That'll definitely put a dent in the electricity bill if it's 24/7

I'm curious about the perceived (or not!) adverse effects of running cpu heavy computations on a MacBook or phone for years? This article doesn't discuss the downsides of running something like BOINC if there are any... But that may be a real issue that shouldn't be ignored...?

That's right. A MacBook user would want to know how this daemon would affect the longevity of his or her device. My own guess is that heat cycling is more detrimental than relatively constant (but below thresholds) usage.

Last time I looked into this I concluded that the effect on hardware life is near zero. Stuff is outdated before it matters unless the gear is running at its thermal limit which is unlikely unless something is wrong.

Power bills on the other hand are a real problem

The article really should mention http://www.gridcoin.us/ A Crypto currency awarded for actual BOINC work units.

I was a big fan of Folding@home back in the day, but now that I have the wonderful responsibility of the electricity bill, I've stopped all work. I wonder if Gridcoin could become a suitable offset for those bills in the future?

The article does.

Wow. I completely missed it.

>There may be a time when we think of computer power as money — Bitcoin is pushing us that way — and we give back by donating a few calculations at a time.

Actually, can't we think of a low/competitively priced resource as a donation? Why can't contributors be compensated? Not thoroughly accounting for value contributed to networks is part of the reason for decreasing employment prospects. Users are taking risk (by spending their time creating value) and uploading it without compensation.

It would already be a step into the right direction if institutes using BOINC promise to release any research papers they do (not only the ones generated using BOINC) completely (without exception) under some creative commons license. If they already do, they should advertize this fact.

In other words: Volunteering computation time just to make Elsevier/Springer/Wiley rich or volunteering computation time to an institute which also supports projects that make Elsevier/Springer/Wiley rich is something that I'm dubious about.

Reply on news.ycombinator.com