[ View menu ]

Saying Goodbye to SETI@home, Wondering What’s Next

When the announcement that SETI@home will enter hibernation and end volunteer processing on March 31 appeared, I assumed that the date wasn’t chosen only because it was the end of a month, but also because it was a Tuesday, which is the weekly maintenance day. For that reason, I also assumed that the last work units will be sent before the servers will be shut down for maintenance, and after they’ll return on-line it’ll all be over and they’ll be configured to only receive results. However, things were a little bit better, since they’ve been skipping the scheduled maintenance towards the end, including on March 31, and what happened on that day was that the last data files were added to the splitter queue, the last new SETI@home work unit actually being sent out the next day, and existing work units will be resent if necessary, until valid results will be received for all of them. What’s more, at that point there were plenty of Astropulse work units still to be created from the same data files, the process going very slowly, apparently due to a lack of space.
It’d have been nice if they’d have allowed that process to complete. In fact, since they got that far, it’d have been even better if they’d have continued until the last data file recorded on March 31 would have gone through the splitter, and some users were hoping that will happen when they saw those March 27 files continue to be added as hours passed on March 31, before the announcement that those were the last ones showed up. However, the servers were then taken off-line, without any announcement, in order for changes to be made, and when they returned on-line the splitter queue was cleared, likely meaning that only SETI@home work units were created out of those last March 27 data files. I’m not sure what it means for that old file that was actually the last one added to the splitter queue, I guess because it was among those they wanted reanalyzed, having been recorded on March 31, 2011, seeing as it had already been through the process and the SETI@home part finished almost as soon as it started, but as far as I’m aware the Astropulse part never got to it to “see” that it had been processed before.

As I already said, the decision to shut down, or at least mothball, the project is awfully disappointing, and that may be putting it mildly, especially considering what seem likely to be the real, or at least main, reasons and “politics” behind it. The fact that it had become far more for many users and a large number of people will therefore truly feel the loss at a time when it’s already particularly difficult to maintain mental health and balance only adds to it, greatly, as does the fact that it popularized distributed computing and likely continued to be the standard bearer for the concept. But even strictly in terms of the science, the original goals, this project was supposed to either find signals sent by intelligent extraterrestrials or prove that there are no detectable radio signals regularly sent our way by them, which obviously wouldn’t have proven that there are no intelligent extraterrestrials out there, but only that this method is not the right one to discover them. In other words, as long as there’s a reason for the radio part of the SETI project to continue, SETI@home should as well. It was never a question of having something to show along the way, but of either detecting and confirming that signal of a probable intelligent origin or scanning the entire sky at all wavelengths likely to be used for this purpose, neither of which having happened.

What’s next for the project, what they’ll do with Nebula or any other plans, I don’t know and at this point can’t say I care. What I know right now is that I have a few dozen work units left and tomorrow I’ll need to look for a project to switch to, most preferably before I’ll run completely out of SETI@home work, so the computer won’t idle if the first or even second one I’ll try won’t be able to provide sufficient work right away or if I’ll decide I can’t commit to it because it uses far too many resources, both scenarios actually being very probable, considering the number of users flocking to other projects, and especially to similar ones, at the moment and the fact that, based on what I saw so far, SETI@home used an extraordinarily low amount of resources compared to the rest. I mean, two SETI@home work units tend to use no more than 75 MB of RAM and write mere kilobytes to disk, while a full queue of 150 work units only takes about 100 MB of space. Admittedly, Astropulse work units have about 8 MB each, so those could add up if there’d be more than a few at a time, but they always were rare, so it didn’t make much of a difference. On the other hand, I’m seeing others say that other projects can easily require around 1 GB of RAM per work unit, and even up to 2 GB, and possibly even more disk space, while also frequently writing large amounts of data to disk, when on my 32-bit system with a total of 4 GB, the listed usable amount being 3.19 GB, I might, barely, be able to spare a total of 0.5 GB at the very most, so 0.25 GB per work unit unless it’s a project that can use both CPU cores for a single one, and would also want to avoid that sort of wear on the HDD, which is over seven years old.
There’s also the fact that one potential option, returning to climateprediction.net, which was the first other project I used when SETI@home couldn’t provide enough work, many years ago, is quite clearly out of the question even without taking the resource use into account, because I noticed recently that I no longer had an account there and the credit for the past work no longer showed up in my BOINC account either, later learning that a few years ago they removed all accounts that hadn’t been active in more than three years! Now I am looking for either a green project, and I have a rather hard time considering it one, or another one having to do with space, so it wouldn’t have been my top choice anyway, plus that a quick look told me that it now easily requires around 0.5 GB of RAM per work unit, and possibly even 1 GB, in which case it’d have been out of the question anyway, but that sort of behavior, wiping away the past work done by somebody without the person requesting it, is something I won’t put up with.

For others in a similar situation, there’s the SETI orphans thread, and also the Which project after? one, but I find the options and discussions there disheartening. Science United, strongly supported by at least one of the SETI@home administrators, apparently the one who pushed to end the project, is clearly out of the question, not allowing the user to select the specific project or projects to attach to. Einstein@home, which seems to be the most popular option among users, is also out of the question due to the resource requirements. Same goes for Rosetta@home, which also seems very popular, though that’d clearly be out of the question anyway, being a medical project when I’m too terrified of serious illnesses to get involved even in such a manner. That also goes for Folding@home, which isn’t even a BOINC project anyway. MilkyWay@home does remain an option, and I have briefly used it before when SETI@home ran out of work, but it all depends on the resources used. Might also have a look at that Africa Rainfall Project on World Community Grid, but even though I have also briefly used it before, I’m uncomfortable with World Community Grid in general, being another layer between BOINC and the actual projects, and that project isn’t actually a green one either, and I highly doubt it’ll have acceptable resource requirements anyway. And then I’m also seeing Asteroids@home, Cosmology@home and Universe@home, which I know nothing about but are at least space projects. But at this point I’m thinking that if MilkyWay@home will require too many resources or fail to provide enough work, which also means enough to store for a number of days in advance, I’ll look through the BOINC projects list again on my own and see what I’ll try next, resource requirements and availability of work being likely to force me to look outside the fields I’d normally care for as well, as much as I’d want to avoid that.

0 Comments

No comments

RSS feed Comments | TrackBack URI

Write Comment

Note: Any comments that are not in English will be immediately deleted.

XHTML: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>