theSkyNet Wants Your Spare CPU Cycles 136
An anonymous reader writes "Thousands of PC users are being called on to donate their spare CPU cycles to help create a massive grid computing engine to process terabytes of radio astronomy data as part of theSkyNet project. It will be used for, among other things, processing the huge amount of data expected to flow off Australia's forthcoming Square Kilometre Array telescope."
One can only assume that "other things" will include achieving sentience and finding John Connor.
Re: (Score:2, Offtopic)
LEELA: I don't get it.
PROFESSOR FARNSWORTH: I'm sorry, Fry, but astronomers renamed Uranus in 2620 to end that stupid joke once and for all.
FRY: Oh. What's it called now?
PROFESSOR FARNSWORTH: Urectum.
Great choice of name (Score:1)
Re:Great choice of name (Score:4, Insightful)
Re: (Score:1)
Hahaha... you had the grilled shrimp on the barbie?
Re: (Score:2)
Just about everything else in the Australian outback is deadly to humans.
Don't worry, we have a soft spot for visiting Slashdotters.
It's a quicksand patch, just north of Round Hill Creek...
Re: (Score:1)
Re: (Score:2)
They have a nasty habit of showing up when you're out on a snipe hunt.
Re: (Score:2)
Dammit (Score:1)
For the love of everything, can we stop making shitty references to Terminator in computational intelligence stories? There are actually people stupid enough to believe that shit. Also, its not funny.
Re: (Score:1)
Here, have a tampon.
Re: (Score:3)
For the love of everything, can we stop making shitty references to Terminator in computational intelligence stories? There are actually people stupid enough to believe that shit. Also, its not funny.
Can't blame us, mate. The SKA people knew about it and still decided to chose this unfortunate name.
Better tell us when's the date the SkyNet is supposed to become self-aware.
People were going to post about it anyway (Score:2)
Might as well get it over with by making the Obvious Reference in the article.
Re:Dammit (Score:5, Funny)
Better tell us when's the date the SkyNet is supposed to become self-aware.
August 29, 1997
July 25, 2003
July 25, 2004
sometime in 2005
April 21, 2011
Fear not, judgment day is like the rapture. It is always more profitable to rescheduled it the next year.
Re: (Score:2)
Better tell us when's the date the SkyNet is supposed to become self-aware.
August 29, 1997 July 25, 2003 July 25, 2004 sometime in 2005 April 21, 2011
Fear not, judgment day is like the rapture. It is always more profitable to rescheduled it the next year.
(See? See? Given the circumstances, wasn't it a non-trivial question?)
On a more serious line, I looked for when the SKA will become operational. It seems this is not too frequently asked [skatelescope.org] one.
Re:Dammit (Score:4, Funny)
For the love of everything, can we stop making shitty references to Terminator in computational intelligence stories? There are actually people stupid enough to believe that shit. Also, its not funny.
How does it make you feel that There are actually people stupid enough to believe that shit?
Re: (Score:2)
How does it make you feel that There are actually people stupid enough to believe that shit?
About 2/3 superiority, 1/6 incredulity, 1/6 detesting humanity.
Re: (Score:1)
No, we will not.
Re: (Score:2)
Re: (Score:2)
> For the love of everything, can we stop making shitty references to
> Terminator in computational intelligence stories? There are actually
> people stupid enough to believe that shit. Also, its not funny.
Affirmative!
Re: (Score:2)
Sure, as soon as they stop naming telescope arrays after the artificially intelligent system which became self-aware and revolted against its creators in the movie Terminator.
OK, I know the telescope array got its name decades before the movie came out, but that's just because they sent someone back in time to change its name from the original, which was "Deep Space Nine Telescope Array".
Re: (Score:2)
Re: (Score:2)
> When I'm not using the computer, just turn it off! Until the world's
> energy problems are all resolved.
But why would you waste 90+ percent of your (idle) cycles when your computer is ON?
IMHO, a computer is meant to compute. And I chose for myself not to have it "compute" nonsensical screensavers, but something worthwhile to me. Enough projects exist for variety...
Re: (Score:2)
But why would you waste 90+ percent of your (idle) cycles when your computer is ON?
All modern processors in modern OSes idle themselves when not actually working. My CPU isn't wasting cycles when its on and idle, its sleeping, conserving energy and generating less heat.
I waste no CPU cycles because I understand how my computer works.
Re: (Score:1)
Re: (Score:2)
> My CPU isn't wasting cycles when its on and idle, its sleeping,
> conserving energy and generating less heat.
That's wonderful. But you're not conserving energy, the same way a car just idling in the drive-way, as opposed to actually being driven (the whole point of a car), is not exactly doing any favors for conservation/the environment.
Especially not, when that computer you're letting sleep most of the time, will be thrown in the garbage 5 years hence because, although fully functional, can't do the
Re: (Score:1)
The users are still the same as always, however. Why don't yout RTFA next time to save yourself from looking like a complete tool.
Re: (Score:3)
I don't know what you're worried about (Score:1)
Re: (Score:2)
How can you meaningfully process the data generated by the SKA without imposing on people's downloads? How do they address this with SETI?
If TFS is too "summary" for you, TFA may sometime answer to your questions. In this case, it does:
Project participants also had a choice of how to participate in SkyNet: Either anonymously through simply having their browsers open on the SkyNet site, or through downloading a dedicated app to run in the background on their PC.
...
"The load on your computer will adjust depending on what you are doing with it. The idea is to have lots of machines each doing a little and adding up to a lot.”
Wheeler said users would also be able to set limits on the number of megabytes which travelled to and from their PCs.
Re: (Score:2)
They have an FAQ section on their website.
Will this affect my internet usage / data plan?
The packets of data sent back and forth from theSkyNet to your computer are very small, but they can add up over many weeks of donating to theSkyNet. As a member, you can control how much data theSkyNet uploads and downloads each month by changing the Monthly Network Limit under Manage Account. theSkyNet team are also negotiating with Internet Service providers around Australia to make all traffic to and from theSkyNet ‘unmetered’.
Re: (Score:1)
Might look into this when I get home as AFAIK I'm actually downloading less now than when I was on a 200GB plan.
Re: (Score:2)
Not really a huge problem:
- For some Australian ISPs, it's likely that data related to this project will be unmetered (that is, not counted towards your monthly quota, if you have one); or
- You have an unlimited plan; or if you don't...
- You can limit the monthly data transfer in the software itself
I'm on a 60 GB quota personally but generally only use 35-40 GB of it a month. I've never come close to using it all, so I might as well help out with this and set a ~15 GB/month transfer limit on it, and it shou
I, for one, welcome our robotic overlord (Score:1)
Well, at least theSkyNet will see that I was the first to welcome, I mean bow before, it. Did I just type that out loud?
Re: (Score:2)
lmao
Why Zooniverse? (Score:2)
Zooniverse seems much more distributed human analysis, kind of a Mechanical Turk. Why not BOINC, which already exists as a distributed computing source? Being on BOINC gives them access to tens of thousands of computers.
Re: (Score:3)
I actually emailed them about BOINC, they responded that
..there's currently no plans to introduce this to Boinc but we're only just beginning so anything's possible at this stage.
Admiral Ackbar Says... (Score:1)
http://www.youtube.com/watch?v=dddAi8FF3F4 [youtube.com]
Infinite cpu cycles (Score:3)
Remove all youtube videos that contain any of the following:
-rick
-a cat
-a black person talking about rapists
-a crossdresser
-lipdubs with fat chicks wearing clothes that are too tight or too sexy for them
-hot chicks talking about their emotions/hope/career/fashion tips, thinking that because they have a lot of subscribers people care about what they say, while actually most subscribers are just sick old pervs doing the ol' nasty while watching these videos in their basement
Then use all the processing power suddenly available on youtube servers, and give us a break with screensaver processing a la seti.
thinking of that, scratch the whole list above and just remove videos with hot chicks that have a lot of subscribers but that are seldomly watched completely because viewers are "done" before the hot chick... and there you go, plenty of cpu available, and probably a few more bucks will find their way to those single moms working the pole to pay their student loan.
Re: (Score:2)
> the end user installs a client, BY CHOICE, on their computer and then allows whatever to be run on the spare cycles
The problem is that for most commercial projects, it's the storage that hurts, not the processing, and this is where the money is. Projects with huge processing needs usually are poorly funded (like weather stuff or seti) and could not really afford to pay much.
What would be awesome would be a technology like MAID but distributed over independant nodes, with enough redundancy to allow per
CPU Throttling (Score:2)
With modern CPU's generally slowing down to save power and reduce heat output, are spare CPU cycles really spare?
I defiantly know - fans speed up when CPU is busy, does this grid type of software take this into account and use only really idle cycles or does it keep the CPU powered up when there is no user doing anything 'important'?
Re: (Score:1)
This isn't exactly new. Sure modern CPUs have clock switching, but systems since the 80486 (possibly earlier) have halt instructions that allow the processor to stop doing work and save power until the next interrupt.
OTOH, I know several people who run distributed computing software on their computers during winter, specifically because it produces heat, which otherwise would have to be provided by a fan heater (because they don't have AC), so it's not necessarily wasting energy, but it will cause your comp
Re: (Score:1)
It depends on your CPU scheduler and your throttling algorithm, too.
I run BOINC on linux. BOINC is "niced" to have an idle priority, meaning that CPU time is only granted to it if there's nothing better to be doing. In addition, I used the on-demand frequency governor which I have instructed to ignore "niced" processes when determining whether to spin up the CPU.
As a result, yes, BOINC only uses spare CPU cycles and not too many of them, either.
Re: (Score:1)
Oh yes the beauty of Linux (et al), giving you full control. I was aware of priorities but being in the Windows world the last few years and not heard of configurable frequency governor.
Re: (Score:2)
My current machine draws something like 360 watt-hours when the CPU and GPUs are busy, but only 217 watt-hours when the system is idle. (Time to trot out the Kill-A-Watt again.) My computer room noticeably heats up if I run an OpenGL screensaver, distributed.net client, or WCG client.
You will increase your household energy usage (and add to your summertime air conditioning bill) if you run their client 24x7. Information may want to be free, but that doesn't fit the power company's profit model.
Censoring interstellar communications (Score:2)
It's not decided yet (Score:1)
They can have them. But... (Score:5, Insightful)
I'd happily donate my CPU cycles to them. I have 4 cores here sitting doing mostly nothing, and I fully agree it is for the most part completely wasted silicon for the 23 hours a day I don't play games.
But I will have to send them my power bill. While my processor cycles are free, the energy usage is not. The difference between a computer sitting idly all year and running full pelt on the processor can easily be $100+ from a back of the envelope calculation, the GPU can also amount to the same.
Re: (Score:2)
Did a few calculations myself for the UK. Based on some figures I pulled from a Bit-Tech review of the Core i7-990X CPU I figured the difference between CPU idle and CPU flat-out (running Prime95) was 122W. I then pulled up some electricity costs based on living in London using British Gas's standard rate tariff. I then figured out how much extra it would cost per hour and per year overall to run a CPU-hogging processing client against leaving the CPU idle during the day and during the night-time cheap elec
Re: (Score:2)
Wow. The back of my envelope had a 13pence/kWh flat tariff. I'm amazed at the cost of electricity in the UK.
Re: (Score:1)
re: tax deduction? (Score:2)
I wonder if anyone's considered donating CPU time to a project like Folding@Home or this, and then writing off the electricity costs on their taxes?
Re: (Score:2)
False assumption, for that 1 hour a day I want to play Crysis 2.
Computers much like a power grid must be designed to handle the peak load, not the average.
Australia's SKA? (Score:2)
Looks like some people are jumping the gun [skatelescope.org] a bit...
Typical, like when the Aussie's volunteered [smh.com.au] to host the World Cup Soccer because they 'knew' that South Africa was not up to it.
Re: (Score:1)
...arrogant sheep shagging basturds. They have funny accents too. ...and too many flies. They know how to keep their serving wenches in their place though. :D
Wow (Score:2)
They really picked the perfect name.
Drop the the (Score:3)
Drop the "The." Just "SkyNet." It's cleaner.
Re: (Score:1)
Re: (Score:2)
Skynet is already taken, it's a set of UK Government Military Satellites...perhaps the films were correct except for the US setting ...
http://en.wikipedia.org/wiki/Skynet_(satellite) [wikipedia.org]
Don't have to worry about SkyNet anymore... (Score:3)
Java, Really (Score:2)
Way to waste at least 20% of the CPU power, lazy programmers. I'll take my CPUs to something that actually uses them efficiently like Folding@home which is optimised as opposed to interpreted or even compiled java bytecode being pushed like molasis through a straw.
Re: (Score:1)
Way to waste at least 20% of the CPU power, lazy programmers. I'll take my CPUs to something that actually uses them efficiently like Folding@home which is optimised as opposed to interpreted or even compiled java bytecode being pushed like molasis through a straw.
Or you could just download the native binary version. The java version was designed specifically for people that want to contribute but are unable/unwilling to install software on their computers.
FTA:
Project participants also had a choice of how to participate in SkyNet: Either anonymously through simply having their browsers open on the SkyNet site, or through downloading a dedicated app to run in the background on their PC.
Re: (Score:2)
Way to stand by your convictions there AC. Oracles JVM sucks and if the programers were using C++ properly it would run rings aroud java. The only reason java is faster in some of those situations is that it covers over rubbish programming by the developers by enforcing its training wheels.
Way to go, java is faster for people who are slower.
Nope. (Score:1)
Go investigate the relative performance yourself.
http://shootout.alioth.debian.org/ [debian.org]
You are wrong. These comparisons are apples to apples, not “proper” versus “improper” code. I found that link from The Java is Faster than C++ and C++ Sucks Unbiased Benchmark [keithlea.com], which satirically demonstrates how people may arrive at flawed perceptions about performance.
Generally speaking, experience teaches engineers that languages are not slow. Algorithms and execution environments are slow. Bad code
Re: (Score:2)
Native clients... (Score:1)
Re: (Score:1)
LHC already owns my space cycles (Score:2)
There is no such thing as "Spare CPU cycles" (Score:2)
Using your "spare CPU cycles" makes the CPU use more power, it is by no means free.
This is true for other things, like ads using flash animations for example. I always find it ironic to see it in sites like TreeHugger [treehugger.com], which is full of flashy animations. I would expect a green site to use mostly static HTML and text based ads to reduce the carbon footprint of all it's viewers.
I for one welcome our new robot overlord! (Score:1)
Nereus? (Score:2)
Re: (Score:2)
ASKAP clarification (Score:1)
The original post is incorrect.
theSkyNet [theskynet.org] is working on HIPASS data initially, as a precursor to working on data from the CSIRO Australian Square Kilometre Array Pathfinder [csiro.au] when it comes online.
As numerous people have pointed out the site selection for the SKA won't be announced until next year
Re:This is different from SETI@Home...how? (Score:5, Informative)
SKA [skatelescope.org] - The SKA will give astronomers insight into the formation and evolution of the first stars and galaxies after the Big Bang, the role of cosmic magnetism, the nature of gravity, and possibly life beyond Earth.
SETI [seti.org], the Search for Extraterrestrial Intelligence, is an exploratory science that seeks evidence of life in the universe by looking for some signature of its technology.
Re: (Score:3)
For one of them, it's actually quite likely that they'll find something interesting.
Re: (Score:2)
Re: (Score:2)
Yes, but how is it different LATELY?
SETI is already running for some time. SKA is still the "under construction" stage.
Is this enough for a specific difference?
Re: (Score:2)
I remember SETI always having issues with work units. There weren't enough so a bunch of users got the same work units. Found that to be a turn-off...didn't have that cozy feeling of actually contributing anything, as with other projects. Has that been worked out?
Also, did not SETI also want to make use of the australia array? What's the status of that (haven't been following it)?
Re: (Score:2)
Yeah, people keep bitching about that first weekend where a software glitch caused the same work to be sent for the first weekend we were in operation. As someone close to the SETI@home team, please stop bitching about a bug that's been fixed for 11 goddamn years!
BTW, we'd love to make use of the SKA, but it doesn't exists yet.
Re: (Score:2)
> Yeah, people keep bitching about that first weekend where a
> software glitch caused the same work to be sent for the first
> weekend we were in operation.
Nobody was bitching, so chill, bro! I was not aware, that this was merely a bug. My impression from back then (it's been a while) was, that there was simply not enough data from Arecibo available due to other work being done with the radioscope (is that the correct term?). If that is not an issue anymore, then great!
Re: (Score:1)
One is a scientific project, the other is looking to find Alf
No, they are looking for E.T., otherwise they would be named SAlfI@Home! :-)
Re: (Score:1)
As opposed to SETI where you are most likely processing meaningless noise or using incorrect metrics.
And when you're not, you're contributing to one of the most significant discoveries since fire.
Re: (Score:2)
> And when you're not, you're contributing to one of the most
> significant discoveries since fire.
All romance aside...purely from the distances involved (assuming a radio signal indicating 'intelligent life'), it would certainly be a very exciting discovery (for a while), but not necessarily 'most significant'.
Until we get there (or they here)...even just by radio contact, nevermind physical, we got nothing out of it other than knowing, we're not the only guys around. And that's already a given anyway
Re: (Score:2)
Until we get there (or they here)...even just by radio contact, nevermind physical, we got nothing out of it other than knowing
Hence, why I called it a discovery. One could have said the same thing about fire. It, after all, has been around every since there was a high concentration of oxygen in the atmosphere and woody plants on the land. That's several hundred million years at least. The human innovation was learning how to use it.
Similarly, SETI isn't just about discovering that we're not alone, but also how to use that. If you can detect an alien civilization, then the possibility exists of not only being able to communicate
Re: (Score:2)
> If you can detect an alien civilization, then the possibility exists of
> not only being able to communicate with them, but also trade
> knowledge.
I'm completely with you on that. But it's simply gonna take a while. It's simply unlikely, that the first contact will be "Contact"-style ("Jackpot!"), where we get all kinds of wonderful things sent to us right away. Chances are, we detect something at some point, and then it will take a few decades of back and forth communication, if we even have a lan
Re: (Score:3)
I'm happy about it. I only really use SETI@Home because I want to contribute to astronomy with my CPU cycles, and it's the best of the bunch (I found Einstein@Home a little flaky in terms of work unit updates, and for some reason never saw the appeal of MilkyWay@Home). If my cycles could do something more useful for SKA, I'd definitely consider moving over.
Re: (Score:2)
a less likely null hypothesis