Cell Phones For Science: BOINC Now Available For Android

Luyseyal writes "BOINC is now available on Android. Many of you may not know, but the Slashdot Users team makes a decent showing on World Community Grid. WCG supports research on AIDS, schistoma, cancer, clean energy, and more. Now is your chance to put your idle charge cycles to good use. Let's do some science!"

Comments

Battery

[Anonymous Coward]

Like I'm not already fighting to keep my battery last a day? :(

Re:Battery

[Saethan]

From the app description: 'BOINC computes only when your device is plugged in and charged, so it won't run down your battery'

Re:Battery

[MickyTheIdiot]

out of the box configuration is to not do calculations on battery... but I know we have 20k more battery comments coming.

Also, it is also set up out of the box so that if the battery gets below 90% it doesn't do calculations at any time. This allow the battery to charge up properly on most devices. Having said that, there are a lot of devices (like the Nexus 7 with four cores) that have problems because the battery drains *EVEN IF* it is on the battery. I have to set mine so that it only uses 2 cores otherwise the thing never charges.

Re:Battery

[MickyTheIdiot]

aw.. crap... ...redo

  Having said that, there are a lot of devices (like the Nexus 7 with four cores) that have problems because the battery drains *EVEN IF* it is on the charger. I have to set mine so that it only uses 2 cores otherwise the thing never charges.

Re:Battery

[mwvdlee]

It will burn a hole in your table though.
Ever noticed how hot smartphones get when left running at full capacity for a while?

Re:Battery

[MightyYar]

I suspect you need a higher-amperage charger, and not one marked higher amperage but made for iThings. If you do have an Apple charger, you need a special "charging" cable for most Android devices, though I think there is a hack to force some into high-current mode without such a beast.

Re:Battery

[MickyTheIdiot]

I actually have a solar-based charging system and charge off of a 12v storage battery, so this isn't really relevant for me.

Boinc is the lower priority for me, really, so if I had to I would get rid of it.

However, for most people your advice is straight on right.

Re:Battery

[Saethan]

Yeah I just threw it on two cores on my Galaxy S2 and it started getting -really hot- -really fast- ... so, one core it is.

Re:Battery

[MickyTheIdiot]

Actually, another nice thing about Android clients is that they suspend at certain CPU temperatures. The official android client suspends at 40 C

Re:Battery

[Luyseyal]

My Galaxy S4 loses power while on a computer USB port but charges correctly on a wall wart. I'm sure there's a reasonable explanation but haven't bothered to look into it seeing as how I'm full to the brim with warts of various sizes and colors.

NARF,
-l

Re:Battery

[Anonymous Coward]

i know my phone runs off battery even when plugged in.. when low, and plugged in to extend a call... it still runs out of juice.. just takes longer... the charge rate is slower than the power consumption of the device.

such a stupid fucking idea.. perfect example of: just because you can (do something) doesn't mean you should.

Re:Battery

[khellendros1984]

A USB port on a computer generally supplies 500mA of power. A wall wart is likely to supply double that amount or more. If your phone is draining its battery faster than the external power supply is charging it, the predictable thing happens.

Re:Battery

[fredgiblet]

It's a checkbox on advertisements. It's not that they need it, it's that it can be advertised as an advantage.





Also: Photoshop.

Re:Battery

[Anonymous Coward]

My phone overheats using Boinc as well. Given that lithium-ion performance is directly tied to the max temperature ever reached by the battery, this does concern me. Well, it would, except work paid for the phone and I have a feeling that if the battery goes out that they'll just pitch me a new phone rather than a new battery. Convenient!

Re:Battery

[khellendros1984]

Who said anything about a smartphone? The Nexus 7 is a tablet. Of course, that doesn't mean that there aren't 4-core phones.

The primary (theoretical) benefit wouldn't be for servicing the needs of the OS, but for providing more performance for demanding applications.

Re:Battery

[Luyseyal]

Yeah, that makes sense. I just wondered if I had a buggy phone or too much crap running. I did put it into developer mode and kill just about everything on a regular basis. My last phone was a Nokia E75 so all this USB charging is new to me.

-l

Re:Battery

[tlhIngan]

It will burn a hole in your table though.
Ever noticed how hot smartphones get when left running at full capacity for a while?

This is especially true on the A15-based smartphones and the "octacore". Truth is, they hit their max temperature (around 125C) after a few minutes if you peg all 4 high-powered (A15) cores in around 5 minutes.

From an analysis I saw, the software had to modulate at that point two of the cores to be around 50% utilization (so 2 going 100%, the other 2 at 50%) in order to maintain 125C junction temperature. If the chip still continued to heat, one core must be turned off and the other to 25%.

Basically, don't give it more than one core unless you put a heatsink on it... two cores will basically render the phone almost useless.

Re:Battery

[RicktheBrick]

I would estimate that the average user would be able to do one result a day. That would require about 8 hours of charging time since a lot of the fa@h work units require over 8 hours of computing time. IBM has already contributed over 49 million results and over 119 thousand yesterday alone. There are 6 members that have contributed around 10 million or more results so there is some series money involved here. There are over 900,000 results yesterday so even if WCG managed to get 100,000 new android users, that would mean maybe a 10% jump in results. Now lets compare that to the recent use of GPU power. The help conquer cancer project did manage to use the GPU and it was getting around a million results a day or more than what WCG is getting today. So a project that involve the GPU would contribute far more without adding a singe volunteer. I have read that WCG's member have the computing power comparable to around a petaflop. But the largest supercomputers now is 30 times that. To me that is telling me that this project is not worth the money it would take to get 5% of the time of the United States fastest supercomputer. I would think that a cure to aids would mean well into the billions of dollars in savings. So is the prospect of finding a cure so low that a number of countries would not contribute enough money for a super computer? They spent over 100 billion on the ISS and are spending a lot of money on the cern collider. The largest supercomputer probably is at least 100 times more energy efficient than the average home computer. This is a project that would cost a significant less total money on a super computer than the total amount being spent by the volunteers. It just requires the faith from a couple of governments. Incidentally since the article bragged about the contribution of the slashdot users, I will brag about me being the number one contributer of the slashdot users or at least I was a little while ago.

Re:Battery

[Luyseyal]

And the other thing is that if PCs start getting less popular (phone + KVM + RDP to the cloud or similar), BOINC will shrink. They're trying to cover their bets with this phone stuff. While I, the submitter, think it's cool and important, I seriously doubt people will want to spend much effort on it -- especially if PC ownership shrinks overall.

Also, you're talking about AIDS and other high profile projects. There are plenty of low profile, tiny projects that benefit a great deal from BOINC at a price point that is affordable.

FAAH has been going since 2005. They've scanned millions of particles against a number of receptor sites. And we have barely got a few molecules making it to the wet lab. I'm not sure that even a billion dollars on dedicated kit back in 2005 would have made similar progress simply due to logistics and the brute force game we're playing.

As you know, not everything works on the GPU. But it sure is cool when it does!

Thank you for your contributions. I've been on the team for a long time, too.

-l

Re:Battery

[hermitdev]

I'm trying to run Crysis? Sorry...couldn't resist.

Re:Battery

[Molochi]

Note to self: Design and market a bigass copper heatsink for mobile phones.

Re:Battery

[EnglishDude]

The phone communicates with the charger - if the phone can't do it, it'll max the charging at 500mA. My phone charges at a different rate when connected to different chargers. I'm sure you already know this, but just in case...

Re:Battery

[kav2k]

Better idea, a Qi charger over a running-water tank.

NativeBOINC

[SecretSquirrel33]

I have been using the Android port NativeBOINC for quite a while now. Just recently they even added support to my favorite program, Einstein@Home.

Re:NativeBOINC

[MickyTheIdiot]

I use NativeBOINC for my android devices that don't have batteries... it is actually a much better client than the new official client, but it has more problems with batteries.

Re:NativeBOINC

[Velex]

That's pretty neat! Mod parent up. Einstein@Home and SETI@Home here.

I just installed it on my phone (Tegra 3) and will be putting it on my tablet (Tegra 2) later, so I'll see how it goes.

Re:WCG on Android

[armanox]

Einstien@Home reports the same thing on my old Droid.

PrimeCoin

[dirtaddshp]

rather find primes with primecoin, and get paid while doing it :-)

Re:PrimeCoin

[MickyTheIdiot]

This is something I think I need to try. Does it use CPUs well, or is this another GPU usage thing...

How about

[koan]

Why not tie this sort of crowd sourcing to Bitcoin or the equivalent, where in the work (blocks) being done for the Bitcoin are actual research blocks.

Not sure if it's technically possible, but it would be interesting.

Re:WCG on Android

[skids]

Yeah this is pretty much par for the course: BOINC runs on just about everything, but individual BOINC applications authors generally won't bother to implement any code for anything other than the top dominant mixes of arch/OS. So if you have yourself a nice giant cluster of old G4s running linux, good luck putting it to use for something other than a very obscure but easily-implemented mathematical constant hunt.

The Idle Cycles Fallacy

[GameboyRMH]

This point gives me a chance to bring up the fallacy of "idle cycles" on modern processors.

There's no such thing any more. There was in the '90s and earlier when CPUs didn't have the power controls they do today, but nowadays your CPU uses exactly what it needs and anything more you give it to do will use extra energy.

So be aware that you're not putting any wasted resource to good use with these things. You're just using more resources. And on a phone that's the last thing you need.

Re:The Idle Cycles Fallacy

[Luyseyal]

Indeed, idle cycles nowadays simply refers to your phone doing all of jack and shit while it's plugged in and recharging all night... apart from checking your email, SMS's, waiting for that phone call, etc.

It does NOT refer to wasted energy cycles as it did in the past. Yes, it absolutely uses more power than it would if you were just charging (or better, you turned it off completely and charged it then).

It is an act of donating a few bucks a year to scientific research.
-l

Re:The Idle Cycles Fallacy

[mmcxii]

It is an act of donating a few bucks a year to scientific research.

Not to bust anyone's bubbles but I think that just donating the money instead of thinking you're a swell guy for letting them use your hardware would be a better gesture.

Re:The Idle Cycles Fallacy

[Luyseyal]

Or both? :)

If it was not worth it to the scientists, BOINC would not exist beyond a curiousity. Instead, many corporations and individuals have donated time and money to make the supercomputer of the people possible.

-l

Re:The Idle Cycles Fallacy

[fredgiblet]

I'm not certain I agree, idle cycles may not be as prominent as they once were, but procs don't slow to 0 when they're idle, they just ramp down a ways. My proce drops to (IIRC) 1.8Ghz at "idle" but I can assure you that my computer isn't using 4 1.8Ghz cores at idle.

Re:The Idle Cycles Fallacy

[Luyseyal]

Not necessarily. A lot of these smaller research teams would have to pay big bucks to get on a decent grid (within a reasonable research timeframe). BOINC affords them that with very little cost.

http://boinc.berkeley.edu/trac/wiki/VolunteerComputing

Why is volunteer computing important?

It's important for several reasons:

• Because of the huge number (> 1 billion) of PCs in the world, volunteer computing can supply more computing power to science than does any other type of computing. This computing power enables scientific research that could not be done otherwise. This advantage will increase over time, because the laws of economics dictate that consumer products such as PCs and game consoles will advance faster than more specialized products, and that there will be more of them.
• Volunteer computing power can't be bought; it must be earned. A research project that has limited funding but large public appeal can get huge computing power. In contrast, traditional supercomputers are extremely expensive, and are available only for applications that can afford them (for example, nuclear weapon design and espionage).
• Volunteer computing encourages public interest in science, and provides the public with voice in determining the directions of scientific research.

Re:The Idle Cycles Fallacy

[Anonymous Coward]

You're confusing/intermixing two different technologies, which is understandable given their roles: CPU clock throttling (which is the dynamic adjustment of clock frequency based on CPU load, ex. Intel SpeedStep or AMD's Cool'n'Quiet), and halting a CPU core or processor when there's nothing more for it to do (usually via the x86 HLT instruction).

Clock frequency adjustments affect the entire CPU (meaning all cores) and is hard to explain exactly. It doesn't happen when "all cores are unused/idle", but has a whole series of logic decisions it makes based upon stuff that's a bit outside my pay grade; meaning clock throttling can/will happen even when opcodes are actively being executed by a core. I'm certain Intel's architecture/processor documentation explain all these conditions, I just haven't taken the time to read it.

Now for HLT. History is important here: in olden days, there was no such opcode -- you'd "block indefinitely" by sitting in a tight loop (e.g. @foo: jmp @foo, poll VBlank status and wait until it began firing, etc.) or finding something else to do in the meantime. Since the introduction of a reliable HLT opcode (this was around the end of the 486 series -- earlier models had it but it was known to be buggy or wasn't used by the OS (ex. DOS)), the OS will execute HLT which will suspend that CPU core. The only way to "wake up" that core is via an IRQ (interrupt) (either hardware (NMI) or software) bound to that core (which is a whole separate subject; SMP affinity and masking of interrupts on x86 is almost a beast in itself). A complex example would be an Ethernet driver, where upon the physical NIC receiving a frame would generate an IRQ (usually via MSI/MSI-X these days) to tell the CPU core bound to that IRQ "wake up, I got something on the wire".

My point is that the two technologies have some relation, but they're separate things.

When BOINC Drive Out the Bugs...

[CAOgdin]

...that cause every Windows system to crash and have to be powered-off to resume...I'll consider putting it on my phone! --CAO

Re:When BOINC Drive Out the Bugs...

[Luyseyal]

I've never had BOINC crash Windows. Were you using the GPU? Nvidia or AMD?

-l

Re:When BOINC Drive Out the Bugs...

[roc97007]

...that cause every Windows system to crash and have to be powered-off to resume...I'll consider putting it on my phone!

--CAO

That was my concern also, and was the reason I very reluctantly stopped contributing to SETI when they switched to BOINC. Every once in awhile they plead with me to rejoin, and every time I say "still using BOINC? Then no." Once bitten, etc etc.

On my phone? Freaking kidding me. The phone is unstable enough as it is.

Re:on your phone?

[Luyseyal]

By default, it only works when it's plugged in, as pointed out ad nauseum here.

-l

WCG

[macraig]

If we had clean-as-in-free energy or I had a better income then I'd still be crunching for WCG. I stopped because I couldn't afford the extra 150W it caused my system to draw 24/7. I don't have a cellphone now because of the monthly cost, so they can't get my contribution that way either.

Re:WCG

[Luyseyal]

Been there and don't sweat it. Thank you for the time you were able to contribute.

-l

Gladly.

[Seumas]

Gladly... how much am I going to get paid?

Re:Gladly.

[Luyseyal]

Nothing. It's a donation. If you don't want to donate, it's your choice. :)

-l

Anybody got benchmarks?

[evilviper]

I'd be very interested to hear how the performance on a variety of smartphones compares to Intel/AMD CPUs.

Anybody got some benchmarks to share?

still stupid

[iggymanz]

silly to risk doing anything to battery on phone when most of have multiple computers at home including a workstation with multiple cores. project wants to borrow my kid's plastic beach sand shovel when I have a bulldozer or three around the house.

Pauses until 90 percent charged

[tepples]

I read the FAQ linked from the featured article. It pauses while the phone is not plugged into a charger, and it pauses until the battery is at least 90 percent charged.

Bill Watterson reference?

[N1tr0u5]

Scientific progress goes BOINC?