Would it be possible to build this application for a generic ARM Linux platforms like arm-unknown-linux-gnueabi (without hardware support for floating-point, like ARMv5) and arm-unknown-linux-gnueabihf (with support for floating-point, like ARMv6)?

Though ARM is not a high-performance processor by today's standards, it may be as fast as the typical PC of a few years ago and comparable to a current Intel Atom.

Other projects like Enigma, OProject, QCN, Radioactive, Yoyo and WUProp already provide an application for such platforms.

I've helped out other projects getting the applications built and tested, as can be seen here. Please, let me know if I can help.

TIA

---

It is possible that such a MW client could be coded, but it unlikely in view of the poor outputs of such chips (I believe). MW crunching power can only come from the application of 100,000s of such devices, and who wants to crunch on their mobile/smart phone?

---

Go away, I was asleep

You might want to look at that thread and eventually contact the one, who already has a working ARM application.

---

I see that Rastemier got a version working on his Galaxy for MW.

I, personally, think it might overburden a mobile, but it is interesting that it has been done. Moreover the numbers of ARM, and other chip based mobiles well exceeds the combined PC and Mac desktop/laptop numbers.

---

Go away, I was asleep

Might be worth to give it a try ...
http://www.hardkernel.com/renewal_2011/products/prdt_info.php?g_code=G135341370451&tab_idx=1

It is possible that such a MW client could be coded, but it unlikely in view of the poor outputs of such chips (I believe). MW crunching power can only come from the application of 100,000s of such devices, and who wants to crunch on their mobile/smart phone?

It's not as bad as it may look; I joined last weekend with my Nexus, installed native Boinc.
As the first project I have choosen theSkyNet POGS. You can check the performance here:
http://ec2-23-23-126-96.compute-1.amazonaws.com/pogs/results.php?hostid=6061
One wu took 6,7 x longer than an i5, and I can run 4 wu's at a time.

I also tried 2 MW wu's; they ran more than a day.

Good to see other devices crunching. One of my friends has converted his smart phone to run work as well ...

---

Go away, I was asleep

Moreover the numbers of ARM, and other chip based mobiles well exceeds the combined PC and Mac desktop/laptop numbers.

Most of the ARM chips out there are dog slow though. Modern cellphone ARM CPU's are the near pinnacle of commonly available ARM CPU's too BTW. Not really worth the effort from a performance stand point but would be a good learning experience for those looking to try their first recompile I guess.

Moreover the numbers of ARM, and other chip based mobiles well exceeds the combined PC and Mac desktop/laptop numbers.

Most of the ARM chips out there are dog slow though. Modern cellphone ARM CPU's are the near pinnacle of commonly available ARM CPU's too BTW. Not really worth the effort from a performance stand point but would be a good learning experience for those looking to try their first recompile I guess.

Yes BUT, there is ALWAYS one of those I think, with some umpteen BILLION cellphones in the World together they could easily surpass the crunching power of the current number of pc's crunching! The key again is to get people interested so they will at least try it.

Yes BUT, there is ALWAYS one of those I think, with some umpteen BILLION cellphones in the World together they could easily surpass the crunching power of the current number of pc's crunching! The key again is to get people interested so they will at least try it.

Give the app an intresting skin, something that they like to show their friends. This is how I found SETI 10 years ago when I saw the 'screensaver' of my hair-cutters PC.

Yes BUT, there is ALWAYS one of those I think, with some umpteen BILLION cellphones in the World together they could easily surpass the crunching power of the current number of pc's crunching! The key again is to get people interested so they will at least try it.

Give the app an intresting skin, something that they like to show their friends. This is how I found SETI 10 years ago when I saw the 'screensaver' of my hair-cutters PC.

Agree..Malaria used to have one where it was from the perspective of a mosquito flying looking for something to bite. I remember MANY years ago there was a pc screensaver where a guy was on an island and every day something happened to him. One day some coconuts came by, another day a large wave came by, etc, etc. Make it interesting and people WILL run it!!

Yes BUT, there is ALWAYS one of those I think, with some umpteen BILLION cellphones in the World together they could easily surpass the crunching power of the current number of pc's crunching! The key again is to get people interested so they will at least try it.

Most of those ARM CPU's are pretty weak, meant to use almost no power for dedicated tasks which don't require much at all, a 386 would probably be far faster. Its only with the modern "super phones" you're going to get anything resembling "decent" performance. Due to "super phone" cost not many in the world actually can afford to have one...

Yes BUT, there is ALWAYS one of those I think, with some umpteen BILLION cellphones in the World together they could easily surpass the crunching power of the current number of pc's crunching! The key again is to get people interested so they will at least try it.

Most of those ARM CPU's are pretty weak, meant to use almost no power for dedicated tasks which don't require much at all, a 386 would probably be far faster. Its only with the modern "super phones" you're going to get anything resembling "decent" performance. Due to "super phone" cost not many in the world actually can afford to have one...

You are correct, but we are at the typical business model scenario...is it better to make a 1 dollar profit on 100 items or a 100 dollar profit on 1 item? With the 1 dollar profit people will be flocking to your door and you will be swamped with customers, with the 100 dollar profit price very few people will even come in and look. Cellphones are the 1 dollar profit price, they are billions of them, so each alone won't be able to do much, but together they can do alot. I person cannot easily flip a car over, but half a dozen people can do it with no problem, more hands makes lighter work for all, etc, etc. BUT it MUST be fun for them, VERY short units with some interesting graphics going on, or people will download it and then delete it almost immediately.

You are correct, but we are at the typical business model scenario...is it better to make a 1 dollar profit on 100 items or a 100 dollar profit on 1 item?

Depends on the item of course.

Many of these ARM CPU's are effectively inaccessible to anyone but the most dedicated number cruncher. We're talking about stuff like micro controllers for autoclaves, washing machines, cars, etc. The vast overwhelming majority of people won't even have a clue as to how you'd even begin getting BOINC up and running on such a device. In many cases not only would drastic software rewrites be required but you'd probably be forced to do some soldering to access the buses and/or BIOS chips on the devices to even begin communicating with them.

With the bar set so very high the only model that could work ("1 dollar profit on 100 items") is now ruled out for accomplishing significant work, much less surpassing or even equaling the number crunching power of current PC's/GPU's.

You are correct, but we are at the typical business model scenario...is it better to make a 1 dollar profit on 100 items or a 100 dollar profit on 1 item?

Depends on the item of course.

Many of these ARM CPU's are effectively inaccessible to anyone but the most dedicated number cruncher. We're talking about stuff like micro controllers for autoclaves, washing machines, cars, etc. The vast overwhelming majority of people won't even have a clue as to how you'd even begin getting BOINC up and running on such a device. In many cases not only would drastic software rewrites be required but you'd probably be forced to do some soldering to access the buses and/or BIOS chips on the devices to even begin communicating with them.

With the bar set so very high the only model that could work ("1 dollar profit on 100 items") is now ruled out for accomplishing significant work, much less surpassing or even equaling the number crunching power of current PC's/GPU's.

There is already available, and released, Boinc software for the ARM devices and it is working right this minute. Here is a link to the release forum showing the changes from version to version:
http://boinc.berkeley.edu/dev/forum_thread.php?id=8203

Several projects have ARM devices actively crunching right this minute for them.

There is already available, and released, Boinc software for the ARM devices and it is working right this minute. Here is a link to the release forum showing the changes from version to version:
http://boinc.berkeley.edu/dev/forum_thread.php?id=8203

That won't work on autoclave, car, washing machine ARM micro controllers, etc. which are still the vast overwhelming majority of ARM devices out there. Most of them don't even have a OS per se much less Android running on them or are even capable of running Android. If they have a OS at all it'll be something like QNX RTOS and stripped to the bare bones with only a command line interface. A straight forward recompile won't even work half the time since many are semi-custom CPU's with no public documentation out there.

I have no doubt you could eventually get something running on these sorts of devices, that isn't in question, but the effort vs. reward isn't there even if you got BOINC running on dozens of them for every PC you had. And even if you were still willing to go through with it that would apply to you and a few other dedicated crunchers. Almost no one else is willing to go through that effort to crunch a handful more WU a day.

Most people can't even be bothered to mod chip their own console and that is way easier than what you're talking about doing. For that matter most people aren't even bothering to run BOINC on their home PC which is pretty damn easy to do.

Several projects have ARM devices actively crunching right this minute for them.

That doesn't mean much of anything though and certainly wasn't in dispute either. None of these projects are even close to having billions of ARM devices running BOINC of any sort for a reason.

There is already available, and released, Boinc software for the ARM devices and it is working right this minute. Here is a link to the release forum showing the changes from version to version:
http://boinc.berkeley.edu/dev/forum_thread.php?id=8203

That won't work on autoclave, car, washing machine ARM micro controllers, etc. which are still the vast overwhelming majority of ARM devices out there. Most of them don't even have a OS per se much less Android running on them or are even capable of running Android. If they have a OS at all it'll be something like QNX RTOS and stripped to the bare bones with only a command line interface. A straight forward recompile won't even work half the time since many are semi-custom CPU's with no public documentation out there.

I have no doubt you could eventually get something running on these sorts of devices, that isn't in question, but the effort vs. reward isn't there even if you got BOINC running on dozens of them for every PC you had. And even if you were still willing to go through with it that would apply to you and a few other dedicated crunchers. Almost no one else is willing to go through that effort to crunch a handful more WU a day.

Most people can't even be bothered to mod chip their own console and that is way easier than what you're talking about doing. For that matter most people aren't even bothering to run BOINC on their home PC which is pretty damn easy to do.

Several projects have ARM devices actively crunching right this minute for them.

That doesn't mean much of anything though and certainly wasn't in dispute either. None of these projects are even close to having billions of ARM devices running BOINC of any sort for a reason.

You are correct on ALL counts, but I would like to point out that when Boinc, or more exactly Seti, first began it started with a handful of Scientists and then grew into the millions of users.

This is my Nexus 7
http://ec2-23-23-126-96.compute-1.amazonaws.com/pogs/show_host_detail.php?hostid=6061
A RAC of 660 - is that poor? Running on 2 of 4 cores.
The number of tablets is increasing. Make an impressive skin and you will have more and more volunteers.

You are correct on ALL counts, but I would like to point out that when Boinc, or more exactly Seti, first began it started with a handful of Scientists and then grew into the millions of users.

There is certainly a lot of potential in the future for ARM chips to become significant contributors. I just don't think its going to come from portable/battery operate ARM devices nor other integrated ARM devices.

I would love to have something like a HPC many core ARM chip in desktop form to crunch on but those don't exist on the consumer level of buying power and aren't likely too either any time soon.

A RAC of 660 - is that poor?

Depends on your definition of poor of course.

In terms of performance vs power consumption or performance vs price on a chip to chip basis its great. In terms of total performance vs modern-ish x86 chips IMO it'd be poor.

Running power hungry apps on devices that are very power limited is unlikely to take off in a big way...

Really I swear I'm not trying to piss on your guys enthusiasm/dreams here but there must be at least some thought given to practicality when forming your expectations.

Really I swear I'm not trying to piss on your guys enthusiasm/dreams here but there must be at least some thought given to practicality when forming your expectations.

But when Science is included practicality often goes out the window. If it works go for it and figure out how to make it efficient later on.