Boinc on smartphones and ePads for crunching s@h may not be productive for the volume of computation results returned, but it all may well be worthwhile for generating inspiration and learning as a form of 'outreach' publicity and teaching.

However, there are other non-compute-intensive Boinc projects that may well directly benefit greatly for science and discovery, beyond just the 'outreach' aspect.


Keep searchin',
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

Boinc on smartphones and ePads for crunching s@h may not be productive for the volume of computation results returned, but it all may well be worthwhile for generating inspiration and learning as a form of 'outreach' publicity and teaching.

However, there are other non-compute-intensive Boinc projects that may well directly benefit greatly for science and discovery, beyond just the 'outreach' aspect.


Keep searchin',
Martin

Given......carry on.

---

"Time is simply the mechanism that keeps everything from happening all at once."

Boinc on smartphones and ePads for crunching s@h may not be productive for the volume of computation results returned, but it all may well be worthwhile for generating inspiration and learning as a form of 'outreach' publicity and teaching.

However, there are other non-compute-intensive Boinc projects that may well directly benefit greatly for science and discovery, beyond just the 'outreach' aspect.


Keep searchin',
Martin

Kinda like I mentioned I think a project that focuses on the phones sensors instead of crunching anything, would be a great use of smartphones!

---

#resist

Oh, nobody's denying your want to TRY it.......
We just know you shall fail for all obvious reasons previously stated.

It's not a question of trying anymore, that's nothing that can fail for any reasons, BOINC actually works already on smartphones, see my posts above in this thread. And as you can see the performance is not that bad, it's comparable to a few years old CPU-only machine. Sure, not much compared to current high-end crunching rigs with multiple GPUs, but that's not what most people have staying around in their homes.

---

Thanks for those links, Link. I see that smart phone is running about 80 RAC right now. That puts it about number 14,400 on the list of 28,500 active MW users (according to BOINCstats). To look at it another way, we would need to add 16,540 more of these phones to make a 1 % increase in the current MW daily RAC (again using today's numbers from BOINCstats). For what it is worth, my dual Pentium II machine does about 80 RAC a day with MW on CPU only, running 12 hours a day and equally sharing the CPUs with Einstein and Climate Prediction.

This is a start, but only a start. I would really like to know if this is just a smart phone running in somebody's pocket, or is it some Frankenphone spread all over a work bench with power supplies, cooling, etc. plugged it to it.

---

Thanks for those links, Link. I see that smart phone is running about 80 RAC right now.

Well, the current RAC value is pretty much useless, it has not done any tasks in the last few days and probably never reached the possible maximum. According to the tasks, that it had in the list at the time when I posted it, the expected RAC (in case of running 24/7) would be around 310-330. And that was only using 1 of the available 2 CPU cores (whatever the reason was). So eventually up to 600 credits/day when using both cores.

---

+ if you get a wattometer I'm sure that's a very favourable credit:watt ratio.

E.g my Dell GX60 + Fujitsu laptop are each about 400RAC:70W.

If you want to mess about with low power computing why don't you get a $30 Rasberry Pi? Rather than devote $500+ cellphone to crunching you can power one of those boards off a wall-wart, drawing a couple of watts, with no cooling issue etc.

http://www.raspberrypi.org/

Similar ARM CPU, runs Linux off a SD card, has USB, Ethernet, video out etc. It's a proper little PC.

Ian

TEMPTING!

ok, this is related to my smart phone which can be called a stupid phone atm because I got the AndroBOINC app (to connect to my client).
When i open it, I have to configure it to a new host, and I have no clue as to how to do that. I was wondering if anyone here knows what to do?

It is asking for Host Nickname, Address and password, using port #31416

Thanks!

If you want to mess about with low power computing why don't you get a $30 Rasberry Pi? Rather than devote $500+ cellphone to crunching you can power one of those boards off a wall-wart, drawing a couple of watts, with no cooling issue etc.

http://www.raspberrypi.org/

Similar ARM CPU, runs Linux off a SD card, has USB, Ethernet, video out etc. It's a proper little PC.

Ian

I would think the idea is to use old phones that you are no longer using or a current current phone while plugged in/charging. Instead of purchasing a dedicated ARM cruncher. Granted there are some that would do it just for the fun of setting up 50 of these little guys.

The processor on that looks to be about the same as what I have in my current phone. If I can ever get the Native-BOINC to run the apps they made for it on android it would be nice to compare the two.

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

I never bought an ARM or Altera-Cyclone development board that I wanted to.

Installing Linux/Boinc on one however does sound fun... I better ask the wife.
*"Hunny can I buy..."*

---

#resist

When i open it, I have to configure it to a new host, and I have no clue as to how to do that. I was wondering if anyone here knows what to do?

It is asking for Host Nickname, Address and password, using port #31416

I have posted your question in the Boinc Client for Android thread at Milkyway forum, since you (as many others) probably would have some difficulties to register in the BOINC@Poland forums due to the language for to aks that there.

BTW: not bad for a phone (HTC EVO 3D).

---

IIRC MW does not use the FPU which is the reason for the relatively fast times. Seti is a completely different animal

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

When i open it, I have to configure it to a new host, and I have no clue as to how to do that. I was wondering if anyone here knows what to do?

It is asking for Host Nickname, Address and password, using port #31416

I have posted your question in the Boinc Client for Android thread at Milkyway forum, since you (as many others) probably would have some difficulties to register in the BOINC@Poland forums due to the language for to aks that there.

BTW: not bad for a phone (HTC EVO 3D).

Thanks Link!

IIRC MW does not use the FPU which is the reason for the relatively fast times. Seti is a completely different animal

Perhaps I should have run that first instead of primegrid. As I'm at 1% after 4 hours.
On the plus side 4 hours of it crunching with the screen on only took the battery from 100% to 40%. At which point it didn't get warm at all.
However plugged in and charging it is now generating some heat. So I took an old P4 heat and turn it upside down and then sat the phone on it. To help draw some heat off of the battery.

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

which primegrid project?

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

which primegrid project?

Cullen/Woodall Prime Search (Sieve). Which is the only one they have an application for right now.

Edit: Something a bit odd I noticed with Android System Info. It shows 100% CPU usage, but doesn't list the science app that is doing the processing. Does anyone use or know of other good "task manager" type apps for Android?

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

the average turn around time on a Cullen woodall sieve is 12.5 hours. I'd say you are going very very very slow. 4 hours for 1% means that for a 12 hour WU it will take you 400 hours. that's around 16.6 days to complete.

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

the average turn around time on a Cullen woodall sieve is 12.5 hours. I'd say you are going very very very slow. 4 hours for 1% means that for a 12 hour WU it will take you 400 hours. that's around 16.6 days to complete.

After a longer run time it comes out to to more like 1.68% per 4 hours or .007% per minute. Which comes out to about 9.9 days. With a task limit of 7 days that is just a bit long.

Doing some math. The RaspberryPi system would probably take about 8.4 days. If it's 700MHz vs my 600MHz is directly comparable. So the sub GHz range may not be viable for a large range of projects at this time. With a 1GHz processor taking an estimated 5.9 days for that task.

I'll see what kind of results I get with MW next. If that doesn't work our well either then hopefully one of the new phones I'll looking at getting this year will with their ~1.5GHz quad-core goodness.

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[