Message boards :
Number crunching :
Could BOINC app be written for Iphones?
Message board moderation
Author | Message |
---|---|
Pilot Send message Joined: 18 May 99 Posts: 534 Credit: 5,475,482 RAC: 0 |
I Phones and Crackberries etc. What a hoot:) It might even get the Macophiles to come back and could be a whole new category of contributors. When we finally figure it all out, all the rules will change and we can start all over again. |
arkayn Send message Joined: 14 May 99 Posts: 4438 Credit: 55,006,323 RAC: 0 |
|
Fred J. Verster Send message Joined: 21 Apr 04 Posts: 3252 Credit: 31,903,643 RAC: 0 |
|
CryptokiD Send message Joined: 2 Dec 00 Posts: 150 Credit: 3,216,632 RAC: 0 |
i have an etch-a-sketch that needs someone to port an app over to it. anyone up to the challenge? |
QSilver Send message Joined: 26 May 99 Posts: 232 Credit: 6,452,764 RAC: 0 |
I Phones and Crackberries etc. What a hoot:) Maybe. But why bother? |
SoNic Send message Joined: 24 Dec 00 Posts: 140 Credit: 2,963,627 RAC: 0 |
Port on iPod? You need a real multitasking SO first :) |
Gecko Send message Joined: 17 Nov 99 Posts: 454 Credit: 6,946,910 RAC: 47 |
Short answer. "sure" it could be written. However, then, someone would have to write a project application (or adjacent project aka iSeti ) for it as well ; >) that would process "shortened" work. If you did run a project you'd have 3 main issues. 1) Suck a battery dry w/in only a few hours (100% load). 2) Drastically reduced battery life span w/ much higher rate of charge/discharge cycles. 3) Battery heat? Lastly...the most important.... Performance would be SOOO painfully slow, (read: wayyy slower than even Intel Atom), that besides the fact that it would be totally undesirable to do it, likely not to meet most project deadlines for return of work. The only way for practical use would be shortened work units that would be specific to iPhone/Touch. There's validity to the methodology of this. Up-teen millions of potential hosts means that even if a project tapped only a small fraction, running shortened work, could still be a substantial contribution to the project. Today, the barriers and practicality are marginal, but in 12-24 short months, handhelds are likely to be substantially more powerful & better capable of necessary project performance. Battery life however, will still be the #1 problem and would relegate most crunching while charging in a dock overnite. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
I Phones and Crackberries etc. What a hoot:) Because there are so very many of them. |
Gary Charpentier Send message Joined: 25 Dec 00 Posts: 30987 Credit: 53,134,872 RAC: 32 |
I Phones and Crackberries etc. What a hoot:) For the iPhone the answer is a big fat no. The iPhone OS will not allow any background tasks. If you did it foreground, when the phone rings, BOINC would stop. When the call is over the user would have to restart the app manually. Then there is the battery issue. Dead real quick. I suppose crunch only when on the charger is a work around. Finally there is the speed issue. The iPhone is a slow CPU compared to a laptop. Yes it is a good thought, but we are a decade of Moore's law away from it being practical. |
Tom Philippart Send message Joined: 29 May 06 Posts: 23 Credit: 949,840 RAC: 0 |
ati support should clearly be a priority |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
I Phones and Crackberries etc. What a hoot:) Single device speed isn't a problem. That's the whole idea of distributed computing. The iphone uses the ARMv6KZ processor, which is also used for the nvidia goforce 6100. Not the most powerful chip, but it could do something. 50,000 cell phones would make a rather nice computing system I'd think. The big problem with using cell phones would be the power and heat. Most are not ment for running full tilt 24/7. I had some little farming game on my previous phone. I would let run overnight while it was plugged in. The phone died after about a year. Which could have had nothing to do with running the application so much, but I suspect it. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Cygnus X-1 Send message Joined: 15 Feb 04 Posts: 75 Credit: 3,732,505 RAC: 175 |
1) Suck a battery dry w/in only a few hours (100% load). The same issue exists on laptops and I don't see this as an obstacle since you can set Boinc to not run when on battery power. Just as with a laptop a phone app would be meant to run while plugged to a charger. I'd like to get a Nokia N900 at some point and when/if I do I'll install the ARM/Linux ports of Boinc and Seti. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
ati support should clearly be a priority I doubt that SETI@Home has heavy resources to devote to either platform, so the one that gets done first will likely be the one where a volunteer steps forward. |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
ati support should clearly be a priority I volunteer to nod aprovingly at any project that allows more processing power to be harvested. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
skildude Send message Joined: 4 Oct 00 Posts: 9541 Credit: 50,759,529 RAC: 60 |
unfortunately the chips in cellphones, as mentioned before are woefully inadequite for Seti since the TAT is about 1 month even on a short WU. I'd assume using a phone would be like using anything prior to the 200 MHZ processor level. IIRC the 200MHZ processor was the first to require a fan to keep the chip cool. In a rich man's house there is no place to spit but his face. Diogenes Of Sinope |
52 Aces Send message Joined: 7 Jan 02 Posts: 497 Credit: 14,261,068 RAC: 67 |
Applets tend not to be frameworks, rather they are.. um, stand alone applets. Downloading additional code is a no-no (one of many dictated by Apple). And presently, only one App can run (of course surrendering control as necessary for phone, sms, email, etc). An individual project (or Apple themselves for PR) could create a template iCharity Applet that consumes custom "conveniently sized" WU's (ie: can complete in 2 hours w/option to be continuous -- yes, less work per device, but 75 million iPhone/iTouch touting tree huggers who love trying out new apps in the top 25 would more than make up the difference). The CAUSE must resonate with the Apple masses. EX: Appstore:
Help Map Climate Change (part of iTunes iCharity) Help End Starvation With New Crops (part of iTunes iCharity)
|
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
Even a relatively old computer is several orders of magnitude faster than an iPhone for floating point calculations. The iPhones have no FPU which has been standard on laptops and desktops since the Pentium first came out. I once tested the difference of an 8086 with no FPU and an 8086 with an FPU on a strictly floating point calculation. Without the FPU I calculated that based on progress it would take about 6 months to complete (never completed as I was not that patient). With the FPU it took about an hour. Mind this was with what was otherwise exactly the same hardware. The iPhone clock rate is lower than that of a modern computer. All in all, I bet that the iPhone will take about 100,000 times as long to complete a Floating Point Intensive task as a desktop. I believe that someone actually ported SETI to a phone. It was discovered that no task from any project that they ported would come anywhere close to making the deadline. The only place that an iPhone might be able to make a good showing was some calculation that was all integer based (Prime Grid perhaps) or network intensive rather than CPU intensive. All in all, given the vastly reduced speed, is it really worth the effort to port a floating point intensive project to any hand held device. BOINC WIKI |
Jonathan Send message Joined: 14 Nov 08 Posts: 1 Credit: 15,205 RAC: 0 |
You might have difficulty getting such an app onto the iPhone App store, however Android and other phones are not as restrictive. I've thought of this a lot, there is a LOT of untapped processor power. The phone might only process while plugged in. I want it on Windows CE/PPC/Mobile too, I have an old 400MHz ARM Windows phone that usually justs sits "charging" I only take it with me for ereading or watching movies so it could be processing the rest of the time... it would help heat my room too :-) (I don't need to turn on the heat, turn on BOINC and my laptop puts out enough heat to keep it above 55 in the winter. Another device... routers. My router has almost as much CPU as my phone, and its only being used when I'm on the net. Its running linux, so the CPU must be pretty standard. |
Cygnus X-1 Send message Joined: 15 Feb 04 Posts: 75 Credit: 3,732,505 RAC: 175 |
I recently preordered a Pandora, I'm curious to see how it will perform with Boinc/Seti. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
All in all, I bet that the iPhone will take about 100,000 times as long to complete a Floating Point Intensive task as a desktop. Regarding SETI it could be not so bad if some algorithm change will be done. We get data for analisys from 2-bit samples. It should be possible to expand them not to FP single precision but to int32 for example and doing at least most part of pipeline in integer calculations instead of floating point. Moreover, AP for example demonstrates strong memory-bound behavior and doing mostly additions/comparisons that could be done in integer instead of FP. That is, maybe it's worth to compare no processor of handheld device but its memory access speed (afraid it will be too low too though). And regarding ~75M of cell phone users... IMO many if not all of them have "big" PC too. Question is - do we see those ~75M hosts here?... so, only smal fraction of cell phones will participate anyway, even if such port will be done... |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.