Message boards : Number crunching : Rosetta and Android
Author | Message |
---|---|
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1994 Credit: 9,594,538 RAC: 8,678 |
Some weeks ago, on Ralph@Home home page: "We are happy to announce that today we added the android arm platform minirosetta_beta version. If you have an android arm device/phone that supports android-9, please download and install the boinc client and attach to our project." (but no wu for android up to now). When do you think to pass Rosetta on Android?? As i ask: Will you create a "specifical counter" in the server status for android wus? |
sgaboinc Send message Joined: 2 Apr 14 Posts: 282 Credit: 208,966 RAC: 0 |
while i'd not like to disappoint the android fans it is difficult to imagine running rosetta on a mobile phone, i guess the battery would be flat before the task ends :o :p lol |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1994 Credit: 9,594,538 RAC: 8,678 |
while i'd not like to disappoint the android fans it is difficult to imagine running rosetta on a mobile phone, i guess the battery would be flat before the task ends :o :p lol The battery it's not a problem. You can configure it to crunch when phone is charging. Other project use boinc on android... The problem, i think, is memory. 500/600 mb per core on phone/tablet is too much |
sgaboinc Send message Joined: 2 Apr 14 Posts: 282 Credit: 208,966 RAC: 0 |
i see, thanks for pointing out :D yup, agree too that for most consumer devices today it would be pushing the limits of memory, storage, processor prowess etc to run r@h on those devices. lol :D |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1994 Credit: 9,594,538 RAC: 8,678 |
i see, thanks for pointing out :D There are some projects on android (wcg, simap, asteroids, seti, einstein, ecc) so is not impossible to crunch on mobile device. The problem is to optimize the code. |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1994 Credit: 9,594,538 RAC: 8,678 |
Now i'm crunching a ralph@home's wu on my smartphone!! |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1994 Credit: 9,594,538 RAC: 8,678 |
Not so good the android version. There will be 3 possibilities: 1) Optimize the code to reduce the usage of ram over 50% 2) Use Android version to run a very specific kind of simulation 3) A waste of time and resource Remain of view that a gpu client is better :-P |
David E K Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 1 Jul 05 Posts: 1018 Credit: 4,334,829 RAC: 0 |
We probably will have an android specific app version that will get specific jobs that don't require as much memory. Looking towards the future, memory, processors, etc.. will get better. This is just the beginning. |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1994 Credit: 9,594,538 RAC: 8,678 |
Looking towards the future, memory, processors, etc.. will get better. This is just the beginning. I know this is just the beginning, but i remain skeptical about boinc over android (not only for rosetta project). For example, a high-level ARM cpu (A9) has the same computational power of a single core cpu of 2004 and 25/30 times less than a medium I7... |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1994 Credit: 9,594,538 RAC: 8,678 |
|
David E K Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 1 Jul 05 Posts: 1018 Credit: 4,334,829 RAC: 0 |
Wow, ARM cpu are the new rac monsters :-P I'm just happy to have a relatively stable version running on Ralph now :). I think the CPU time reporting is not consistent with the actual CPU usage time so the granted credits are not correct. At least it isn't on my device. I think the BOINC client has some bugs/kinks to work out. Android puts the tasks to sleep often even though the BOINC client assumes the tasks are running. Things will improve with time and effort... |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1994 Credit: 9,594,538 RAC: 8,678 |
I'm just happy to have a relatively stable version running on Ralph now :). I think the CPU time reporting is not consistent with the actual CPU usage time so the granted credits are not correct. After some days of android:
:-( |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1994 Credit: 9,594,538 RAC: 8,678 |
Rosetta has released offical app for android (version 3.58) https://boinc.bakerlab.org/rosetta/apps.php Has someone tried it? |
FalconFly Send message Joined: 11 Jan 08 Posts: 23 Credit: 2,163,056 RAC: 0 |
Rosetta has released offical app for android (version 3.58) I'll give it a shot for a few Tasks, will report back tomorrow :) |
FalconFly Send message Joined: 11 Jan 08 Posts: 23 Credit: 2,163,056 RAC: 0 |
Got the Android Client running over night - and the results aren't good. The Application basically starts off good but apparently ceases to utilize CPU within seconds. After that, the Task sits at what looks like idle and eventually times out after wasting about 150% of the alotted target CPU time. It basically doesn't process anything, just blocks a CPU core and fails. I've just restarted a Task that was sitting idle for >9 hours on a 6 hours task - at least right now it seems to actually process data. Let's see for how long... RAM usage seems very low, my Android Device has plenty of free RAM remaining while processing 2x SIMAP + 1x Rosetta. With a bit of luck it was only the 1st tasks that failed (maybe some project init problems on the device). |
FalconFly Send message Joined: 11 Jan 08 Posts: 23 Credit: 2,163,056 RAC: 0 |
Allright, the conclusion after a day of testing is : the Android Client just doesn't work. The 2nd workunit apparently completed without issues (nothing bad visible in logfile either) but ended up as a computing error as well - just like the one of my wingman. |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1994 Credit: 9,594,538 RAC: 8,678 |
I continue to consider "rosetta android" a waste of time. Much more sense, for example, introduce avx/avx2 extension in the rosetta code. |
xdarma Send message Joined: 20 Jan 08 Posts: 5 Credit: 5,043,924 RAC: 2,766 |
Much more sense, for example, introduce avx/avx2 extension in the rosetta code. +1 IMO, the lack of support for AVX is a loss of potential starting from today. The lack of support for ARM is (maybe) a loss from 2016. In my opinion, the development priorities should be revised. Thanks. |
FalconFly Send message Joined: 11 Jan 08 Posts: 23 Credit: 2,163,056 RAC: 0 |
Generally I wouldn't discount the Android Client, not for its performance potential per device - but due to the sheer number of devices. But of course, if Rosetta could be optimized for the latest SIMD or even OpenCL, that would be a massive performance gain... |
David E K Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 1 Jul 05 Posts: 1018 Credit: 4,334,829 RAC: 0 |
Yes, there are definitely issues with android and boinc apps. The main issues now I believe are with the BOINC client and current android versions which put background processes to sleep. For now, I am not going to spend much time on our android version until they fix this issue. The motivation for an android arm version has come from BOINC and their partnership with HTC power to give. Samsung is also interested in running R@h on their phones. |
Message boards :
Number crunching :
Rosetta and Android
©2024 University of Washington
https://www.bakerlab.org