So it's been close enough to two weeks of pretty much just Rosetta@Home work units (although I did give a few other projects a try for a bit), and.... I don't know, it's hard to say what my actual thoughts are about running it on one's phone. Obviously I wind up comparing it against my 4770k which as of late has been averaging around 690 PPD per thread (but is showing a RAC of 1,074.35 due to the fact I'm currently running ten projects on it (two of the POTM's, three that I'm using to try and get more bronze stars in WUProp, and my five regulars).... But in all honesty, if you look at it from a PPD/thread perspective (I can't compare PPD/watt due to lack of a kill-a-watt or other power meter) it technically isn't that bad....
4770k: 690 PPD/thread (currently displayed RAC for the machine: 1,074.35)
HTC One M8: 151.06 PPD/thread (currently displayed RAC for this device: 302.12 - not bad for a phone that's only crunching at most 4.5 days a week)
Samsung Galaxy S3: 149.36 PPD
Interestingly enough, I either didn't have enough data for my previous estimated RAC/PPD calculations, work units are different, getting your RAC to stabilize takes longer than two weeks, or it could very well do with how the Samsung modified 4.4.2 Android tends to basically just shut itself down and stop crunching if I don't turn the phone's screen on at least every 12 hours. Hard to say, but there are a wide variety of variables at play and I'm mainly just doing this as a long term experiment to justify the cost of a few developer boards (Raspberry Pi 3's, UP Board 4GB/32GB, ODroid C2 (and maybe the XU4's too) and the UDOO x86 Advanced/Ultra editions).
Anyways, I am still surprised though that my estimates for the Galaxy S3 are 3.45x higher than what I'm seeing, and for the HTC One M8 the estimates were 3.03x higher than what's currently showing (then again, I calculated the estimated PPD based upon 24/7 BOINCing, and the weekly on 4.5 days so it's somewhat skewed). Either way, here's the data that I have so far on the two phones - at least what I can pull from the Rosetta@Home website right now, since I didn't fully document it all. I probably should do that, just pull all the data every day, have it in a Google Sheets spreadsheet and then be able to pull from a much larger pool of data... That would definitely help me narrow down a few variables!
.... ....Oh, pulling up the computer details reveals an interesting fact about my S3 - it's only crunching 66.8% of the time, I'll have to find out why this is happening.
...Samsung Galaxy S3...
Max threads running: 1 (I'm looking into seeing if it's possible to get a custom kernel for this phone that will let me use both)
Tasks completed: 6
Runtime: 5,744.66 - 21,951.30 seconds (average is 13,976.94 seconds)
Credit: 62.58 - 145.16 (average is 87.11)
Estimated PPD based upon averages: 538.48
Estimated weekly points output: 3,769.36
...HTC One M8...
Max threads running: 2
Tasks completed: 14
Runtime: 8,211.93 - 26,786.88 (average is 16,912.64 seconds)
Credit: 28.48 - 154.72 (average is 77.4)
Estimated PPD based upon averages: 790.81
Estimated actual PPD when factoring in only 4.5 days of crunching due to work: 508.38
Estimated weekly points output: 3,558.65
It is interesting to me to see that a good portion of the averages have remained similar, and that despite the current issues it seems to regularly face, my Galaxy S3 has been fairly steady, and even increased it's estimated PPD somewhat. What is also interesting is that my HTC One M8 has had it's estimated 24/7 PPD and especially PPW numbers drop, but I suspect part of that is because I'm also playing Monster Legends on a regular basis, and the two don't play well together for some strange reason. It's almost like Monster Legends at full screen brightness makes the phone hot or something and I turn off BOINC while I'm playing the game..... .....And then occasionally forget to turn BOINC back on when I exit out of the game
EDIT (10 July 13:34 -06:00 GMT): Got a Google Spreadsheet going for the data. I'll do my best to update it daily, but right now I can't figure out why calls to the individual sheets aren't working so I'm having to manually enter data for Daily/Weekly/Monthly stuff. Take a look here if you're curious
EDIT 2 (10 July 15:24 -06:00 GMT): Oh..... There's an advanced section in the preferences..... And I'm now running two cores on the S3, four cores on the M8.... And allowed my devices more usage of the CPU and RAM. Let's see what happens over the next few days ^_^Edited by WhiteWulfe - 7/10/16 at 2:25pm