Multicore smartphones - how to use their full potential

Message boards : Number crunching : Multicore smartphones - how to use their full potential
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3

AuthorMessage
Profile Ageless
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 14314
Credit: 3,578,526
RAC: 654
Netherlands
Message 1907932 - Posted: 19 Dec 2017, 0:06:21 UTC

Direct comparison.

Task https://setiathome.berkeley.edu/result.php?resultid=6236351563
run time 225,892.29 CPU time 103,635.70

Task https://setiathome.berkeley.edu/result.php?resultid=6236425769
run time 227,298.17 CPU time 103,882.20

Not even going to do the ratios, you can see these are slower. Also armv7-vfpv3-nopie, but this time run with the display off. CPU temp as far as I saw, 30 - 35C.
Jord

Ancient Astronaut Theorists can tell you that I do not help with tech questions via private message. Please use the forums for that.
ID: 1907932 · Report as offensive     Reply Quote
Profile Pierre A Renaud
Avatar

Send message
Joined: 3 Apr 99
Posts: 90
Credit: 2,527,925
RAC: 6,314
Canada
Message 1911481 - Posted: 7 Jan 2018, 14:00:22 UTC - in response to Message 1906991.  

3 weeks l8r, on my Samsung TAB E tablet, my average turnaround time of 1.38 days increased to 1.52 and on my Polaroid A4 smarthone it increased from 2.23 to 2.96. I'll uninstall the apps shortly, since they are not improving things and aren't apparently forcing the devices in high-power state.

I should have controlled for WU types, but considering the low stakes involved and the short 3 week trial period, I chose not to bother with that.

I will test Keep Screen On for my aging Polaroid A4 android smartphone, which can disable my maximum imposed 30 minutes screen saver time (No Screen Off is simple to use, but can't bypass the 30 minutes imposed limit). In a week or two I'll be able to see if it can significantly improve on my Average turnaround time of 2.23 days.

On my Galaxy Tab E No Screen Off seems to work, and it remains to be seen if its Average turnaround time of 1.38 days can be improved upon.

So, the biggest question is what tools we have to adjust such power-saving features of Android devices?
Maybe device firmware can be fooled somehow to force in high-power state...
Over Christmas I'm planning to test with No Screen Off and if that doesn't work, with XPosed Framework so that the screen doesn't turn off - and hopefully the CPU doesn't either. I'll have to find something to put the phone in so the screen is covered.

   Team Carl Sagan on Facebook | Contact
ID: 1911481 · Report as offensive     Reply Quote
Karl Öberg

Send message
Joined: 8 Jun 99
Posts: 10
Credit: 20,528,694
RAC: 18,149
Sweden
Message 1911740 - Posted: 8 Jan 2018, 10:49:53 UTC - in response to Message 1911481.  

How did it go?
ID: 1911740 · Report as offensive     Reply Quote
Profile Pierre A Renaud
Avatar

Send message
Joined: 3 Apr 99
Posts: 90
Credit: 2,527,925
RAC: 6,314
Canada
Message 1911788 - Posted: 8 Jan 2018, 18:21:42 UTC - in response to Message 1911740.  
Last modified: 8 Jan 2018, 18:46:01 UTC

How did it go?
Not good: for my TAB E Samsung tablet, the daily WU output decreased by 10%, and for the Polaroid A4 smartphone it decreased it by 33%. The Android app retained for both was Keep Screen On (No Screen Off could not prevent the A4's screen from going blank after 30 mins).

Allowing their screens to turn off then probably insures that more ressources can be allocated to the SETI@home app without activating (or more than compensating) a sleep mode thought to be significantly detrimental to the SETI@home app's performance - at least for these mobile models.

I'm glad to report that all devices have now been returned intact to their standardized space-time fishy continuums.


   Team Carl Sagan on Facebook | Contact
ID: 1911788 · Report as offensive     Reply Quote
Previous · 1 · 2 · 3

Message boards : Number crunching : Multicore smartphones - how to use their full potential


 
©2018 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.