Attend XDA's Second Annual Developer Conference, XDA:DevCon 2014!
5,731,053 Members 46,447 Now Online
XDA Developers Android and Mobile Development Forum

[Q] Highest score on your antutu benchmark

Tip us?
 
sleekmason
Old
(Last edited by sleekmason; 23rd February 2014 at 10:53 PM.)
#21  
sleekmason's Avatar
Senior Member
Thanks Meter 183
Posts: 382
Join Date: Dec 2013
Location: Fayetteville
@whoamigriffiths is right that it is a number of small tweaks on the system that make the difference. Here's my two cents.

The device comes stock with performance governer as the default. The switch to Ondemand should show no difference in benchmark tests as Ondemand will ramp up under pressure to the highest state possible. Ondemand for our device may provide slightly better battery life as it allows the cpu to operate at levels below the max where performance does not. It may also extend the life of the device, but that probably won't matter to anybody. Using interactive requires more than just switching in on as default. If the device manufactures didn't pre-tweak it for the device it might not perform well. mileage may vary.

As far as the stuttering goes, this is affected more by the scheduler, and by writes to the log/cache/etc taking up time/energy. In our device, Deadline is far better than any of the other schedulers and will smooth things out considerably. You can think of deadline as being just that. It tells the device how much it needs for any given application(s) and wants it now! no messing around! lol.

I would not set your cpu to the maximum frequency for your minimum value. Because of thermal throttling on our device, you will see a negative value for performance as well as battery life. Raising it say, from 384 to 594 might feel slightly snappier, but probably not. Devices that benefit from this are generally phones with a poor processer, with a minimmum default in the 128/256 range. For ours, you are just wasting battery.

The afore mentioned fix for Dalvik and Bionic may well work for the 500 device, but for those on the 510 as whoamigriffiths said in an earlier post, you shouldn't bother as it is already optimized for the device. I tried both and noticed a larger battery drain and poor benchmark results down in the 23000-24000 range. My device averages about 26500 using the sleekai3 kernel with no patches.

Changing the delay time in the developer section for screen functions is a good idea. I think they put it there just to make us feel better.
The Following User Says Thank You to sleekmason For This Useful Post: [ Click to Expand ]
 
sleekmason
Old
(Last edited by sleekmason; 27th February 2014 at 04:54 PM.)
#22  
sleekmason's Avatar
Senior Member
Thanks Meter 183
Posts: 382
Join Date: Dec 2013
Location: Fayetteville
Quote:
Originally Posted by whoamigriffiths View Post
Haha, using dyn kernel r3and mahdi rom



Sent from my LG-V500 using XDA Premium HD app
Looks like that zip-align bit may make a difference after all! 27474

Knocking out a $600 device feels pretty good.

Still out of the box, running sleekai4 w/ init.d support.
Attached Thumbnails
Click image for larger version

Name:	antutu22714.png
Views:	55
Size:	200.6 KB
ID:	2603976  
 
whoamigriffiths
Old
#23  
whoamigriffiths's Avatar
Senior Member
Thanks Meter 630
Posts: 894
Join Date: Oct 2012
Location: Behind You
Quote:
Originally Posted by sleekmason View Post
Looks like that zip-align bit may make a difference after all! 27474

Still out of the box, running sleekai4 w/ init.d support.
Imba! Nerf KK! :P

Sent from my LG-V500 using XDA Premium HD app
"It's psychosomatic. You need a lobotomy. I'll get a saw."
The Following User Says Thank You to whoamigriffiths For This Useful Post: [ Click to Expand ]
Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes