Is ocing the a72 cores worth it?
imho, not really.
I don't really think that it would bring a noticeable improvement in performance.
But why? Is something working not smooth enough? Or just for antutu points?
If you want to bench , it's not very effective :
I'm on cleanrom with zetsoubou (allowing oc to 2000mhz ) and i tried with and without oc (performance governor of course ) and for 90% of the bench score is Ux scoring and overall optimisation of the rom
81k antutu with 1800 , 83k on 2000mhz
2910 on vellamo (1800mhz ) vs 2970 (2000mhz ) vs 3100 without oc on radon 3.2 and f2fs /system partition
Related
is it possible to overclock GPU? when i had xoom the GPU was OCed to 400mhz from 300.. and with 1.6 ghz CPU it was super good..
with the touchwiz update Gtab is better even with the stock GPU and CPU but i still want to OC -especially- the GPU.. i believe it will help in the games.. for example dungeon hunter 2 hd is so laggy.. is it possible -and easy- to make a kernel just capable of OC ing for both GPU and CPU?
and one more thing; i think there are two limitations of GPU.. if the moving objects are not too fast the device tries to lock the FPS to 30, if they are fast enough it is limited with 60 FPS.. this is whan i can observe from quadrant test.. so, is it also possible to unlock this 30FPS limitation and let the GPU try to handle everything in 60 FPS?
just my observation: Quadrant run for touchpad only gives score of around 2400... my nook color with 1.3ghz single core gives score of over 2600 all the time!!! i looked into the device info, it seems like it only using 1 core!!! does cm7 only recognizing one core??? what r ur guys thought??? is there a fix for it or any kernels that utilize both cores of touchpad!!!
The Quadrant results on the TP are skewed because the GPU code has not yet been implemented (or optimized) which primarily affects the FPS section (looks like is is only doing about 4 frames per second) where all the other graphics demos (especially the planet one) do 20-60 frames per second which is a 300%-600% improvement over the Nook Color running 1.2GHz.
I would be interested in other benchmarks that are not dependent on the GPU code.
try using Antutu - free from market
it separates the scores according to each test
(also shows total)
i got around 5000 - OC 1.7 ghz
Maybe it is because we are running an un-optimized alpha build...Don't worry about synthetic benchmarks anyway.
Both cores are already being used, what is your processor clocked too?
I'm completely aware benchmarks don't really mean much, but for curiosities sake, out of 3 runs, I averaged 3228 in quadrant. I am overclocked to 1782mhz ondemand.
I suggest you run SunSpider also.
So I got an S2, running a custom ICS Samsung based ROM, I want the best gaming performance, I'm using this configuration for ExTweaks:
Scaling_MAX_FREQ:1.2GHZ
Scaling_MIN_FREQ:500MHZ
GPU:
Step1:108MHZ 900MV
Step2:160MHZ 950MV
Step3: 267MHZ 1V
Step4: 330MHZ 1.1V
Step5: 400MHZ 1.15V
And I'm still getting lag spikes on games like GTA 3.. What can I improve? And are these settings I've listed above are safe? I'm not concerned about battery life.
Thank you.
LifeLessGT said:
So I got an S2, running a custom ICS Samsung based ROM, I want the best gaming performance, I'm using this configuration for ExTweaks:
Scaling_MAX_FREQ:1.2GHZ
Scaling_MIN_FREQ:500MHZ
GPU:
Step1:108MHZ 900MV
Step2:160MHZ 950MV
Step3: 267MHZ 1V
Step4: 330MHZ 1.1V
Step5: 400MHZ 1.15V
And I'm still getting lag spikes on games like GTA 3.. What can I improve? And are these settings I've listed above are safe? I'm not concerned about battery life.
Thank you.
Click to expand...
Click to collapse
i play gta 3 at much lower than that and no lags at all...perhaps u should use "mali status app" to see which frequencies are being used during game play
i notice u using 330 as well so i assume u using siyah with vpll enabled...try disabling vpll
I have seen the benchmarks of htc one v, it has the same specs as lww, mini etc. , the quadrant score for the i/o was 2000 approx, but it is 1000 approx on my lww, why?
I tried to change all the possible i/o schedulers in mesa kernel via set cpu, but the score always comes between 900 and 1000.
Questions:
1) is i/o scheduler even related with the i/o benchmark scores?
2) how can i/o performance increase?
3) why gaming on one v is way smoother than lww?
Do not depend on benchmark scores. The true rating of the phone will be with how it runs for the user.
Best Regards,
AJ
Wow did a antutu test on both of my phones and was expecting my L 9 to beat my escape. Boy was I suprised! My l 9 got a 10,568 ( pretty good compared to some). Then I ran the escape. My jaw dropped when I saw the amazing score of 15139. WHOA, it beat it in every catagory except RAM operation. My l 9 is running kumas rom with pure performance tweaks with cpu overclocked at 1.2 ghz. The escape is running sphinx rom with no cpu tweaks. Strange thing is every stats info I see on the escape says it has 1.2 ghz dc snapdragon. However im getting 1.5 ghz??? hmm Whats everyone else scores looking like for their L 9 ?
I get around 11k running mateo's CM 10.2 ROM.
Sent from my LG-P765 using Tapatalk
I have 9,616 running CyanogenMod 10.2 official (19/11/2013) :crying:
DarwinDeLaRosa said:
I have 9,616 running CyanogenMod 10.2 official (19/11/2013) :crying:
Click to expand...
Click to collapse
That's because custom rom overclock both CPU & GPU so it has higher score.
My L9 on CM 10.2 by mateo with stock cpu and gpu speeds i have 9843 pts
With CPU at 1,2 GHz i have 10671
With stock CPU and GPU at 384 MHz i have 9732 pts
With CPU at 1,2 GHz and GPU at 384 MHz i have 10706
With CPU at 1,35 GHz and GPU at 384 MHz i have 11502
lwg45714 said:
My l 9 is running kumas rom with pure performance tweaks with cpu overclocked at 1.2 ghz. The escape is running sphinx rom with no cpu tweaks. Strange thing is every stats info I see on the escape says it has 1.2 ghz dc snapdragon. However im getting 1.5 ghz??? hmm Whats everyone else scores looking like for their L 9 ?
Click to expand...
Click to collapse
Your LG Escape has got a better cpu and gpu and it would have shown better performance anyway. My L9 running stock showed maximum of 9600 with pure performance tweaks. But I don't think Antutu points can give you a good idea of how fast and smooth your phone is
madheado said:
Your LG Escape has got a better cpu and gpu and it would have shown better performance anyway. My L9 running stock showed maximum of 9600 with pure performance tweaks. But I don't think Antutu points can give you a good idea of how fast and smooth your phone is
Click to expand...
Click to collapse
Thats what I was getting at. Despite having a better cpu and gpu on the escape using both phones side by side in real world use the L9 performs better ALL around. The gpu was a big part of the escapes high score. However in side by side comparisons the L9S graphics seems to look better and out perform the escape. So I honestly believe all benchmark scores reflects what the phones software/hardware is CAPABLE of performing not really HOW it performs with daily use. This would explain why a score can as much as double by simply running a custom rom. I would never solely base my opinion of a phones performance on a benchmark score. Sure it can give you a idea of what its software/hardware is capable of. However you cant honestly get a true feel for the performance until you try the phone. So my escape murders my L9 in benchmark scores. BUT my L9 murders my escape in real life usable performance. Anyone else agree with this?
lwg45714 said:
BUT my L9 murders my escape in real life usable performance. Anyone else agree with this?
Click to expand...
Click to collapse
Yeah, I got your point! I guess, the only good thing about the Escape's cpu and gpu is that it's more battery-friendly, isn't it?
madheado said:
Yeah, I got your point! I guess, the only good thing about the Escape's cpu and gpu is that it's more battery-friendly, isn't it?
Click to expand...
Click to collapse
Yep, and both phones use the exact same battery. The escapes standby battery life is much longer than the L9 . During normal use the escape lasts longer as well by not to the degree of the standby difference.