Hi everyone!
Last night I was sitting with a friend of mine and tried to tell him that the Nexus 4 was faster than his Galaxy S3. He wanted to test it out and we both downloaded Quadrant Standard and ran the test...
Let's just say that I had to eat my words. I had a fresh install of CM10.1's latest nightly installed. The strange thing was that I was expecting the CPU to max out at 1500~MHz but it still stated 1100~ whilst the S3 was around 1400Mhz.
I went into the Performance menu and set it to PERFORMANCE but still the score was lower. My friend's S3 also had a custom ROM so was thinking that maybe the extra 1GB of RAM in the N4 may give a higher score.
What happened?
so because the s3 scores higher in quadrant, youre thinking that its a faster phone?
Quadrant scores literally mean nothing. You can't compare scores from different hardware on different versions of android running different rooms.
The s3 is a great all around performer that scores above average in all categories. The Nexus 4 scores higher in some categories (mostly hardware and software) but scores lower in others (camera, speaker)
thermal throttling
ZAnwar said:
Hi everyone!
Last night I was sitting with a friend of mine and tried to tell him that the Nexus 4 was faster than his Galaxy S3. He wanted to test it out and we both downloaded Quadrant Standard and ran the test...
Let's just say that I had to eat my words. I had a fresh install of CM10.1's latest nightly installed. The strange thing was that I was expecting the CPU to max out at 1500~MHz but it still stated 1100~ whilst the S3 was around 1400Mhz.
I went into the Performance menu and set it to PERFORMANCE but still the score was lower. My friend's S3 also had a custom ROM so was thinking that maybe the extra 1GB of RAM in the N4 may give a higher score.
What happened?
Click to expand...
Click to collapse
The Nexus 4 has got some heavy thermal throttling problems, search for this term on these forums and you'll find plenty of info about it. Basically, the Nexus 4 is the only phone that is unable to keep running at the stated max clock frequenzy for more than ~5-10 minutes, because it would overheat. After a few minutes, it will throttle the performance. This is also the reason it sucks so hard on benchmarks, compared to other phones with exactly the same hardware.
IINexusII said:
so because the s3 scores higher in quadrant, youre thinking that its a faster phone?
Click to expand...
Click to collapse
Not at all. I was just wondering what would cause the large discrepancy in the score.
jb91 said:
The Nexus 4 has got some heavy thermal throttling problems, search for this term on these forums and you'll find plenty of info about it. Basically, the Nexus 4 is the only phone that is unable to keep running at the stated max clock frequenzy for more than ~5-10 minutes, because it would overheat. After a few minutes, it will throttle the performance. This is also the reason it sucks so hard on benchmarks, compared to other phones with exactly the same hardware.
Click to expand...
Click to collapse
I read a little bit about the throttling but didn't think that it was that prevalent.
Thanks for the info!
http://www.anandtech.com/show/6440/google-nexus-4-review/3
check out the benchmarks here when anandtech put the N4 in a freezer. It tops every android phone on almost every test(and it still suffered from thermal throttling ). Any benchmark done for the browser will be slow bc chrome sucks for mobile phones. The AOSP browser will have higher results
jeff5891 said:
thermal throttling
Click to expand...
Click to collapse
It does NOT throttle in quadrant ... It just sucks on android 4.2
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
I watched it throughout the test ... It did NOT throttle and scored 4641
I noticed that it didn't use cores 3-4 in graphics tests however
Sent from my Nexus 4 using Tapatalk 2
italia0101 said:
It does NOT throttle in quadrant ... It just sucks on android 4.2
I watched it throughout the test ... It did NOT throttle and scored 4641
I noticed that it didn't use cores 3-4 in graphics tests however
Sent from my Nexus 4 using Tapatalk 2
Click to expand...
Click to collapse
It's not necessary to use CPU at full power for graphics tests. The GPU is what is put under stress.
How do you say
throughout the test ... It did NOT throttle and scored 4641
Click to expand...
Click to collapse
and then say
it didn't use cores 3-4 in graphics tests however
Click to expand...
Click to collapse
thats called throttling
jeff5891 said:
http://www.anandtech.com/show/6440/google-nexus-4-review/3
check out the benchmarks here when anandtech put the N4 in a freezer. It tops every android phone on almost every test(and it still suffered from thermal throttling ). Any benchmark done for the browser will be slow bc chrome sucks for mobile phones. The AOSP browser will have higher results
Click to expand...
Click to collapse
Holy ****. That's incredibly... worrying. Like someone added a .5 to reference volts or something. It really shouldn't get that hot.
But also incredibly nice cause the scores as so tasty. For 2 minutes at a time. Imagine what its doing to the battery as well!
Sent from my GT-I9300 using Tapatalk 2
jeff5891 said:
How do you say
and then say
thats called throttling
Click to expand...
Click to collapse
No it isn't .. It doesn't use all the cores all the time for everything..
even if I fire up GTA 3 , straight away it isn't using all the cores, and in fact doesn't use cores 3 and 4 at any time in GTA3 , it isnt throttling here,
The only benchmark iv seen the Nexus 4 throttle on is glbenchmark when ALL the tests are ticked, which is like a 15-20 min test at full load
Dsmboost was correct in the previous post as well , hence why not all the cores are being used when in games.
Can pretty much bet anything that its not throttling in 1 run of quadrant ...
Sent from my Nexus 4 using Tapatalk 2
italia0101 said:
No it isn't .. It doesn't use all the cores all the time for everything..
even if I fire up GTA 3 , straight away it isn't using all the cores, and in fact doesn't use cores 3 and 4 at any time in GTA3 , it isnt throttling here,
The only benchmark iv seen the Nexus 4 throttle on is glbenchmark when ALL the tests are ticked, which is like a 15-20 min test at full load
Dsmboost was correct in the previous post as well , hence why not all the cores are being used when in games.
Can pretty much bet anything that its not throttling in 1 run of quadrant ...
Sent from my Nexus 4 using Tapatalk 2
Click to expand...
Click to collapse
As further evidence, I see no difference in frame rate when limiting my gs2's cpu at 800mhz vs 1.2ghz when playing graphically intense games.
Related
I know that this tool isn't a good indicator of performance... but when I see other people with the EVO getting 1200+ and I am at 800-900, I am starting to wonder if there is a setting wrong. My phone occasionally feels sluggish. Any idea's what the cause could be? using an AOSP rom with Tiamat Kernel
- Just realized I wasn't in Q&A forums Sorry! - please delete if needed
Stuke00 said:
I know that this tool isn't a good indicator of performance... but when I see other people with the EVO getting 1200+ and I am at 800-900, I am starting to wonder if there is a setting wrong. My phone occasionally feels sluggish. Any idea's what the cause could be? using an AOSP rom with Tiamat Kernel
Click to expand...
Click to collapse
Good question!
Ive always wondered why some score so low on good combos.
I'm on CM nightly rc3 build 18 with savgzen 1.0.0 cfs-havs and scoring very well
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
new quad you will get low scores
imho quad and all the other benchmark test mean nothing
CheesyNutz said:
new quad you will get low scores
imho quad and all the other benchmark test mean nothing
Click to expand...
Click to collapse
I agree when comparing to other phones... but 2 phones of the same model should get similar scores right?
Stuke00 said:
I agree when comparing to other phones... but 2 phones of the same model should get similar scores right?
Click to expand...
Click to collapse
Have you ran it off of a fresh reboot?
I find my best scores were when I rebooted let it sit for 3-5 min after reboot, and then fire it up.
Mark_in_Tulsa said:
Have you ran it off of a fresh reboot?
I find my best scores were when I rebooted let it sit for 3-5 min after reboot, and then fire it up.
Click to expand...
Click to collapse
Trying it again after the phones been on a few minutes.
I used to use SETCPU and I thought maybe that screwed things up
If you want to increase scores why don't you put your Max frequency at 1190 and min at 1190 with setcpu. High scores are a thing of the past.
~ d3rk
dirkyd3rk said:
If you want to increase scores why don't you put your Max frequency at 1190 and min at 1190 with setcpu. High scores are a thing of the past.
~ d3rk
Click to expand...
Click to collapse
That is with my CPU at 1190. I am not trying to get a high score, but a score my phone should be able to get, you know? I see someone with an EVO get 1500 and then I am thinking something is wrong with my phone... I have the same phone..
Stuke00 said:
That is with my CPU at 1190. I am not trying to get a high score, but a score my phone should be able to get, you know? I see someone with an EVO get 1500 and then I am thinking something is wrong with my phone... I have the same phone..
Click to expand...
Click to collapse
Putting stock in scores is foolish though. 936 does seem a Tad bit low for sure. My set up is cm7 #36, tiamat kernel 3.3.5, JIT on, setcpu 1113 Max, 422 min smartass governor. I just ran a linpack and got 37mflops and 1200 in benchmark. This is the same set up I usually run on any rom or kernel. I try not to clock so high because it just drains battery. My evo actually can perform flawlessly clocked at even 998mhz. The only thing I notice different is that scores change and I don't really take a performance hit. Underclocking is the new overclocking lol. My last suggestion would be to play around with different kernels to achieve what you desire. Try savage zen!
Edit: just ran these 2 test just for you right now. Same results and I'm using same kernel you are.
~ d3rk
Run, it once, and then run it again right after w/out quitting. The if you run it multiple times (at least on my phone), all subsequent scores are on par with the second one which is usually 1-300 more than the first. Won't mean anything any different, but at least you might get some higher scores.
Also, you might need to try some different kernel combinations as well as different clock speeds. Mine usually does fine at 998 sometimes even at 921, but if I go too high it will start to slow things down. If you're not getting good results at 1190, try backing it off one setting at a time and see if you can get improved scores.
As far as comparing to other phones, even of the same make/model, you can't really do that. As long as they're in the same ball park, within a couple hundred, you're probably OK. Every phone reacts differently.
It also depends on the version of quadrant. You def can't compare diff phones running diff versions. I usually don't use it to compare to other phones, just diff setups on my phone. I know if I'm getting 1200 avg and then I all of the sudden flash a ROM and kernel and can't break 800 that something is probably wrong.
The most important thing is how the phone feels though. The fact that you say it feels sluggish leads me to believe you might need to try a diff kernel or see if something is eating up your CPU. Also, when you're over clocking, the min clock speed can effect your scores just as much as the max speed.
Ya give savage a go.
Kernel CPU setting:
Ondemand governor
Min 245
Max 998
My Device scores high runs smooth battery drain not a problem.
I just tested mine and got 1326. With these settings.. CM7 Nightly 36 with 2.6.37.6-Tiamat -v3.3.5 kernal.
Governor - Conservative
Min - 384
Max 1036
heres an article from androidcommunity.com http://androidcommunity.com/nexus-s...nce-benchmarks-with-trinity-and-cm7-20110530/
the way i see it, the real winner here is Trinity and our own morfic(@morfique on twitter) as Trinity for the g2x and for the nexus s is on top for both phones.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Are there other comparison benchmarks beside Quadrant? I take quadrant scores about as seriously as stories from AboveTopSecret
Also, overclocking to 1.5ghz.. What do the dual cores do when overclocked? Where are those comparisons?
Also.. Really.. Is a 1.5ghz overclock really an "every day use" type of configuration?
I'm not saying this isn't impressive but you're comparing a Mustang with a racing kit in it to it to a stock Camaro.
styckx said:
Are there other comparison benchmarks beside Quadrant? I take quadrant scores about as seriously as stories from AboveTopSecret
Also, overclocking to 1.5ghz.. What do the dual cores do when overclocked?
Also.. Really.. Is a 1.5ghz overclock really an "every day use" type of configuration?
Click to expand...
Click to collapse
heres a smartbench benchmark. and no, 1.5ghz isnt a daily. its more to play with or use when gaming. as you can imagine, battery life wouldnt be the greatest. i use the Trinity 1.1ghz UV as my daily. oh, and the top g2x(dual core) quadrant benchmark is 5587, also using trinity but for the g2x. btw, thats morfics g2x with the top g2x score.
Gotcha.. Don't get me wrong.. It is impressive and I'm all about getting the most out of hardware whether it's from a phone or a PC.
Damn. Now that is impressive. I wonder what the 3D score will be like once they o overclock it. It's going to be awesome!!!!
Most of it is coming from I/O due to ext4, quadrant benchmarks are far from accurate. Anything higher then 2000 (maybe less) is SF.
Dual core phones are definitely better, as well as their GPUs. I can run most of the tegra 2 games just fine tho.
look at that I/O score clearly nexus s it king
1.5 on single core is not faster than 1ghz dual core
Sgs2 ATM is best android out...period it doesn't need ext4 exploit to score 4700 on quadrant
Sent from my Nexus S using XDA App
lvnatic said:
Most of it is coming from I/O due to ext4, quadrant benchmarks are far from accurate. Anything higher then 2000 (maybe less) is SF.
Dual core phones are definitely better, as well as their GPUs. I can run most of the tegra 2 games just fine tho.
Click to expand...
Click to collapse
The Sgx540 beats the Geforce ulp gpu in the Tegra 2.
So there is no reason that it wouldn't work.
Sent from my Nexus S using XDA App
not saying its not impressive...your CPU OC is impressive that's about it
And only seen a handful of nexus S that can handle these speeds so to say nexus s is still king based on this is laughable im sorry thats my opinion
cant believe androidcommunity made an article on this
also removing the FPS cap on the kernel doesn't do anything for games the screen refresh rate cant handle such high FPS so you dont even see these extra frames when gaming all you're doing is killing your battery
tegra 2 dual core has a huge advantage on gaming.
Just becasue you gain a better score on a benchmark the tegra has an extra core to take some of workload in 3d rendering thats why games like Riptide run great on tegra and slow on our sgx540
im not bashing you simms i respect you and everything you do for the community just think you have a misunderstanding how things are working
demo23019 said:
im not bashing you simms i respect you and everything you do for the community just think you have a misunderstanding how things are working
Click to expand...
Click to collapse
i hope that you dont think that i think that a quadrant is an accurate snapshot of this situation, because i dont. quadrant is only meant to benchmark a single core, it even makes the field even between single and dual cores because of that. but, since the community likes to use quadrant as a "standard" for benching, ill throw these benchmarks up to look at. kinda like looking at a freak of nature, not being able to take your eyes off. anyways, since people like to use quadrants to compare android penises, the nexus s has the biggest(for now).
I get what you're saying and were you're coming from. I wont but in anymore enjoy you 1.5ghz wish my phone could run that speed
Sent from my Nexus S using XDA App
Quandrant scores are not a good way to judge device performance. This article explains it all:
http://www.androidcentral.com/how-cheat-android-benchmarks
As the article explains, the big give away is the length of the I/O part of the bar in the OP picture, which shows that the phone has manipulated quadrant benchmark into providing a higher score.
It is still pretty neat to see a +5000 quadrant score, even if it doesn't really mean much.
Like I said, anything higher then 1800-2000 is a fake score.
I can run riptide gt just fine with a 1.3ghz overclocked CPU btw
lvnatic said:
Like I said, anything higher then 1800-2000 is a fake score.
I can run riptide gt just fine with a 1.3ghz overclocked CPU btw
Click to expand...
Click to collapse
anything higher than 1800-2000 is a fake score??? then you havent ever been oc'ed to 1.3ghz.. never. btw, at stock speeds you should be getting 2300-2600. what phone are you using? my oc'ed original droid gets 1800-2100
lvnatic said:
Like I said, anything higher then 1800-2000 is a fake score.
I can run riptide gt just fine with a 1.3ghz overclocked CPU btw
Click to expand...
Click to collapse
Those scores are not faked although I do understand that quadrant is not the best barometer for benchmarking. On my Nexus S, stock, I get around 2100. With my G2x, stock, I get about 2700 (and that's mounted on the crappy ext3 filesystem). When using an ext4 kernel my score jumps to 3500.
demo23019 said:
look at that I/O score clearly nexus s it king
1.5 on single core is not faster than 1ghz dual core
Sgs2 ATM is best android out...period it doesn't need ext4 exploit to score 4700 on quadrant
Sent from my Nexus S using XDA App
Click to expand...
Click to collapse
A point of correction. The SGS2 out of the box is mounted as ext4, which is why it scores great on Quadrant. It's also clocked @ 1.2ghz. 4700 is a bit high for a stock SGS2, its typically around 3700 which is certainly nothing to sneeze out.
Hello guys ...
I have recently bought an S3 ( box packed ) with a Touch screen starter kit as well ... which came as a gift along with my phone .. there are few very handy things inside .. will post the pics of em soon ....
The set is from UK and is an international version ...
The confusing part is .. i am getting the score 3693 on quadrant .. which is u know almost 1000 less then the HTC One X ..
My phn has
android 4.0.4
Baseband I9300XXLG8
Build number I9300XXBLG8
Can any one help in this ??
Sent from my GT-I9300 using Tapatalk 2
I get 5264.
Not that I care for benchmark scores anyway.
Sent from my GT-I9300 using xda app-developers app
how many times did you try quadrant?
Please for god's sake stop using quadrant.
Does the phone feel good to you? If yes enjoy it, if no get a new one.
Check if power safe mode is enabled.
Running jb leak no oc, but had almost the same, I'd not even a bit better at ics tw roms.
But the scores don't matter anyway, if it works flawless, don't care about the score, quadrant isn't really reliable.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Gesendet von meinem GT-I9300 mit Tapatalk 2
Tectas said:
Check if power safe mode is enabled.
Running jb leak no oc, but had almost the same, I'd not even a bit better at ics tw roms.
But the scores don't matter anyway, if it works flawless, don't care about the score, quadrant isn't really reliable.
Gesendet von meinem GT-I9300 mit Tapatalk 2
Click to expand...
Click to collapse
Yes .. power saver mode was active ... i turned it off ... and wowwww ...
Now the scores are 3730 ... lmao
Sent from my GT-I9300 using Tapatalk 2
watsa said:
how many times did you try quadrant?
Please for god's sake stop using quadrant.
Does the phone feel good to you? If yes enjoy it, if no get a new one.
Click to expand...
Click to collapse
I tried it several times ... bro .. its quite embarassing when a friend of u sitting with u along with Htc One X and showing u quadrant scores ....
Sent from my GT-I9300 using Tapatalk 2
salmanghiyas said:
Yes .. power saver mode was active ... i turned it off ... and wowwww ...
Now the scores are 3730 ... lmao
Sent from my GT-I9300 using Tapatalk 2
Click to expand...
Click to collapse
HOw's the temp in the area you're testing? I've seen my score drop from 14k in Antutu to 7K if the temp of my device went above 35 degs. It might be due to throttling if the cpu is getting above certain temp. My scores on o/c 1800 are in my sig.
That's the reason why noone shouldn't trust quatrant scores, because some people thinks that higher scores means more powerful phone. There is many things affecting those scores, software, kernel settings, android version etc...
Nothing to do with phones real gaming performance.
I can't believe that after years of xda there are still a lot of people taking care of benchmarkig score.
Inviato dal mio GT-I9300 con Tapatalk 2
its like saying P4 3.6GHz is better than i7 1.8GHz
OP all i can say is that SGS3 benchmarks around 12000 12400 mark in AnTuTu .
Quadrant is known as a useless benchmark anyway .
jje
9992 on Antutu :/ ... thats not fair
And guys guyss guyssss ... just relax .. i know benchmark results doesnt mean ... but its just ... u knowww .. anyway ..
Sent from my GT-I9300 using Tapatalk 2
salmanghiyas said:
9992 on Antutu :/ ... thats not fair
And guys guyss guyssss ... just relax .. i know benchmark results doesnt mean ... but its just ... u knowww .. anyway ..
Sent from my GT-I9300 using Tapatalk 2
Click to expand...
Click to collapse
It's just...what??? Is there something wrong on phones real performance?
JJEgan said:
Quadrant is known as a useless benchmark anyway .
Click to expand...
Click to collapse
Every benchmark on a phone is useless. But maybe that's just me and my already big virtual penis...
Precisely. Benchmarks are useless I strongly agree. And this is coming from a One X user. The joke is that Quadrant shows the dual core Snapdragon S4 with a higher score than our Exynos/Tegra quad core, which is bull**** conaidering our real life quad core performance thrashes the **** out of the dual core S4. What matters most is that we know SIII and One X are the most powerful phones on the market. Period.
Sent from my HTC One X using xda premium
It's like benchmarking a gaming console: useless. Would you benchmark your Xbox/ps3 and think it's better than the other one, because it got higher points? But how come the other console still performs better on some games? Like I said: useless.
Hi,
@ salmanghiyas , stock/custom rom? Stock/custom kernel?
In case of custom kernel: mainly governor used? Other settings?
pähkinä said:
Every benchmark on a phone is useless.
Click to expand...
Click to collapse
Not exactly. Comparing benchmarks from different phones is useless. Comparing 5%-improvements-in-benchmark from different ROMs is useless. Comparing kernels using (only) 5% differences in benchmarks is useless. Claiming your phone is better because you have bigger scores is just idiotic. However, a very low benchmark score does point to something being wrong (thermal throttling happening too soon, governor problems, tasks using up the CPU when they shouldn't, etc.)
Imagine that your car has a problem with the axles. This causes many hard-to-notice problems, such as stability problems, some noises, increased heat, extra fuel usage, etc; none of those problems are easy to diagnose. However, if this axle failure makes your Ferrari unable to go faster than 200km/h, you will see clearly that there is something wrong. Even though you wouldn't go faster than 200km/h on your daily commute, this "benchmark" showed you that your car had a big problem.
Benchmarks are a tool. You need to understand how it works, why it works, and how to use it. Bragging about 50000+ points on Antudrant 3D Plus 2018 is ridiculous. Wondering why you get 5000 points when phones identical to yours get 20000 is a very valid question.
(incidentally, wondering why you get 18000 points when everybody gets 20000 is almost as ridiculous as bragging for getting 25000)
pähkinä said:
Every benchmark on a phone is useless. But maybe that's just me and my already big virtual penis...
Click to expand...
Click to collapse
Nope used as a test they are very useful as a bench mark .
Tells me a lot about LH1 LH4 and LH6 as to build completion .Likewise kernels scripts and other modifications can be bench tested .
Used to show who has a bigger score is useless .
jje
salmanghiyas said:
I tried it several times ... bro .. its quite embarassing when a friend of u sitting with u along with Htc One X and showing u quadrant scores ....
Sent from my GT-I9300 using Tapatalk 2
Click to expand...
Click to collapse
Tell him he's an idiot, ***** slap him and enjoy your damn phone rather than obsessing over benchmarks that mean NOTHING.
Sent from my iPad using Tapatalk HD
I'm running stock ROM, rooted and unlocked bootloader.
I ran antutu and benchmark pi on stock ROM+stock kernel and I got:
Antutu:18096
Benchmark pi:440
I flashed Franco r143 and ran antutu and benchmark pi and got:
Antutu:15526
Benchmark pi:557
I read everywhere that franco's kernel gave better performance so.....something wrong with my phone?
Sent from my Nexus 4 using xda premium
Nothing is wrong with your phone.
Stop doing benchmarks on your phone. Benchmarks are sh!t.
Sent from my Nexus 4 using Tapatalk 2
If you honestly judge performance by benchmarks then I don't know what to say.
But how else would I brag to my friends?
Sent from my Nexus 4 using xda premium
I'm not judging performance based on benchmarks
Just wondering why the scores decreased
Sent from my Nexus 4 using xda premium
I'd say the battery performance of Franco kernel comes with a price.
I just ran a test and got this on r143
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Sent from my Nexus 4
benchmarks dont prove anything
I've always found the stock kernel to be smoother. I only use Franco because I get double the battery life out of it.
Benchmarks are just a number. It means nothing when it comes to daily performance.
Franco Kernel is the same as stock kernel in terms of performance, but most likely faster. If its not faster then its something going on with your setup such as: ROM, bad apps, bad configuration when it comes to settings [did you try clearing app cache for franco kernel updater etc]
Why does gs4/htc-one get higher benchmark scores? Because they are faster.
OP you are probably getting lower benchmarks because the kernels governors are set frugally to save battery.
meangreenie said:
Why does gs4/htc-one get higher benchmark scores? Because they are faster.
OP you are probably getting lower benchmarks because the kernels governors are set frugally to save battery.
Click to expand...
Click to collapse
And yet I have yet to see my N4 lag whilst every video of the S4 I see shows lag at some points and every review I've read confirms random lag. Don't try and defend benchmarks, they mean absolutely nothing.
Properly done benchmarks means whatever it is designed to mean. For example, a super computer benchmarked at 1.4 teraflop makes the designer who developed it to run at 1.4 TF extremely happy.
An apps that spew out a number that is supposed to quantify your phone's quality is bull****.
However, I do agree that stock kernel is perceivably smoother than Franco kernel though
Sent from my Nexus 4 using xda app-developers app
KyraOfFire said:
Properly done benchmarks means whatever it is designed to mean. For example, a super computer benchmarked at 1.4 teraflop makes the designer who developed it to run at 1.4 TF extremely happy.
An apps that spew out a number that is supposed to quantify your phone's quality is bull****.
However, I do agree that stock kernel is perceivably smoother than Franco kernel though
Sent from my Nexus 4 using xda app-developers app
Click to expand...
Click to collapse
It's probably because stock kernel is using Ondemand as it's default Governor. Frequencies go on-line more often than Franco's hotplug settings on Interactive.
Nigeldg said:
And yet I have yet to see my N4 lag whilst every video of the S4 I see shows lag at some points and every review I've read confirms random lag. Don't try and defend benchmarks, they mean absolutely nothing.
Click to expand...
Click to collapse
No, the gs4 and htc-one really are faster as the benchmarks show... Software is a different matter.
Lag is generally caused by governors set to race to idle too aggressively on android. Some launchers cover this up better than others.
meangreenie said:
No, the gs4 and htc-one really are faster as the benchmarks show... Software is a different matter.
Lag is generally caused by governors set to race to idle too aggressively on android. Some launchers cover this up better than others.
Click to expand...
Click to collapse
Their hardware might be faster but their software is not which completely proves my point that benchmarks are not an accurate reflection of performance. The fact that results vary so much on the same device with different software shows that they aren't a way to measure hardware performance either.
Benchmarks were never meant to be an 'accurate reflection of performance'
Benchmarks are a very accurate reflection of the performance potential and hardware of said device.
Devices with better performance 'potential' and hardware are faster devices, period. Whatever hang ups software is causing has nothing to do with it, and benchmarks are not meaningless.. some people just get the meaning wrong.
Benchmarks are not meaningless, but there is no way to effectively measure benchmarks on a cellphone.
Sent from my Nexus 4 using Tapatalk 4 Beta
Bencharks are differnet cause stock kernel and custom kernel works differently..also..mpdecision are gone from custom kernels...they are enabling cpu's differently from stock kernel...if you wanna brag about it at your friends...simply show them your version number at 10th June..and give them your nexus 4 for 5-10minutes..without say nothing to them..
Whats your G3 Antutu Benchmark?
Mine is 26481 taken with airplane mode.
That's too low, isn't it?
It's really irrelevent. Open up a game and play it and how does it play?
Do you have any lag transitioning between apps?
Highest I got after a few tries is 30000.
Whats yours?
firecard said:
Highest I got after a few tries is 30000.
Whats yours?
Click to expand...
Click to collapse
I was going to start a new thread about it. I don't really care about benchmarks but always check them for point of reference between my various tests.
See my screenshot of latest Antutu scores.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Also attached GSM Arena scores as point of reference.
I've not boosted, photoshopped these scores, hell no development yet, but small tweaks here and there.
Sent from G3
firecard said:
Highest I got after a few tries is 30000.
Whats yours?
Click to expand...
Click to collapse
dont worry about it there low because of the screen ,longs as apps load fast and play well thats all what matters
I got a 0 when I used Antutu X
2gb/16gb UK version around 31000
Sent From My LG G3 Using Tapatalk
I now cannot get through the test without the phone saying "too hot shutting down"
Not rooted!!
Gesendet von meinem LG-D855 mit Tapatalk
i got 35813 on my last try
Rooted stock, no performance tweaks, just visual stuff with tweaksbox. Not outstanding, but I'm pretty pleased, in game i also didn't notice any lags yet (e.g. at asphalt 8), also haven't faced any throttling yet at longer periods of high demand.
Here are some additional stats of 3dmark
Gesendet von meinem LG-D855 mit Tapatalk
I got 36030. Pretty alright.
Stock rom, rooted with some LG bloat removed, default thermal settings retained (aka nothing modified) on ART I get 26K (the test, I believe runs better on Dalvik so I am assuming that is why my score was lower than 30K)
I decided to run Antutu X and it just gave me a score of 32K (on ART, which is usually about 5K lower than Dalvik)
After a few more runs.
Why so much big difference between the phones? I see scores from 26000 to 37000!! Why? I can understand what gives so much power in some phones.
ninjazx6 said:
Why so much big difference between the phones? I see scores from 26000 to 37000!! Why? I can understand what gives so much power in some phones.
Click to expand...
Click to collapse
it's another example of why these benchmarks arent the end all be all.
ninjazx6 said:
Why so much big difference between the phones? I see scores from 26000 to 37000!! Why? I can understand what gives so much power in some phones.
Click to expand...
Click to collapse
different people running phones differently, ART, Dalvik, certain ram draining apps all has an impact.
Generally Dalvik will give higher scores than ART as will having more 'free ram' (whilst giving you a higher score will not necessarily imrprove actual performance).
Temperature and battery level of the device when you run the test will also have an impact on the score.
Also I hope everyone is using Antutu X rather than Antutu - this is designed to stop cheating.
Lennyuk said:
different people running phones differently, ART, Dalvik, certain ram draining apps all has an impact.
Generally Dalvik will give higher scores than ART as will having more 'free ram' (whilst giving you a higher score will not necessarily imrprove actual performance).
Temperature and battery level of the device when you run the test will also have an impact on the score.
Also I hope everyone is using Antutu X rather than Antutu - this is designed to stop cheating.
Click to expand...
Click to collapse
Agreed. In my experience even changing a launcher can increase your score.
Sent from G3
This is my result
37k
Tarzan- said:
I was going to start a new thread about it. I don't really care about benchmarks but always check them for point of reference between my various tests.
Click to expand...
Click to collapse
Why start a new thread? It's then so hard to compare results, say, G3 vers Nexus 5 vers Note 4. 'Tis easier to post where all results are
theraker007 said:
It's really irrelevent. Open up a game and play it and how does it play?
Click to expand...
Click to collapse
Well, you're not exactly right, try to run this one benchmark and post here or there screen of your highest score. It will be much-much lower than that of XZ2, or HTC M8. And some people do play themselves or give their phones to kids for they can play. And low fps in games like in example. I'm not even mention GTA SA which is he11 of bad code itself and lags on almost everything but that's different story
ninjazx6 said:
Why so much big difference between the phones? I see scores from 26000 to 37000!! Why? I can understand what gives so much power in some phones.
Click to expand...
Click to collapse
Why? 'Tis so easy to understand, it's called throttling and discussed in this nearby topic