By 
Here at Phandroid, we never put much emphasis on benchmarks. While sometimes a fun way to measure general device performance, they don’t really give us an accurate measure of real world speed from a device. More than often benchmarks are simply a tool for fanboys to tout their device superiority over another — and we think that’s super lame, bros.
It wasn’t too long ago it was discovered that Samsung was inflating benchmarks scores for the Samsung Galaxy S4. Using a special benchmark boosting app found inside their custom software, Samsung was able to make the S4 fire on all cylinders. This caused the GPU to hit a frequency outside of what was available to the end user when using the device for normal things (like playing a video game). While not completely deceptive, it was a little misleading and we had little reason to believe this would change in the Samsung Galaxy Note 3. Looks like we were right.
When comparing scores on the LG G2 and Samsung Galaxy Note 3 — 2 devices using the same Snapdragon 800 SoC — ArsTechnica discovered the Note scoring abnormally high for a phone with the same processor. Lo and behold, Samsung was up to their same old tricks. After diving a little deeper, some line of code was discovered as the culprit and causing the Note 3 to operate in “high-power CPU mode” only when a benchmarking app is opened.
To get around this and measure performance that is actually offered to the average Note 3 users, Ars renamed Geekbench so that it would fly under the radar of Samsung’s benchmark boosting software. It’s ironic that even without the benchmark boosting, the Note 3 still outperformed the LG G2, (again, both devices equipped with the same SoC), so it appears there was little reason for Samsung to do this unless they’re prepping for the unforeseen Android competition in the future.
Where as Samsung’s official stance on the S4’s benchmark boosting app is that it was also available on a variety of Samsung-made system apps, it appears the Note 3’s is reserved exclusively for benchmarks. No gallery, no video player, no browser. Whether or not Samsung is being shady in providing users with the highest possible performance of their hardware — despite not being available in normal use — is up to you.
WOMP womp….
Its like you can’t trust anybody these days..shady practices hurt the market.
xperia Z Ultra with no boosting software
Badass phone!
Or people could just start renaming all their apps after benchmark apps and watch this thing really fly
Delete
My nexus 5 isn’t doing very well….
hmmmmmmm
is this real?
OK I gotta come clean, this is a optimus g pro. For some reason antutu labels it as a nexus 5. I’m not sure why.
That’s finny as hell xD
Funny too! :D
You running a ROM?
Yes carbon ROM 4.3 nightlies and wootevers custom kernel. It runs great and have had zero problems.
So… This is what I’ll get when I overclock. Got it.
More like unlock, flash a ROM/kernel and allow your phone to reach its full potential.
Yup,which is what we do anyway. I think it’s kinda cool. Shows you what you can achieve and why you should root, unlock, ROM, & kernel. Not to mention de-bloat if you decide to run stock. All good things. :)
Some people just want a device that they don’t have to mess around with. To work perfectly out-of-the-box, without root or any of that stuff. I’m definitely in that camp, even though I always end up ROMing all my phones after a few weeks. :P
Same story as before. There’s no inflation of the stats. This is simply the speed at which the SOC is rated. It’s also still an intent that other apps could call on besides benchmark apps. Once again the question you should be asking is why can’t the SOC achieve full power on a regular basis. Where is it being limited, and why?
Please tell me you aren’t that obtuse. When you run electronics at their maximum output, you are using a lot more power and straining your components more. This will greatly decrease your battery life, will make the device hotter, and will decrease the reliability of all the components in the phone.
The whole point of the benchmark is to compare phones. When you are comparing one phone running at max to another phone running at average everyday use, you aren’t comparing apples to apples (no pun intended).
For a company to intentionally attempt to deceive their customers says a lot about their integrity.
The complaint is specifically that without this additional app/intent the phone CAN’T run at full power rather than “average”, ever. If the CPU/GPU clock speed NEVER runs at the full rate as it is documented for that is of greater concern and question to me than an app/intent built in to actually make use of it.
I don’t buy 2GB of RAM for my PC and expect to only every be able to use 1.5GB. A 2TB hard drive and accept only 1.75TB of useable space. So why should I get an underclocked SOC?
Again if Samsung were cutting corners in the math operations, forcing the renderer into a lower resolution or OVERCLOCKING the SOC that would be a different story. I would consider that to be “cheating”.
I LOL’D
Not surprised…I now own both a GS4 and an Iphone 5… hoping to sell the GS4 for the Note…. because really I hated that gesture crap… it was definitely buggy…the iphone 5 runs games just as smoothly…Im sure they all exaggerate a bit.
I’m not fond of the gestures and that’s why I turned them off.
Apple and Motorola are the only two not trying to cheat consumers. Just to clarify it. At least those two are honest.
Motorola software has a lot of bugs and I never seen a quad core phone by them. I own a atrix HD and its very buggy and dies in a few hours. Its just a horrible experience from Motorola that won’t make me like there phones.
I don’t know what they have to gain by doing this. If they get caught (which they have already), anybody who actually cares about these benchmarks is going to know about it. It’s like trying to stick your hand back in the cookie jar after it’s already been swatted.
I go back and forth occasionally between android and Iphone…and I really feel that android still has that choppy…beta feel almost to it and did so even with the s4…the screen is beautiful ..I like the size…but many of the extra features just took up all of the space on the phone…16gb…had 8.5gb left right out of the box…SD is a cool idea but you cannot put many of the apps on it so what good is that? The Note 2 I was a big fan of…Definitely gonna check out the Note 3 when it comes out
Weird. I got between my G2, HTC One, and iPhone 5 and found that Android is every bit as smooth as iOS. Maybe it’s because of iOS 7, I don’t know. I also prefer the Holo look of Android and the way apps behave (action bar, side menu, etc) over iOS.
lol, funny thing is, it’s still the fastest without the boost.
I can’t believe Samsung is using the Note’s hardware to boost scores.
it’s about twice as fast as a s4 in all but memory without boost? nice…in theory could a rom be made to use those higher settings as it’s normal settings? since it can do it when it wants to and all, and just isn’t used on stock
(Shaking my finger) shame on you, Sammy, shame on you. Now give me a Note 3.
Lame! Why? Really that many virgins care Samsung about the benchmark score? Why?
This is coming from the haters of the Samsung devices. The Note 3 is without a doubt the most powerful Smartphone built to date. It is a beast no matter what way you flip it. Mine is on order and will be here very soon. Thank You Samsung.
Yes, the Note 3 is an excellent device, but that doesn’t make it okay for Samsung to manipulate benchmarks.
Obligatory car analogy: Would it be okay for a car manufacturer to state inflated numbers for acceleration and horsepower just because it’s an otherwise great car?
As a consumer, I want to know how long the battery is going to last as well as how fast the processor is. If Samsung says the battery will last X hours under normal to high usage than the CPU benchmarks should be run under those parameters. Not cool Samsung. I hope no one else follows their bad example.
instead of bashing samsung for using codes to utilize hardware performance at maximum capacity…why not ask other smartphone makers to do the same and compare numbers at full capacity
Fool. Why should they do that. People go to a lot of effort to write benchmark tests so unbiased comparisons of hardware can be made. Why should EVERY handset make then manipulate their OS just for the benchmarks. Perhaps if Samsung, et al just didn’t go to the effort of manipulating the results this would be better for all.
Kinda giving Samsung the benefit of the doubt, that the engineers assumed popular benchmark app tests should measure a device at its best always. But we also need to measure how a device performs when set conservatively or anywhere in between if adjusting.. I wouldn’t fault Samsung for leaving this as default behavior, they just need to provide the option to disable this, um, feature.
And the opposite… let users set individual apps to run at the highest performance, not just benchmark apps.
IDGAF.
i wish these geniuses realize that everyone last one of them(so-call cheats)and thats a fact ask that brian guy whatever his name is,he’ll talk to you. ty and a good day sir.
When did benchmark became a real world performance?
Benchmark is a “limit break”, it is a process of where you will try to see the limit of a certain thing. Like asking, what is the highest possible that this phone can do.. And obviously that is not a real world performance.
For example.. “Weight LIFTING” .. Here, we benchmark a human strength..its limit.. If a man can lift a 200KG weights then that does not mean a real world performance because you cannot carry that in everyday situation unless your adrenaline push you to.
TL’DR… BENCHMARK != REAL WORLD PERFORMANCE
Typical Fandroid apologist. Had Apple done it, you guys would be crucifying them.
Typical double-standard. Benchmarks now “suddenly” don’t matter to the likes of you. How convenient.
They haven’t done it because they cant. Simple as that.
No, like Google/Motorola, Apple has shown integrity in choosing not to game and cheat the benchmarks. There is nothing stopping them (outside of ethics) from putting in similar benchmark detection code.
It’s used to be able to compare between devices. It appears Samsung realize this and you don’t.
They have already done it before; I don’t see any reason they won’t be doing it again.
That’s what Samsung is like. They want to stay on the top, no matter what it takes even if they had to cheat.
It’s not only in smartphones; it happens in major sport events too.
They don’t want to stay on top….their ONLY aim is to make money. They only exist to grab money….that is why they operate in so many markets from phones to oil tankers.
This company lacks any real passion for technology….only for money.
tinyurl.com/l3cselt
v
Er I don’t know I mean this is not the kind of thing when I think of “inflating benchmarks”. Any PC user knows to set the performance settings to maximum before benchmarking their systems. The fact that Samsung disables power saving features is not what I would consider “cheating”. I think everyone should be running benchmarks at maximum settings in order to properly gauge performance. Cheating is when you would hard code in return values in drivers or implement algorithms in hardware specifically for synthetic tests that is not what is happening here.
Your an idiot.
Benchmarks are used to compare handsets. They are designed to run at the speed you would expect the device to run under normal operations so making it a good litmus test. If one company changes the results and others don’t or they manipulate how it runs then the results are meaningless and the benchmark becomes pointless to be used as a way of being able to compare between devices.
Great way to make friends on the internet by saying “Your an idiot”. I think you meant to say “You’re an idiot”, idiot.
So if Samsung had a low power mode as a settings option and a full speed mode which setting would be appropriate for a benchmark? Every PC benchmark done is always compared relative to the competition by placing the system in full speed mode. No one would benchmark a laptop unplugged for example. I guarantee you plug that phone into the wall and run the “stealth test” you will get the same high scores as the power saving features will be disabled. Samsung is simply trying to protect itself against companies that use different power saving technology. It’s not even necessarily a failure of the technology that it doesn’t automatically activate all cores when running a benchmark as it must strike a balance between performance and battery life. Other technology may work differently and enable all cores will running it benchmark and offer far worse power saving potential. Still other phones may deploy sneakier methods at recognizing when benchmarks are being run and not get caught. I’m VERY suspicious of Apples java benchmarks (sunspider) for example as they are much lower than the android scores.
The only way to compare devices in terms of performance is to run them all at full speed which any user can obtain by simply plugging the device into a wall. Otherwise devices with inferior batter saving tech will look relatively better than devices that are otherwise superior.
You miss the point. There’s no way to make the Note 3 run any app except a benchmarking app run at “full speed” (using your wording here), any other app runs in “low power mode” (again using your wording). So you’re only measuring what the device can do in a benchmark, not what it can do in real apps.
Your comments would make more sense if the user could choose which “mode” an app runs in.
Here’s what I’m saying:
For most everyday applications we don’t need the thing running at 100% CPU that is crazy since it would kill battery. So when we run a benchmark we are asking (or at least I am asking) “What is the fastest that this thing can go?”
Now if I had an application that required max performance I would plug the device into a wall. That would (or at least should) disable any power saving features and I could run said application at full tilt.
The article wasn’t clear as to whether it was bench marked plugged in or not, I assume it wasn’t plugged in because power saving feature seem to be on. If the benchmark was run with the device plugged in and there was a difference between the stealth and non-stealth varieties of the test then I conceded the argument i.e. they cheated. I just got my note 3 and I’ll try it out for myself.