Looking at CPU/GPU Benchmark Optimizations in Galaxy S 4

http://www.anandtech.com/show/7187/looking-at-cpugpu-benchmark-optimizations-galaxy-s-4

by MBCook • 13 years ago

216 78 13 years ago

rsynnott• 13 years ago

Samsung is merely being retro: http://blogs.msdn.com/b/oldnewthing/archive/2012/04/30/10298...

mokus3• 13 years ago

If the higher clock speeds are truly never used for anything but benchmarks then that's downright dishonest. But is that established to be the case? Perhaps those speeds are available for extremely heavy loads under specific circumstances (wired power, etc) that are rare but can actually occur. If they are, then this does seem like a fairly reasonable thing to do. Still mildly shady, but no more so than the norm when it comes to benchmark optimization.

alec• 13 years ago

The AnandTech link in the article says they found the string "BenchmarkBooster" in the apk alongside the class names of several well-known benchmarks.

http://anandtech.com/show/7187/looking-at-cpugpu-benchmark-o...

mokus3• 13 years ago

The fact that they found a code path that ensures particular programs get the full hardware speed does not mean that no other code path enables that performance level. I'm asking whether it is established that there is no other way that performance level is activated, as it does not seem clear to me from the article that that is the case.

corresation2• 13 years ago

The processor/GPUs do absolutely scale by load, however in this case they're scaling for specific apps, presumably because scaling that high for any other use (even where the throughput would be beneficial) would yield too hot of a device/kill battery.

It is incredibly dishonest, and there is simply no rationalizing it -- they are cheating, and everyone involved with this is kind of a scumbag.

megablast3• 13 years ago

You can certainly try to dig your head into the sand.

0x03• 13 years ago

Note to self: Make sure to use "com.antutu.ABenchMark" as the internal app name for my next android game.

JoshTriplett• 13 years ago

You probably wouldn't want to, even if you needed the extra performance. You'd get complaints from users about battery life, device heat, and similar things that people running benchmarks often don't check for.

stephengillie• 13 years ago

Maybe it could be an option in the settings?

[checkbox] Enable "High performance option"? Caution: will drain battery life and may cause device to overheat!!

cscheid• 13 years ago

It's Quack3 all over again! http://techreport.com/review/5226/further-nvidia-optimizatio...

bobbles• 13 years ago

Man, can't believe this was 12 years ago.. History repeats itself

MBCook2• 13 years ago

It's so sad to see this kind of stuff. It's not like the Galaxy S4 is a slow device. It reminds me of Nokia used a fake video [1] to sell the Lumia 920, which was supposed to have a pretty good camera. It just ends up being an unfortunate distraction that causes people to mistrust you.

Ars is right that this isn't new. I remember ATI being caught in the Quake/Quack thing [2] a dozen years ago.

[1] http://www.theverge.com/2012/9/5/3294545/nokias-pureview-ads...

[2] http://www.hardocp.com/article/2001/10/23/optimizing_or_chea...

ngoel36• 13 years ago

Even worse - when Microsoft ran Xbox One demos using high end NVIDIA PC's...

spartango• 13 years ago

At least those weren't benchmarks that were part of reviews, where people are making decisions based on 1-1 comparisons...

simonh• 13 years ago

That reminds me of when NT first came out. They had a demo where they'd show bar charts of CPU utilisation for multiple processors and show them ramping up and down exactly in sync as they ran more software on the box. Except the demo was a fake. It was actually showing CPU utilisation for one CPU on both bars. I actually saw someone from Microsoft show this at a trade show and I did a lot of work with Sun multicore boxes and it looked weird how the bars were exactly in step, like far too perfect to be a real world demo, but I figured they were just using a demo app that was rigged to use exactly the same cycles on both CPUs or something, not that the whole thing was a lie.

Boy that dates me, apart from the fact this was NT maybe 3.5 or so, what happened to PC trade shows?

Khaine• 13 years ago

Stay classy Samsung

enraged_camel2• 13 years ago

This is not the first time Samsung has been caught in a disgusting lie:

http://www.techspot.com/news/52274-samsung-admits-to-posting...

I don't understand why people keep buying this company's god damn products. I mean, holy shit.

czhiddy• 13 years ago

Why? Because they spend a fortune on advertising.

Anyone who's used a S3/S4 knows most of the S-"features" are pure gimmicks that a fairly competent engineer could implement in less than a week. HTC phones are much more sturdy and better designed, IMO.

I was vacationing in Hong Kong earlier this month, and the subway stations were just completely plastered with various 15-foot tall Galaxy posters. Some of the features they were promoting were just laughable - simultaneous photos with the front and back camera composed into a picture-in-picture result? Really?

acchow• 13 years ago

S-Voice. lol. It almost feels like they had a handwritten rules-based speech to text.

mrich• 13 years ago

While I agree that Samsung phones have quite a mishmash of features just for the heck of it, I think the picture in picture feature is quite nice and is something the competition (especially Apple) would heavily advertise had they come up with it themselves.

nmridul• 13 years ago

> simultaneous photos with the front and back camera composed into a picture-in-picture result?

I really wanted this feature for a long time and wondered why phones that had 2 cameras did not allow it.

It was specially useful when taking photos of my 1 year old daughter. It helps capture our emotions at the same time we were capturing her. So now you can see photos where she is trying to learn walking and at the same time see my face that is at times smiling (when she walks couple of steps) and at times scared (when she is about to fall) in the photos.

pachydermic• 13 years ago

Who cares about the gimmicks?

Samsung galaxy s phones have generally good hardware compared to the competition as well as a few really important hardware features which their competitors don't have (all of the features at once, that is): 4g lte support, a removable battery, and (for me the deal(maker||breaker)) an SD card slot. Those aren't just gimmicks. They're real differences. Many people like me really care.

I couldn't care less about the default software - there are a wide range of roms which address most (unfortunately not all for super picky people like myself) issues you might have.

ajross3• 13 years ago

Pretty much all major hardware vendors selling into a performance space get caught in this kind of game. This certainly isn't new behavior, though it's the first time I've seen a handset vendor try it.

But as far as why people keep buying the products: sadly it's because they're better. I'm looking at the GS4 (i9505, not the Exynos one) right now vs. the HTC One and Xperia Z. On paper, they're all basically the same phone (2G quad core snapdragons with 1920x1080 screens). But all the tiebreakers (really, all of them) go to the Samsung phone. The GS4 is slightly smaller, thinner, and lighter yet its screen is slightly larger. It's CPU is clocked higher and its battery is larger. It's cameras are better (much more so than the HTC, slightly more so than the Sony). It has a removable battery where the others don't. It has a sdcard slot where the HTC doesn't. It supports the "T-Mobile" HSPA+-on-AWS-band frequencies. It has a thermometer, hygrometer and barometer built-in.

It's just a better phone on everything quantifiable. Not so much so as to kill the competition, but to anyone paying attention and recommending hardware to their friends, they're doing an awful lot right.

xlayn• 13 years ago

You are right, and then it comes android over it, more of the same for them all, now for the average user that makes the business profitable, this means nothing... (so I'm with you)

so still why? advertizing .... average user would look to the numbers and buy it (as somehow people have started to realize are the same...)

mirsadm• 13 years ago

It is only a better phone if you print out the specs and compare them side by side. If you go by that method than the iPhone can also never win. But in reality the S4 is a horrible phone (in my opinion of course). I've got one here for development along side the Nexus 4 and HTC One. Out of the branded phones the HTC One is definitely the best. It does the things most people want out of a device very well. The S4 is a mishmash of poorly designed applications and features. Full of gimmicks and awful design decisions.

uchi• 13 years ago

HTC One's camera isn't bad at all. It was built that way for a purpose. You're doing the straight up Megapixel comparison between the phones.

http://www.anandtech.com/show/6747/htc-one-review/4

Every phone has its features. It's just at a point where a simple chart-by-chart comparison won't work for most people anymore. Practically every reviewer likes the One despite the fact that on paper it's a downgrade to the S4 in every way. Basically, it is like you said, the devil is in the details, but bigger is not necessarily better.

Personally, out of those three phones, I'd rather go with the HTC One only because I like the aluminum body chassis (along with how the screen looks) and the fact that their stock Android version is 50 dollars cheaper than Samsung's. The decently sized front speakers are also a plus. I very much dislike the look of the GS4 and TouchWiz which, in comparison with HTC's Sense, has less cooler whistles and doodads IMO. But ultimately I like stock Android the best.

The beats audio chip on the One is a con for me though (http://www.youtube.com/watch?v=Cdbn_pmxFic).

I haven't given the Xperia Z much of a chance, but aside from the fact that it looks aesthetically pleasing to me and I can use it while showering, I don't see much of an advantage. The GS4 came out with a water and dust resistant version, but it's slightly bigger and the audio jack is off so it won't fit most S4 cases. The Z's overlay is much more stock Android (GS4 version is TouchWiz only). So I'll pick it over Samsung because of that.

Basically, I really don't like TouchWiz. For the most part it just feels cluttered and half baked and a pain to use.

workbench3• 13 years ago

You're a marketers dream. So easily convinced by some "specs"

nivla• 13 years ago

I don't understand why is Samsung so hellbend on doing such things. Copying designs, faking reviews, faking benchmarks... Its not like they are at the bottom of the barrel, infact they are the best in mobile space right now. Are they getting too greedy?

megablast3• 13 years ago

A lot of this stuff, copying designs, they,did to get popular. They are, and always have been a dirty company, willing to do what it takes to win.

eonil2• 13 years ago

Copying and faking is the only thing Samsung can do. They're not good or creative, but also not stupid. They know what they can do and cannot, and where they are exceptional and sucks. So they're just keep doing what they're best - copying and faking.

Do you really think they can change suddenly? I don't think so. Copying and faking is internal system and philosophy of Samsung, can it won't be changed.

dangrossman• 13 years ago

@drivebyacct2: You're hellbanned.

justin66• 13 years ago

In the big scheme of corporate misbehavior, posting fake user reviews barely even registers. Which is depressing, but still.

steve19• 13 years ago

I don't understand why people keep buying this company's god damn products. I mean, holy shit.

Hyperbole much?

Because they make very decent phones that can easily be made to run alternative Android operating systems that allow me to do whatever I like with my phone.

For example, you can't plug a USB Software Defined Radio (SDR) module into an iPhone to monitor and decode spectrum.

antimagic• 13 years ago

Well you can but you would have to jailbreak the iPhone first, but then its the same for the S4 too, isn't it? (That was an honest question, I don't follow the state of Samsung phones too closely)

slacka• 13 years ago

disgusting lie

Is a little extreme. This is what companies do in the computer biz. Nvidia, ATI have been doing this BS for decades now. As cscheid said it's quack.exe all over again.

If they just gave the user the option to enable this performance tweak for games, I wouldn't mind. The problem is it's only enabled for benchmarks.

eonil2• 13 years ago

You talk like Samsung.

Others also do cannot justify anything.

solnyshok2• 13 years ago

I promised myself not to buy samsung phone after they forked me with Omnia HD i8910 - the flagship phone they made on a Symbian which they abandoned after couple of months, without a single bugfix. But then, I also promised not to buy anything from Motorola after they abandoned a flagship XT720 (xenon flash, full HD video) without bigfixes, withou upgrades (stuck on Android 2.1 forever with washed out photos, jerky 720p and half the RAM (all reviewers said it had 512 to be unlocked with upgrade) and locked bootloader. THE ISSUE - as much as I would like to buy Nexus 7 2013, it is not sold in my country by Goog, it is coming maybe in October, November, and xda forums are full of people who exchanged theirs 2-3 times to find unit without dead pixels, erratic touch, random reboots or light bleed. So, I may have no other option, but to buy tab from Sammy, because I can touch it in any shop, exchange or return, etc... EndOfRant.

diroussel4• 13 years ago

You could by an Apple device. That would almost certainly receive updates and fixes for several years.

m_darkTemplar• 13 years ago

Their marketing division might be shady, but they also make very solid devices. I've had my s3 for awhile and haven't felt the need to upgrade at all since it still performs super well.

As an Android developer I really appreciate what Samsung did in raising the bar for Android devices, and really bringing them to the mass market. High quality Android devices have been around for awhile, but it was far more common to see most people walking around with the low-mid tier Android phones. Samsung made it "cool" to own an S3, and now you can release and know that many of your users will actually own a solid Android device.

kenster07• 13 years ago

I don't condone lying in any sense, but I don't really care about the phone's benchmarks, and neither do most phone buyers.

In an era where patent trolling is rampant, an exaggerated benchmark of this nature is sadly, child's play. And it certainly would not have stopped me from buying the product if I was in market.

edderly• 13 years ago

On mobile devices it would seem sensible to also run your benchmark until the battery is "flat", it might be a crude heuristic but it probably the best you can do with product devices where you can't easily break out power rails and the like.

You should treat benchmarks with a heavy pinch of salt, simply because it's caught up in the bullshit messaging world of marketing where the aim is just to produce the bigger or smaller number where appropriate.

The necessity to satisfy marketing ignores a lot of things. Benchmarks often translate poorly to real use.

Does GLbenchmark 2.7 provide any useful insight into the performance of the top fifty games in the Android mobile market? Probably not with a few 3d heavy exceptions.

Some benchmarks will never die, you'll still see Dhrystone trotted out to this day. http://www.eejournal.com/archives/articles/20090602_coremark...

And then you hit a grey area where software can behave badly on one piece of hardware versus the other because of implementation. So if the benchmark hits this area, is it a hardware problem or a benchmark software problem?

fidz• 13 years ago

Reminds me about Nvidia with their fake benchmark several years ago

sczkid• 13 years ago

This is a minor point, but I'm really glad the article title uses "allegedly". Even if it's quite probable Samsung is guilty here, I'm glad the article doesn't adjudicate the way so many news articles do.

megablast3• 13 years ago

It comes off to me like these sites are scared of Samsung, they are a big player now, with a lot of pull.

tadfisher• 13 years ago

AnandTech is not one of those sites:

Today, we are large enough to avoid these petty discussions of withholding review samples. Most manufacturers know that one way or another we'll get our hands on a product for review and don't try to play these sorts of games. Rarely we are faced with a manufacturer or advertiser who is looking to influence our content. We have a firm internal policy in place to deliver honest, balanced reviews to the best of our ability - regardless of external pressures. Fortunately, as I mentioned earlier, we have been around long enough and are large enough to avoid this being an issue in the vast majority of situations.

http://www.anandtech.com/home/about

induscreep• 13 years ago

Anyone know if iPhone or any other phone does this kind of stuff? Downright disgusting.

songco• 13 years ago

Huawei's K3v2...

workbench3• 13 years ago

Since when has Apple relied on spec checklists to sell a product

corresation2• 13 years ago

Every iPhone release is a sea of specs: Number of GPU cores, pixel density, thinnest, lightest, made of raw holistic unobtanium, number of apps in the App Store, display metrics, etc.

Despite the truly ridiculous claims to the contrary, the iPhone has always been about specs, where in many cases it is the leader (such as having the perennial market leading GPUs)

Dwolb• 13 years ago

From an EE perspective, benchmark tests should include power consumption and thermal profile measured by third party, calibrated equipment.

There are some errors induced through the act of measuring, but power consumption and thermal profile checks would cut down on the amount of CPU/GPU benchmark tampering.

DominikR2• 13 years ago

Samsung has responded that the higher GPU speeds are also available certain other applications that run in full screen. (browser, camera app, video app - which are all apps that are used frequently)

If that is the case, then I fail to see anything immoral. Bad engineering - yes - but there's nothing immoral in arbitrarily boosting the GPU to run the browser or the camera smoother. (and once you regularly do this for some standard apps, it would also be appropriate to run benchmarks with the fastest speed)

Link: http://www.theverge.com/2013/7/31/4574210/samsung-benchmark-...

Edit: I also don't understand why anyone would investigate with a hex editor. (like Anandtech does)

Samsung is quite open with its Android devices, and regularly releases the source code (Link: http://opensource.samsung.com/) It also would be quite strange to cheat, and then release the source code so everyone can see.

diroussel4• 13 years ago

I also don't understand why anyone would investigate with a hex editor. (like Anandtech does)

The binaries were on the device, but the source code was not, so the binaries are a natural place to begin the investigation.

Is it ever clear which source from http://opensource.samsung.com/ is on the device Brian was investigating?

DominikR2• 13 years ago

It is absolutely clear which source is on the device because you can look up Kernel version, Baseband version and the Build number in Settings->About Phone.

Anyways, what I would have liked to see in this article is if the Intent they are referring to (the one that triggers the increase of the GPU clock) is only used for cheating on benchmarks, or actually a feature that is used on many apps - not just speculation/baseless accusation.

simula67• 13 years ago

I wonder the extend to which we can trust performance benchmark results about other closed source software like web browsers, operating systems, servers etc.

Looking at CPU/GPU Benchmark Optimizations in Galaxy S 4

Related Stories