Samsung allegedly boosting benchmark performance

Anandtech digs deep on a forum allegation and finds depressing results.

Read the whole story

 

[ScifiGeek]

Musafir_86[/url]":1si1uo5s]-Hmm, so how about what Intel's (allegedly) doing? http://www.eetimes.com/author.asp?secti ... id=1318894 (via http://legitreviews.com/news/15752/).

Allegedly is a more reasonable word in the Intel case.

It looks a lot more black and white in the Samsung case.

 

[Adam Starkey]

truepusk[/url]":3ul2dcfl]
Best Buy probably isn't the broader market anymore. I don't think many analysts think it will be long before they go the way of Circuit City and those who don't are probably betting on Best Buy to be able to make major changes to adapt.

No disagreement there. I was using BestBuy purely as a stand-in for generic gizmo shopping. The experience would be broadly the same at a Verizon store.

If/when BestBuy keels over, their customers will simply move to another store where they will continue to be only loosely informed about the devices they are purchasing, and only marginally more interested. The majority of the population really doesn't know or care much about their phones and tablets. They know they don't want to buy the wrong thing, and are only too happy to buy what everyone else is buying because there's safety in numbers. Of the small chunk of the market that is engaged with their purchases, a decent proportion of that group is really running on received wisdom. They don't know or understand half as much as they think they do, and their gut is doing as much thinking as their head. The remaining subset of the market makes what could be properly regarded as an informed choice.

As someone mentioned above, there is power in influencing the latter two groups because they (we) can have some effect on the larger market, but not half as much as a decent price (preferably one that appears to be a discount), and a recognized brand name.

This is of course as true of TVs, washing machines, cars, and power tools, as it is of phones, tablets, or laptops. None of us have enough time to *really* know what we're doing when we make half the purchases we make, and only some of us are interested enough in any of those things to make the time.

I doubt this will hurt Samsung any in the short term. In the long term it may begin to coalesce into a broader narrative but that'll take a while, and in the interim Samsung is making money hand over fist.

 

[Adam Starkey]

therealankit[/url]":2v5kujsw]
Also, if some benchmarking applications make the device clock higher, isn't it possible if developers make applications that makes the device faster all the time, to get the best usage. That will be good for the users.

No it wont. Samsung aren't throttling the performance of their hardware for the lulz. Their testing obviously led them to set an upper safe/functional bound on their devices. Lots of apps a) messing around masquerading as other apps, and b) forcing a device to run out of specification for protracted periods of time would be all round bad.

At best it's a monumental waste of developer time that only impacts some Android devices, at a cost of significant battery life. At worst, it's a whole lot of users dealing with crashes, reboots, and shorted device life caused by overheating.

 

[nomorehw]

Now you guys know how Koreans do business, everyone who has worked with Koreans in the industry knows, it's always "whatever it takes"

 

[f0xik]

Firehawke[/url]":3u16mcz5]ATi and nVidia were both caught in the past doing this with video card benchmarks, too. I particularly remember one incident involving Quake 3 benchmark optimizations about twelve years ago where ATi was doing some specific optimizations that would only trigger when it detected the calling application was "quake3.exe"

Made it really easy to prove, too. Just rename the EXE and you'd see performance drop considerably.

nVidia lists in their driver release notes for which games they improved the performance. You can also edit the application profiles through their control panel. I believe that's transparent enough.

I also think that nVidia cares about the user by testing various games and fixing the driver or tuning the profiles, while most of the game developers don't care about that.

 

[Adam Starkey]

f0xik[/url]"04w5gcw]
nVidia lists in their driver release notes for which games they improved the performance. You can also edit the application profiles through their control panel. I believe that's transparent enough.

Tyler X. Durden already explained why what nVidia/ATI were doing was not just tweaking performance for popular games.

 

[Mydrrin]

Samsung...aren't these the guys that pay people to talk badly about other hardware. Anyone else surprised by even more despicable behaviour?

Morals? Stealing? Fraud? It's part of the business method.

 

[jb_ddg]

truepusk[/url]":kncb9fex]

ShlomoAbraham[/url]":kncb9fex]Any theories as to why? If the chip is there, why not let other apps use it? Save battery life?

Yes. Battery life.

Not exactly. If it were just battery life, you'd get full performance when you were on AC, and you don't. The other culprit is heat; which the phone is probably capable of generating enough to damage itself. Benchmarks typically don't run very long, so it's a calculated risk.

 

[Fr0sty514]

Agree that it's false when (as was mentioned here) the benchmark is elevated in a mode that normal usage never could, such as overclocking the processor especially for the benchmark but not having that clockspeed set by the governor for anything else.

I think benchmark cheating isn't the problem - it's the symptom. Benchmarks just aren't real world measurements. Technically you could even make the argument that if the hardware is capable of running the benchmark at a certain level, then the system as a whole IS capable of that performance, even if it is specific to that benchmark. I mean, the benchmark can't force the hardware to do something it isn't physically possible for it to do.

As for video card drivers, again I have no problem with that, bearing in mind that benchmark measurements are almost meaningless. Graphics card manufacturers tweaking their drivers for optimum performance in certain apps - I don't see anything wrong with it. If they can squeeze out the extra FPS without losing visual quality, based on a bit more in-depth knowledge of what the application does with the graphics card, why not? It's kind of like saying that optimizing compilers are 'cheating' with respect to CPU performance.

 

[matthewslyman]

Would this be grounds to retroactively reduce all of Samsung's recent history of benchmarking results by a factor of 533/480?

 

[Grimmash]

deathBOB[/url]":1zcv48i0]Boy I sure don't care about mobile benchmarks.

To play devil's advocate, mobile computing is still an area where performance is still something that could be a factor, depending on application. For desktops, most people can afford a computer that far exceeds the demands of actually critical software. While Ars might be a hotbed of people tweaking to get Crysis or some similar application running at 120 FPS or something, mobile devices can still fail to provide the proper sweet spot of speed and battery life. I know the Macbook Air I am typing this on is woefully unable to do many things my desktop, a full two years older, can just shrug off.

Cheating mobile benchmarking could have actual consequences when mobile phones can cost as much as a video card with contract, and a video card + ram + something else without contract.

 

[gdinero]

Tyler X. Durden[/url]":2ewlpn0n]

F22Rapture[/url]":2ewlpn0n]

Firehawke[/url]":2ewlpn0n]ATi and nVidia were both caught in the past doing this with video card benchmarks, too. I particularly remember one incident involving Quake 3 benchmark optimizations about twelve years ago where ATi was doing some specific optimizations that would only trigger when it detected the calling application was "quake3.exe"

Made it really easy to prove, too. Just rename the EXE and you'd see performance drop considerably.

Was that specific to the benchmark itself though? Making game-specific optimizations isn't really cheating. Cheating would be putting in extra effort to make the benchmark run better than the game does (optimizing a JS engine for Sunspider), or unlocking extra hardware functions not available to anything else (as Samsung seems to have done), or some other hack which misleads the benchmark (such as ATi reducing default image quality and accuracy to get a higher FPS).

If it is the incident I'm think of it I'm pretty sure it involved pre-loading and maintaining in cache certain textures that showed up in a demo sequence in those games that was commonly used for benchmarking (because it was a consistent use and exercising of the game engine that could be duplicated over and over).

So the optimizations would really only work for that particular map and to a certain extent that path.

http://techreport.com/review/3089/how-a ... -quake-iii

The gist is that if the drivers detected quake3.exe running, they would fudge with texture quality to get better framerates.

 

[Dadlyedly]

sd4f[/url]":1kdas9dl]This in a way reinforces my gripe with the mobile market. Companies aren't as interested in making excellent consumer products as they are making sales.

I suppose it's a chicken and egg thing except making a good phone which will build good will is a lot slower than attacking the spec sheet and getting immediate sales.

I think samsung have demonstrated that they prefer the latter; win the contest on tech specs, make the sales and abandon the phone when the next one nears release.

This is what we call "modern capitalism." Ever since the theory that CEOs and Boards are only obliged to keep stock prices high for stockholders, this is what happens. Those that realize they have obligations to their customers, employees, and the company itself do better in the long run, but who pays attention to anything beyond the end of the quarter anyway?

 

[Cerberus™]

I'm still interested to hear an explanation of why Samsung doesn't allow this high 533 MHz for all demanding applications, instead of only for these benchmarks. If the phone is technically capable of running at 533 MHz occasionally, how does it help Samsung to throttle it? There has to be an economic theory that can truly explain this.

I don't think it can be the battery life, because it would only be triggered when a game demands it, just like any regular CPU governor. Or am I mistaken?

Is it that using this high frequency for a prolonged time overheats the device or makes it unstable? Then they can just put a time limit on it, throttle it automatically when it reaches a certain temperature. Again, I believe this is how standard CPY governors work, isn't it?

The only situation where it would make sense to limit this frequency to benchmarks is if a short period of 533 MHz were somehow a *scarce resource* that is "used up". How is that possible? The only thing I can think of is that *any* period of 533 MHz slightly damages the CPU or has a slight chance to damage it, so that the total time the CPU can run at that speed is very limited; if, say, they have calculated that 3 hours at 533 MHz reduces the average life time of the CPU by 20 %, or something, you have a reason to limit it not just within a short period, but over the entire lifetime of the phone. If you let people use 533 MHz at will, they kill their phone; if they only use it a couple of times for benchmarks, there is little damage, and you profit from the misrepresentation of what the phone is actually capable of.

Does that make sense at all?

 

[Walt French]

lowlymarine[/url]":9skrmgrc]This is, regrettably, the way of the industry now.

You must buy from a different industry than I do.

Benchmarking “realistic” workloads is useful. Showing results that are possible, and not mentioning how they'd affect claimed battery life, or even whether they could affect the thermal safety/uptime of the device, is an obvious attempt at deception.

Most of the firms I buy from “put their best food forward.” When they jigger their product to make it appear better than it really will be, I let them sell their stuff to people who are happy to deal with thieves and liars.

If you say, “regrettably,” then it's incumbent on you—helpful TO you—to help fix the problem. Boycott Sammy, demand retractions/denunciations from less tech-savvy shops like TheVerge who swallowed the Samsung jizz naïvely, whatever works for you. Otherwise, you're just saying, “OK, I'm bent over. Now what?”

 

[Walt French]

Cerberus™[/url]":3etrtvz4]I'm still interested to hear an explanation of why Samsung doesn't allow this high 533 MHz for all demanding applications, instead of only for these benchmarks. If the phone is technically capable of running at 533 MHz occasionally, how does it help Samsung to throttle it? There has to be an economic theory that can truly explain this.

I don't think it can be the battery life, because it would only be triggered when a game demands it, just like any regular CPU governor. Or am I mistaken?

Is it that using this high frequency for a prolonged time overheats the device or makes it unstable? Then they can just put a time limit on it, throttle it automatically when it reaches a certain temperature. Again, I believe this is how standard CPY governors work, isn't it?

The only situation where it would make sense to limit this frequency to benchmarks is if a short period of 533 MHz were somehow a *scarce resource* that is "used up". How is that possible? The only thing I can think of is that *any* period of 533 MHz slightly damages the CPU or has a slight chance to damage it, so that the total time the CPU can run at that speed is very limited; if, say, they have calculated that 3 hours at 533 MHz reduces the average life time of the CPU by 20 %, or something, you have a reason to limit it not just within a short period, but over the entire lifetime of the phone. If you let people use 533 MHz at will, they kill their phone; if they only use it a couple of times for benchmarks, there is little damage, and you profit from the misrepresentation of what the phone is actually capable of.

Does that make sense at all?

Game developers, not knowing the fine details of the particular version of the hardware (this model has two of the 11K Android options in the wild), will mostly go for the highest performance possible. I suspect the GPU cranks pretty hard on ordinary videos, and again, developers won't necessarily be able to fine-tune the higher voltage / power budget for a video — especially if some other task, say blinky ads, are ALSO going on.

My crude SWAG is that running the GPU at full tilt might (a) increase overall power for videos by 10%–20%, thereby reducing battery life an hour or two, and possibly (b) raise the temperature to overheat conditions that'd force-shutdown or damage the device, especially when used in warmer environments, eg, in the summertime sunlight. Neither of those are acceptable for your average user.

OTOH, overclockers might be delighted by this info; they can brag how their little gizmo gets a few extra fps, and be an expert for their friends—showing them how to get 10% better outcomes with a bit of wizardry. Samsung could turn this into a bonanza by claiming this capability was only meant for power users with extra-big balls, The Ones Who Care®. The capability is an Easter Egg.

 

[Walt French]

matthewslyman[/url]":3iwx8tcy]Would this be grounds to retroactively reduce all of Samsung's recent history of benchmarking results by a factor of 533/480?

Following on their astroturf campaign, I think Samsung has now earned a “guilty until proven innocent” approach that your idea suggests, but I'd rather have Samsung's Director of Engineering describe why this arrangement was deemed in the customers' best interest.

We might ask some of the publications who like to use benchmarks so extensively, what assertions they'll demand from device makers before they're reviewed. With any luck a bit of shaming will nip the problem in the bud.

[edit: rewrote the “guilty…” part to clarify my intent]

 

[Walt French]

Dadlyedly[/url]":36rwr8sh]

sd4f[/url]":36rwr8sh]This in a way reinforces my gripe with the mobile market. Companies aren't as interested in making excellent consumer products as they are making sales.

I suppose it's a chicken and egg thing except making a good phone which will build good will is a lot slower than attacking the spec sheet and getting immediate sales.

I think samsung have demonstrated that they prefer the latter; win the contest on tech specs, make the sales and abandon the phone when the next one nears release.

This is what we call "modern capitalism." Ever since the theory that CEOs and Boards are only obliged to keep stock prices high for stockholders, this is what happens. Those that realize they have obligations to their customers, employees, and the company itself do better in the long run, but who pays attention to anything beyond the end of the quarter anyway?

Huh, “they all do it,” eh?

I have the pleasure of working in a regulated industry, where dishonest claims — simple deception, not even as bad as what Sammy's up to — often enough results in big fines and occasionally people barred from working in the industry for life.

I'm curious what business the customers are either so stupid that they expect to be cheated, or they're so naïve that they don't realize how scammers get all their money, so there's no benefit to the honest business from making accurate/honest claims. Claims such as how long the device will run when doing “typical” things like video/gaming at the rates implied by benchmark results that were consciously and conspicuously chosen, since battery life and responsiveness are probably the number one and number two criteria that many consumers use in evaluating a smartphone.

 

[TheDissolver]

the benchmark means fuck-all to user experience. so I give precisely one tenth of a shit, because it is a douchey thing to do.
if you're worried about 'cheating' overclocks... do what overclockers do. run burn-in tests and see how long your review samples will last running benchmark loops.

the relevant numbers (how long it plays netflix videos, how quickly it loads actual pages) in nominally 'controlled' environments are still valid. the cpu benchmarks do convince true-believers and fanboys, I'm sure... but for all intents and purposes, they're getting an accurate number: you want to see numbers for raw performance of the SOC? we can give you that. you want to see how fast it is in actual daily usage, where battery life and heat are important? You can test that, too.

it's not like GM producing a Camaro with a V-8 for the horsepower tests and a V-4 for the mileage tests. It *would* be like electronically removing the rev limiter anytime you were in proximity of a dyno or a drag strip--places where fake number seem important in ways that have little or no significance to daily performance and utility of the car.

 

[sunshinerag]

Not surprising at all.

Looking at the name of the APK - TwDVFSApp.apk.... it sounds like touchwiz is a cover to sneak in such stuff over base android.

Sorry for all the S4 buyers who meticulously looked at benchmarks before their purchase.

 

[Marlor_AU]

This isn't anything new. LG were caught doing the same with their fridges in Australia.

In Australia, all whitegoods must have a big sticker on them with an energy efficiency rating.

LG took advantage of this. Their fridges could somehow detect when they were in the energy efficiency testing lab, and throttle their cooling to achieve impossibly impressive results. Suddenly, LG had the most efficient fridges on the market, and the prominent government-mandated labels emphasised this fact.

Until you got them home. Then they actually turned on properly, and had low-to-middling energy efficiency.

This didn't last for long. People found out about the strategy, and LG were ordered to compensate customers for their extra power bills.

 

[Otus]

Firehawke[/url]":3nutmr35]ATi and nVidia were both caught in the past doing this with video card benchmarks, too. I particularly remember one incident involving Quake 3 benchmark optimizations about twelve years ago where ATi was doing some specific optimizations that would only trigger when it detected the calling application was "quake3.exe"

Made it really easy to prove, too. Just rename the EXE and you'd see performance drop considerably.

That's a somewhat legitimate optimization, since quake3 was something you might have bought the product for. These days both companies have extensive game profiles that optimize how the complex GPU drivers behave in particular games. Especially multi-GPU performance depends a lot on that.

Optimizing for e.g. WoW benefits those who play it. Something like Sunspider, OTOH, will not benefit anyone but their marketing.

 

[madmilk]

I suspect that the GPUs are not actually binned/guaranteed to be stable at 533MHz, especially at high temperatures. It'd be a highly annoying if the phones crashed occasionally during day to day use. I doubt anyone runs benchmarks over and over without some resting time in between, so Samsung can fudge the clock speeds here a bit. That, or they're using high voltages to ensure stability, which would definitely affect lifetime.

 

[Batmanuel]

Is this tweak valid on the US models? Those use a quad-core Qualcomm chip with a totally different GPU. Has anyone checked them for similar code?

 

[hobgoblin]

Musafir_86[/url]":at6v6vko]-Hmm, so how about what Intel's (allegedly) doing? http://www.eetimes.com/author.asp?secti ... id=1318894 (via http://legitreviews.com/news/15752/).

Regards.

That could be claimed to be a smart compiler. But then the Intel compiler have in the past produced code that would drop back to 386 instructions if the CPU didn't id itself as "Genuine Intel"...

At this point, the only compiler i would trust to produce impartial binary is GCC.

 

[koolraap]

I don't see why people saying "battery life" are being down voted -- there isn't a one-word answer that sums everything up (even the pithy "marketing" may not be completely correct).

By ramping up the clock in benchmarks Samsung can claim with a reasonably straight face the phone is:
1. fast
2. has decent battery life.

You buy the phone, you run the benchmark! Woo! My phone is fast. And it lasts about how long it says on the box. Woot.

As for who's responsible? It could be the marketing department. It could just be Samsung's Programmer of Legend(tm) who squeezed 10% more out of the phone during one busy coding frenzy. "See boss? It's fast. And it's got the battery life it says on the box."

Back to downvoting: Ars provides a system to punish commentators who have an unpopular opinion with humiliation, and rewards people who don't like people who think differently to themselves by allowing them to anonymously kick those people with public humiliation. It's great.

 

[SedsAtArs]

Grimmash[/url]":3jy8ll70]

deathBOB[/url]":3jy8ll70]Boy I sure don't care about mobile benchmarks.

To play devil's advocate, mobile computing is still an area where performance is still something that could be a factor, depending on application. For desktops, most people can afford a computer that far exceeds the demands of actually critical software. While Ars might be a hotbed of people tweaking to get Crysis or some similar application running at 120 FPS or something, mobile devices can still fail to provide the proper sweet spot of speed and battery life. I know the Macbook Air I am typing this on is woefully unable to do many things my desktop, a full two years older, can just shrug off.

Cheating mobile benchmarking could have actual consequences when mobile phones can cost as much as a video card with contract, and a video card + ram + something else without contract.

While I agree what he posted won't be true for all users, and that it's a shady practice by Samsung in any case, it could still very well be true for the poster you quoted.

I know that myself, I mostly use very simple apps, which don't tax my phone much. Google Maps and Ingress are probably the two slowest running ones, and I don't feel I can fully blame the slowness of Ingress on my phone.
The day that docking a phone in and using it as one does a laptop for work now comes (assuming it does), this might well change for me.

 

[MrMickS]

Perhaps its time that tech web sites, such as Ars, stopped using benchmarks in their phone reviews. They could include a short paragraph to say why.

I don't care about the benchmark performance of my phone. I do care how responsive the UI is, how quickly it can render web pages, how it performs in a low signal area. What happens when it can't get that WiFi or LTE connection. Instead of reviewing a smartphone as a computer, come up with a series of tests that test its ability as a smartphone.

 

[Zarsus]

hobgoblin[/url]":bzy3so3x]

Musafir_86[/url]":bzy3so3x]-Hmm, so how about what Intel's (allegedly) doing? http://www.eetimes.com/author.asp?secti ... id=1318894 (via http://legitreviews.com/news/15752/).

Regards.

That could be claimed to be a smart compiler. But then the Intel compiler have in the past produced code that would drop back to 386 instructions if the CPU didn't id itself as "Genuine Intel"...

At this point, the only compiler i would trust to produce impartial binary is GCC.

Intel's compiler would degrade performance for non-Intel CPUs, but this doesn't make the compiler bad, it just makes the compiler bad for other CPUs, it would certainly produce valid output. Intel doesn't cheat on this regard, it's a testament to their compiler that it's able to identify useless benchmarking code that ultimately doesn't produce any visible results and removing or re-architect them. For an example you can check out the comment of this StackOverflow question:

Intel Compiler 11 does something miraculous. It interchanges the two loops, thereby hoisting the unpredictable branch to the outer loop. So not only is it immune the mispredictions, it is also twice as fast as whatever VC++ and GCC can generate! In other words, ICC took advantage of the test-loop to defeat the benchmark...

 

[pitmonster]

Adam Starkey[/url]":3ifmz8jk]
I totally agree. The auto industry sending out ringers to magazines to test is pretty much the accepted norm. It's wrong, and if no-one calls them out on it, then there's no incentive for anyone to play fair.

At least one journalist did call Ferrari out for this : he is now banned from testing their cars.

http://jalopnik.com/5760248/how-ferrari-spins

 

[ikhtiar]

We all make decisions by evaluating trade offs. When you buy a car, a bigger engine (more CC) will run the car faster but will also consume more fuel. We all want a bigger faster car which consumes less fuel. So if any car maker can claim that, you will buy that. To do that, Car makers will have to tweak the performance for the demos but have to revert back to normal as in real life to deliver the promised fuel consumption.

For the Handsets, it is even easier. You will look to benchmark reports directly if you are a geek or build favorable perception by reading positive reviews as tech media uses the benchmark reports to decide if the handset is fast or not. In reality, the handset is not as fast as it should be but it delivered the battery life as promosied.

Samsung is guilty of deception but we consumers also to be blamed. Why we blindly rely on others to tell which handset is better? Just go to a shop and try the handsets for 30 seconds. That should give you very good idea. Tech media should rely on actual usage and only then recommend a product. Just passing judgment based on Benchmark reports is not journalism, it is like re-printing of press release.

 

[Lwio]

Disgraceful actions by Samsung, it's lucky that most people couldn't give a damn.

 

[hobgoblin]

Zarsus[/url]":3s6rthu1]

hobgoblin[/url]":3s6rthu1]

Musafir_86[/url]":3s6rthu1]-Hmm, so how about what Intel's (allegedly) doing? http://www.eetimes.com/author.asp?secti ... id=1318894 (via http://legitreviews.com/news/15752/).

Regards.

That could be claimed to be a smart compiler. But then the Intel compiler have in the past produced code that would drop back to 386 instructions if the CPU didn't id itself as "Genuine Intel"...

At this point, the only compiler i would trust to produce impartial binary is GCC.

Intel's compiler would degrade performance for non-Intel CPUs, but this doesn't make the compiler bad, it just makes the compiler bad for other CPUs, it would certainly produce valid output. Intel doesn't cheat on this regard, it's a testament to their compiler that it's able to identify useless benchmarking code that ultimately doesn't produce any visible results and removing or re-architect them. For an example you can check out the comment of this StackOverflow question:

Intel Compiler 11 does something miraculous. It interchanges the two loops, thereby hoisting the unpredictable branch to the outer loop. So not only is it immune the mispredictions, it is also twice as fast as whatever VC++ and GCC can generate! In other words, ICC took advantage of the test-loop to defeat the benchmark...

Ok, so it was clever in regards to the Antutu test loop. But the incidence i was referring to was how ICC would ignore agreed upon feature markers for things like SSE, instead focusing on the supposedly cosmetic "Genuine Intel" label. There is one example where ICC compiled code ran 30% faster on a Via CPU that faked the "Genuine Intel" label. That, IMO, makes ICC very much untrustworthy for use with benchmarks or anything similar.

 

[abhi_beckert]

lowlymarine[/url]":2h377hke]SunSpider these days is a test of how well your JS engine cheats at SunSpider, nothing more.

I'm pretty sure the primary goal of sun spider is to make sure every feature or bugfix added to WebKit either makes it faster or has no performance impact.

My understanding is they run sun spider as an automated test for every change to the code, and if your change results in even 1 millisecond slower performance then it will be rejected automatically.

Team policy is every change must make the engine faster or have no measurable impact. Otherwise it is not allowed to be on the master branch.

 

[doppio]

Dick move by Samsung, trolling the tech enthusiasts like that.

 

[Violynne]

I've never been concerned about benchmark data. Regardless how well the tests are designed, they simply don't represent real-world use. Over time, software has a way of slowing down speeds and taking up memory simply because programs just aren't designed well anymore.

Usually, software is simply "upgraded", taking old code and applying bandaids to get it to work for the new system.

Over time, these start to gum up the inner workings. Even as I was reading up on Google Android, one of the most prominent comments I read was my phone *will* degrade over time due to apps.

I'm going to try and stave this issue by reducing the apps I install (so far, only 8), but to me, this seems to be counter-intuitive of using a *smart*phone.

 

[eikonoklastes]

This calls for Booster Justice.

 

[doppio]

Violynne[/url]":2qvwfw65]...
Over time, these start to gum up the inner workings. Even as I was reading up on Google Android, one of the most prominent comments I read was my phone *will* degrade over time due to apps.

I'm going to try and stave this issue by reducing the apps I install (so far, only 8), but to me, this seems to be counter-intuitive of using a *smart*phone.

That's not entirely true though; your phone will slow down as long as it is running many apps concurrently, but it will not run any slower if you install an app but don't run it. I have over 60 apps, and my phone is just as responsive as on day 1 (well, it's actually faster, since I removed some of the always-on apps that came with it).

You do need to take care not to clog the phone with autostart and background apps. Not much different from having to clean your house and walk out the thrash occasionally -- the alternative is renting a room in a hotel where someone else cleans up for you, with the added inconvenience that you're never really at home, you lack some amenities, and you pay extra.

 

[fuzzyfuzzyfungus]

doppio[/url]":114dkblu]

Violynne[/url]":114dkblu]...
Over time, these start to gum up the inner workings. Even as I was reading up on Google Android, one of the most prominent comments I read was my phone *will* degrade over time due to apps.

I'm going to try and stave this issue by reducing the apps I install (so far, only 8), but to me, this seems to be counter-intuitive of using a *smart*phone.

That's not entirely true though; your phone will slow down as long as it is running many apps concurrently, but it will not run any slower if you install an app but don't run it. I have over 60 apps, and my phone is just as responsive as on day 1 (well, it's actually faster, since I removed some of the always-on apps that came with it).

You do need to take care not to clog the phone with autostart and background apps. Not much different from having to clean your house and walk out the thrash occasionally -- the alternative is renting a room in a hotel where someone else cleans up for you, with the added inconvenience that you're never really at home, you lack some amenities, and you pay extra.

Pre 4.3, Android didn't do TRIM on internal flash memory (or SD cards; but apps are less commonly run from those, and those are easier to nuke-and-pave if needed), so 'slowdown' was a definite thing. The Nexus 7 apparently has a lot of high-profile trouble with this. It's related to disk activity, not apps per-se, so 1 app that spends its time writing and releasing chunks of internal storage could cause trouble where 50 that just poke at the internet and live in RAM might not; but it is a known problem.

Barring issues with specific flash controllers, though, 4.3 is supposed to do for Android what TRIM support did for SSDs on the desktop.

 

[Walt French]

koolraap[/url]":218wyopv]As for who's responsible? It could be the marketing department. It could just be Samsung's Programmer of Legend(tm) who squeezed 10% more out of the phone during one busy coding frenzy. "See boss? It's fast. And it's got the battery life it says on the box."

OK, so there's more than one possible explanation.

All this could be easily cleared up by Samsung letting the world know how it happened that they shipped code—apparently, carefully vetted and approved by carriers' rigorous quality-control processes that delay updates, BTW—that has its biggest impact on deceiving consumers about the speed and efficiency of how the phone will perform for them.

Whether or not that was their intent, consumers worldwide generally think that when ordinary business integrity fails, their nations' laws protect them from deceptive practices. Even if there's no law on the books that says you can't tweak the code so it only looks fast in burst mode in a way that fools the naïfs at TheVerge. I don't see this as any less deceptive than hiring astroturfers in Taiwan, so we've now moved from the “See boss?” having slipped thru QC, into a broader pattern of deceiving customers. It's not likely to end well.