Geekbench extends ban to the Galaxy Tab S8 series, Samsung CEO apologizes for throttling...

Humza

Posts: 1,023   +171
Staff member
In context: Samsung landed in hot water recently after its latest flagship, the Galaxy S22, was discovered to be selectively throttling apps using Samsung’s Game Optimizing Service, while excluding other apps and benchmarking tools like those from Geekbench. The latter determined it as benchmark manipulation and proceeded to ban all Galaxy S series models using GOS, all the way back to 2019’s Galaxy S10.

During their in-depth testing of the Galaxy Tab S8 series, Android Police found both the S8 Ultra and S8+ variant to be throttling performance when they ran a modified version of Geekbench, renaming it to the popular game Genshin Impact.

The publication observed the tablets' performance declining by 18-24 percent for single-threaded CPU scores and 6-11 percent for multi-threaded scores. Moreover, older Galaxy Tab models, such as the S7, S7 FE and S5e didn't demonstrate this behavior during testing.

The culprit is believed to be Samsung's Game Optimizing Service and an associated list of 10,000 apps, which resulted in Geekbench delisting four generations of Galaxy S phones, plus legal troubles for Samsung.

Interestingly, Android Police notes that GOS doesn't affect all apps and is a sophisticated service that takes into account device temperature, power consumption and other system factors. Nevertheless, Geekbench promptly moved to delist the Tab S8 series for benchmark manipulation.

The app throttling controversy also resulted in an apology from JH Han, Samsung's CEO of Device Experience, during a recent shareholders meeting. Although Han defended the use of GOS by noting that the service only limited CPU and GPU performance to an extent while not disturbing the gaming experience, the CEO promised that such an incident wouldn't happen again.

Samsung has also recently started pushing out a GOS update in South Korea for the Galaxy S22 to address this concern. While a global rollout for other S22 devices can be expected soon, it remains to be seen if this update also becomes available to older, Geekbench-delisted Galaxy S phones and the Tab S8 series.

Permalink to story.

 

Dimitriid

Posts: 2,216   +4,268
What's funny is that the only reason anyone cares about geekbench 5 on mobile devices is because of the tech arms race promoted by companies like Samsung and Apple. In the past they were able to get away with it by using clever bait and switch: all their press materials were based on Qualcomm chips but then selling most international models with their own, noticeable slower Exynos chips.

I say noticeable slower but the caveat is that they're slower for benchmarks mainly. There's actually nothing wrong with Exynos phones they work just as well as Galaxy flagships because most of those Galaxy flagships can go back 4 or even 5 years before most users actually notice performance issues and most of those are likely at least partially related to battery degradation and the fact they probably throttle more frequently as battery capacity naturally gets worst over time.

At least Apple has shifted towards pushing their premium tech advances to places where you know, it actually might matter like laptops and desktops oriented to a "prosumer" type of customer that is expected to be a content creator as well as a user.

Truth is faster chips are nice, but user replaceable and larger batteries could account for what 99% of the people who buy even flagship phones actually use day to day.
 

Theinsanegamer

Posts: 3,443   +5,764
What's funny is that the only reason anyone cares about geekbench 5 on mobile devices is because of the tech arms race promoted by companies like Samsung and Apple. In the past they were able to get away with it by using clever bait and switch: all their press materials were based on Qualcomm chips but then selling most international models with their own, noticeable slower Exynos chips.

I say noticeable slower but the caveat is that they're slower for benchmarks mainly. There's actually nothing wrong with Exynos phones they work just as well as Galaxy flagships because most of those Galaxy flagships can go back 4 or even 5 years before most users actually notice performance issues and most of those are likely at least partially related to battery degradation and the fact they probably throttle more frequently as battery capacity naturally gets worst over time.

At least Apple has shifted towards pushing their premium tech advances to places where you know, it actually might matter like laptops and desktops oriented to a "prosumer" type of customer that is expected to be a content creator as well as a user.

Truth is faster chips are nice, but user replaceable and larger batteries could account for what 99% of the people who buy even flagship phones actually use day to day.
People care because we use benchmarks to compare devices, not solely because of some push by samsung and apple.
 

Plutoisaplanet

Posts: 735   +1,162
If Apple did this the unwashed would be out in force demanding their blood.
Apple does throttle CPU performance to extend battery life, and they did it at first without telling anyone. It applies to older phones based on the battery aging due to reduced peak energy capability, preventing unexpected shutdowns. Four years ago was when they came clean about it offering battery replacements for $29 for a year period, which would solve the issue temporarily. https://arstechnica.com/gadgets/201...-to-preserve-battery-life-benchmarks-suggest/
 

Sausagemeat

Posts: 1,597   +1,422
Apple does throttle CPU performance to extend battery life, and they did it at first without telling anyone. It applies to older phones based on the battery aging due to reduced peak energy capability, preventing unexpected shutdowns. Four years ago was when they came clean about it offering battery replacements for $29 for a year period, which would solve the issue temporarily. https://arstechnica.com/gadgets/201...-to-preserve-battery-life-benchmarks-suggest/
Yes I remember. But it was only on older devices. You didn’t buy a brand new iPhone and found that the performance didnt match its own benchmarks. And Apples intentions were to preserve the battery, not to deceive people with fake benchmark scores. Also I believe several other manufacturers have d
one the same thing, as batteries get older they degrade. I would rather see a lower clock speed on an old device than have it cut out all the time.

Apples “batterygate” is not the same as what’s going on here with Samsung,
 

Watzupken

Posts: 644   +526
The truth is that all modern mobile SOCs will throttle to keep power and heat in check. May it be Samsung, Qualcomm or even Apple SOCs, they will all throttle. The issue at hand is that Samsung chose to allow the SOCs to run at full speed in benchmarks, which in real life usage is nowhere close in terms of performance. In fact the discrepancies is very significant, that it became such a big issue. Which is why I feel we should always take benchmark results with a pinch of salt since they are easily be manipulated.
 

Watzupken

Posts: 644   +526
Apple does throttle CPU performance to extend battery life, and they did it at first without telling anyone. It applies to older phones based on the battery aging due to reduced peak energy capability, preventing unexpected shutdowns. Four years ago was when they came clean about it offering battery replacements for $29 for a year period, which would solve the issue temporarily. https://arstechnica.com/gadgets/201...-to-preserve-battery-life-benchmarks-suggest/
This is a different problem. The main reason Apple provided for throttling older devices is based on their battery health so as to make sure the phone does not shut down due to sudden power spikes. The throttling in this case would be consistent, whether you are running day to day application, games or benchmarks. Whereas here, Samsung is throttling both new and old devices to keep power and heat in check. Yet they chose to selectively allow the SOCs to run at max or close to max potential just so that they provide some impressive benchmark numbers. And the performance difference between full potential and throttled is very significant as shown in testings. So as a consumer, you think you are buying a mobile device because it offers “great performance” based on benchmark scores. But in reality, your last 1 or 2 gen phone may be just as fast if not faster.