It was discovered yesterday that Samsung allegedly tampered with the Galaxy S 4 in order to provide the best possible benchmark scores in different apps. However, Samsung has addressed the allegations and says that they haven’t done such a thing, saying that they didn’t use any sly tricks to achieve higher benchmark scores. AnandTech reported
There’s no doubt that on paper, the way Sony’s designed the PS4 and the AMD components it’s selected offer 50 per cent more power in the GPU department over the Xbox One. But does it really make any difference when you get down to playing a game?
     |
The Samsung Galaxy S4 is a damn fine phone. It’s fast, it’s big, it’s thin and it’s completely deserving to be in most people’s pockets. But apparently that’s not enough for Samsung. Samsung has gotten caught cheating benchmark tests to make itself look better. Basically, the S4 is specifically tweaked to be better at benchmarking than anything else it does.
| |
Benchmarks often catch the ire of critics because their scores don’t necessarily reflect real-world experience, and it appears that some Samsung devices aren’t helping that reputation out at all. Our friends at Anandtech started sifting through data for two Exynos octacore variants of the Galaxy S 4, and discovered multiple benchmarks that appear to run those devices at higher CPU and GPU clock speeds than they normally run. Some of the apps included in the report are GLBench 2.5.1, AnTuTu, Quadrant and Linpack. Coincidentally, the site also found code within the GS4 that indicates the existence of a program called “BenchmarkBooster,” which is responsible for overclocking the processors when certain apps are running. Is this an evil plot by Samsung to take over the world using rogue benchmarking results? Is this something other manufacturers are doing? It’s hard to say, but there is certainly something curious afoot. Check out the source to get all of the skinny on what exactly is going on.
Filed under: Cellphones, Wireless, Mobile, Samsung
Via: 9to5google
Source: Anandtech
Samsung tailored its Galaxy S 4 to deliver the best possible scores on popular Android benchmarking tools, investigations have revealed, despite apps potentially not getting the same power for real-world use. The AnandTech research was sparked by claims Samsung was reserving its fastest graphics chip speeds for select benchmarking apps alone, with games and other software only ever seeing slower performance from the Exynos 5 Octa processor found in select models of the Galaxy S 4. The motivation behind the tinkering appears to be to ensure the flagship smartphone posts consistently high benchmarking numbers for comparison with other devices, even if that doesn’t necessarily translate to its everyday abilities.
Concerns about the clock speed the GPU ran at during testing began after it was noticed by Beyond3D users that the Galaxy S 4 ran its graphics chip at 533MHz when certain benchmarking apps were used. During the rest of the time, however, the GPU ran somewhat slower, at 480MHz.
A similar process was spotted during CPU testing, with the Galaxy S 4 automatically switched to a certain clock speed when select benchmarking applications were running. When AnTuTu, Linpack, Benchmark Pi, GFXBench 2.7, or Quadrant were loaded, the Galaxy S 4 would push its processor to the maximum frequency supported by each of the four cores. The behavior was spotted on both the Exynos 5 Octa and Qualcomm Snapdragon powered versions of the handset.
The claim is that Samsung has specifically tailored how the Galaxy S 4 reacts to benchmarking by the user, aiming to make sure the phone always looks its best. In reality, the situation is somewhat mixed: the CPU, even in its locked state, never reaches a speed that’s unobtainable to individual applications.
However, on the GPU side, the 533MHz reached during testing is not, apparently, made available for users’ apps. Samsung, it’s pointed out, never actually promises a certain GPU clock speed from the phone, but it raises questions about misleading expectations when on-paper performance doesn’t translate to real-world performance.
Benchmarking has always been a dark art, with questionable relevance for most users. Nonetheless, there are some device owners who enjoy knowing how their smartphones and tablets compare to the rest of the market, and it seems Samsung is doing them a disservice by not being entirely transparent about how its devices treat such testing.
We’ve contacted Samsung, which tells us that there is not currently an official comment on the report. We’ll update when we hear more. Update: Samsung has commented on the benchmark findings.
Samsung Galaxy S 4 artificially tuned for benchmarks research spots is written by Chris Davies & originally posted on SlashGear.
© 2005 – 2013, SlashGear. All right reserved.
This evening, the new Nexus 7 was spotted in a couple shots posted by a Reddit user, complete with packaging and a timestamp. A few short hours later, the folks over at Android Police have produced a full roster of specifications and benchmarks for the 7-inch slate, giving the most detailed look at the upcoming Nexus 7 yet.
The Nexus 7 is codenamed Razor, and according to the benchmarks features a 7-inch 1920 x 1200 display, as various leaks have suggested. With the bar on the screen, the actual resolution works out to 1200 x 1824 pixels. Under the hood, there’s a Qualcomm Snapdragon S4 Pro quad-core 1.5GHz processor and 2GB of RAM. Andreno 320 is pegged for the GPU.
Also as leaks have suggested, the slate runs Android 4.3, but we’ll have to wait a bit longer to find out what exactly the changes that come with that entail. As far as the hardware goes, the new Nexus 7 is pretty much on par with the Nexus 4, although the Quadrant benchmarks for the slate show it scoring a bit higher at around 5500 rather than 4200.
The Quadrant benchmarking tool is akin to the more commonly known AnTuTu, and looks at 3D graphics, I/O, and CPU. In the screenshot, we see a comparison between this device and some other handsets, with it scoring above the HTC One X, Motorola ATRIX 4G, Galaxy Tab 10.1, Galaxy Nexus, Optimus 2X, and HTC Desire HD.
A 3DMark graphics benchmark was also run, with Android Police stating they ran it in “extreme” mode with a 1080p action sequence. During the test, the new Nexus 7 scored 7,188, with it staying around the 30fps mark. This is in comparison to the Nexus 4, which scored around 6300, and the previous Nexus 7, which scored 1877 on the Ice Store Extreme test.
SOURCE: Android Police
Nexus 7 2 full specs and benchmarks surface with codename “Razor” is written by Brittany Hillen & originally posted on SlashGear.
© 2005 – 2013, SlashGear. All right reserved.
Earlier this month, the AnTuTu benchmarking tool raised a bit of a firestorm when it showed Intel‘s Atom Z2580 out performing offerings from Qualcomm, NVIDIA, and Samsung. This led to BDTI, a consulting firm, showing that not all instructions were being executed by the Intel processor, causing a false boost in results. As such, AnTuTu has issued a revised version of the benchmarking tool, and Intel has fallen 20-percent as a result.
Reportedly, this entire issue was the result of the ICC compiler, which was introduced in version 2.9.4 and utilized for Intel processors, while other processors utilized a GCC compiler. The issue caused the RAM test to be incomplete, causing the artificially high results. The revised version is 3.2.2 and was released this past Wednesday.
The ICC compiler is still used for Intel processors in the new version, but the new benchmarks scores are vastly different, with the Atom Z2580 taking a hit of about 20-percent and the RAM score in particular falling by half. Other scores are said to have been marginally affected, but the big changes center in these two areas.
Meanwhile, the results for the Qualcomm Snapdragon 600 and the Samsung S4 Oct were relatively unchanged, with the Exynos 5 Octa now ranking higher in performance the Intel Atom Z2580. Next month AnTuTu plans to issue new testing standards, at which point we’ll get a chance to see what kind of result changes might surface with the next big release.
AnTuTu did not reveal what changes it made to the newly released version to correct these issues, and some are questioning whether they’re now accurate. Looking at other benchmarks for the processors, however, the new AnTuTu results fall more in line with what we’ve seen elsewhere.
SOURCE: EE Times
Intel Atom Z2580 AnTuTu benchmark falls 20% following revision is written by Brittany Hillen & originally posted on SlashGear.
© 2005 – 2013, SlashGear. All right reserved.
While most high-end Android phones currently sport Qualcomm’s Snapdragon 600, there’s another chip announced earlier this year waiting to hit the scene: the Snapdragon 800
| |
Qualcomm may have introduced the Snapdragon 800 processor a little while back, and while we suspect some may be wondering why we have yet to see any real announcements since that point in time, it seems there was an actual plan in action. Just to give a bit of background, the plan was to announce
This is site is run by Sascha Endlicher, M.A., during ungodly late night hours. Wanna know more about him? Connect via Social Media by jumping to about.me/sascha.endlicher.
