updated 01:35 am EDT, Wed May 7, 2014
Testing of the newest Galaxy phone shows strong graphic, everyday performance
In the review for the Samsung Galaxy S5, we covered the ins and outs of the phone as regards features. What wasn't done at the time was raw performance testing measured through benchmarks. Going through the phone as it was reviewed, we noticed that the Galaxy S5 was swift because of its Snapdragon processor, but just how fast was it? More than that, how does the phone stack up in areas of everyday use compared to other phones on the market?
3DMark is the go-to benchmark when it comes to graphical testing, be it for desktop PCs or handheld devices. The Icestorm test for phones and tablets features a 720p or 1080p graphics test, and a physics test to gauge a phone's ability to render real-time gaming situations using OpenGL ES 2.0.
Testing 3DMark for Android touches on the high-end graphical capabilities of the device. Both of the lower modes of the Icestorm benchmark capped out for the Galaxy S5, leaving only the unlimited run to push the phone to a measurable score. Scores from graphic processing physics are averaged to make an overall score.
For these specific tests, we ran the benchmark five times, disregarding the first and second runs. The median score should thus represent an average user. This score for the Galaxy S5 came in at 18,338, which scored it lower than the Samsung Galaxy J and Galaxy Note 3 by 392 and 424, respectively. Historically the Note series has run slightly faster than the S series, so this test offers no real surprise. However, the ability to top 104.6 FPS in the first graphics test and maintain 49.8 FPS during its physics run is a significant feat.
The Galaxy S5 should prove to be quite capable running gaming or graphic intensive applications despite the lower scores, thanks to the Adreno 330.
Rightware BaseMark X
BaseMark X makes use of the Unity 4.2 game engine that is featured in numerous Android games, including Dead Trigger 2 and Deus Ex: The Fall. Two different tests, one in a desert scenario and another in an airport base, test heavy graphics rendering, including dynamic shadows and overlay textures. Rightware even provides the equation for score calculations, for the curious. Comparative results for devices can also be accessed on Rightware's Power Board.
Runs through BaseMark X are approached in the same manner as 3DMark. Five test runs are taken, then the first two scores are disregarded and the median of the remaining scores is selected as the result. The benchmark was run in high quality, since the phone had given the indication it could handle high rendering loads in the 3DMark test. Unfortunately, it still wasn't able to outpace the iPhone 5s, which holds the crown. The iPhone was able to beat it by 1368 points, a margin of 11.6 percent.
Oddly, the Verizon version of the Galaxy S5 slightly outpaced the international unlocked version. These phones often manage to squeak out more power, but the result here is little more than 0.21 percent decrease from the subsidized phone. It will be interesting to see what the developer version of the phone is capable of.
Rightware BaseMark OS II
BaseMark OS II is a all-in-one tool used to measure the general performance of the device. Factors in the final score are determined by four categories of tests consisting of system, memory, graphics and web browsing. Free and paid versions are available to download. For the purpose of this test, we used the free version. As with the previous trials, tests five runs are taken. The first two are discarded and the median score selected as the official result.
The Samsung Galaxy S5 proves to have good system balance overall, coming in second in the results. While doing the test runs, Rightware actually awarded the phone the silver medal consecutively in test runs with results of 1104 and 1108. The only phone to outpace it so far is the Sony Xperia Z2, which has a 121 point lead on it.
Scores for the iPhone 5S put it 84 points behind the S5. While not huge numbers over the Xperia Z2, it still marks an 8.2 percent improvement over last year's Apple model.
Passmark is another tool that measures the overall performance of a smartphone. The benchmark conducts 17 different tests focused on the CPU, onboard storage, memory and 2D/3D graphics. Unlike other tests, Passmark makes use of all cores when testing the device. Passmark lists the results of all tests at the website, but it also mixes tablet and smartphone scores together. Our results will only list Android operating systems.
Results for the Galaxy S5 were a little mixed, considering it scored lower than the Galaxy S4. Finding valid comparisons for the phone provide to be troublesome as well, since performance of the phone sits right at the edge of mid- to low-end tablets. It was expected that the Galaxy Note 3 and Galaxy J would outscore it, but the S4 scored 252 points, or six percent, better than the newer phone.
The HTC One M8 also provided better results than the Galaxy S5, but still not enough to overcome the Galaxy S4.
Vellamo was developed to test the capabilities of Android operating system chips, starting with version 2.3 pertaining to graphics and the web. Fourteen tests focus on running the gamut of HTML 5 for web browser performance, while it tackles CPU subsystem performance in six different tests. While Vellamo doesn't provide comparative scores for the results, it does issue a score to the test. In this case, we will show all five runs to display the consistency of the device.
It should be mentioned that Vellamo claims that the Android 4.4 web engine may not be fully optimized, and this will reflect in the scores. The warning speaks the truth, as devices running Android 4.1 and 4.2 easily outpaced it. End-over-end tests did provide good results, showing that the Galaxy S5 will consistently perform in CPU calculations and web applications. Metal tests only saw a 22 point gap from the high to low score, while the distance between the high and low for HTML was only 14 points.
Ranking-wise, Vellamo places the Galaxy S5 between the Galaxy Note 3 and the Sony Xperia X Ultra for its CPU testing. This is another area where the phone performed as expected. One would think the increase from the S5 to the Note 3 should only be between eight and 10 percent. The HTML 5 ranking was curious, as the S5 ended up pretty far down the list between the Motorola RAZR i (Android 4.1.2) and the HTC One X (Android 4.2.2).
In this test, we show all five results to measure consistency. All tests are done using 3G cellular networks, or LTE when available, rather than Wi-Fi. All results are measured in milliseconds, collecting the results of five different tests.
The Galaxy S5 burned through these tests, offering times from 6623.1ms to 7006.2ms. That's only a 5.7 percent difference from the fastest to slowest times. Offhand testing places the results anywhere from 3000ms to 9000ms difference, depending on the phone. For example, an iPhone 5c in a high saturation area provides results over 15,000, while a RAZR Maxx HD pushed 10,000 in a low saturation LTE area. The Galaxy S5 was tested on the same area, but on different days.
Battery Drain / Charge Times
Three long battery tests were conducted to get a picture of the real day runtime of the device. A music test is conducted, which consists of playing a 256kbps MP3 on repeat through the device speaker, as if someone was listening to music off of the phone without paying much attention to the device. Video testing consists of a 1920x1080 HD video running on a constant loop, meant to simulate watching a video from device memory. Finally, a web test was done that cycled through 10 websites and set intervals to simulate web surfing over Wi-Fi. All tests are run from 100 percent to 15 percent battery capacity while the screen remains active with Wi-Fi, cellular/LTE and Bluetooth enabled.
A charging test is done for comparative results, and to give potential customers an idea of how long they will have to wait for their phone to charge in an emergency. The device is thrown on a charger at 10 percent and isn't removed until 100 percent. No trickle charging is involved, as the intent is to show a normal charging cycle.
When it comes to charging, the Galaxy S5 is quick. Thanks to USB 3.0, charging was completed in a little over 103 minutes. This is impressive, considering many modern phones can take closer to 2 hours for a charge. While this may only shave 20 minutes of charge time from some devices, it should be pointed out that the phone has a 2800 mAh battery. To put it in perspective, the iPhone 4S only has a 1432 mAh battery and takes longer to charge.
Use times for the phone are also well beyond what was expected given, that these results are recorded while the screen was active. These times were accomplished without the use of the ultra power-saving mode that Samsung has built into the phone as well. Performance in this area is strong: if 745 minutes of web browsing left a phone with 10 percent battery remaining, users could put the S5 into the ultra-low-power mode and still be able to access their phones in case of emergency for another 10 or more hours.
A basic volume test, much in the same manner we test the relative loudness of Bluetooth speakers, was run to get a feeling of how a device would perform as if someone was casually listening to music. The phone is placed on a desk with a decibel meter, set to dBA at high, that records the peak decibel reading of three songs at a distance of one foot. Each song is decreased in increments of approximately 25 percent to get a range a device is capable of.
The Galaxy S5 didn't allow exactly 25 percent intervals between testing runs, as the device has 15 clicks from the top end of volume to the low end. This leaves the 25 percent mark a little on the low side for the sake of testing.
The design of the S5 really hurts it in this area, since the phone emits sound from the back. When the phone is set down on a desk, the sound isn't given the chance to disperse very well. It doesn't sit as flush to a surface as an iPhone, but sound doesn't travel far even at a foot away. This problem is exacerbated by the speaker on the phone being weak to begin with.
The Samsung Galaxy S5 is a powerful phone no matter how one intends to use it, though it was never the fastest/best in any of the tests we conducted. The benchmarks tell the tale of a phone that has strong processing capability under its hood, which is good for basic consumers and those who like to game a lot on the go. On the back of the Qualcomm Snapdragon 801, the Galaxy S5 will be a phone consumers will be able to use for some considerable time without noticing a large drop in performance.
Customers should be aware that these benchmarks are only for information regarding a device's general performance in the areas where you might need it most-- gaming and the everyday. In the past, phones have been known to be geared to game the system for the benchmarks, most notably Samsung and HTC in recent history. The HTC One M8 is actually banned from 3DMark benchmarking for overlocking the processors by 20 percent, while Samsung almost faced the same problem for 15 percent inflated results in processing. Samsung has since admitted the cheat, and withdrawn it.
By Jordan Anderson