Mobile Benchmark Tests Explained!
Benchmarking is not a new term in computing world, It has been around for decades. Starting from the same spirit to measure and quantify the performance of PCs to servers, and in the last couple of years, Mobile Benchmarks have gained significant ground to compare smartphones and tablets from all over the industry.
Of course in the race to prove who’s best and by how much, there has been a widespread proliferation of misunderstood and biased benchmarking in the mobile industry. Now the purpose of this writing is very clear so that next time you come around those flashy ‘Benchmark Scores’, you really understand what they mean, their significance and then judge a device. You definitely would have heard many ‘benchmarking scores’ from AnTuTu or GeekBench being the most famous ones. Ever tried to go deeper and look at what they consider, while giving certain scores to a device? and what they don’t? How significant and influential these benchmark scores should be is very crucial to your understanding, perplexed? Don’t worry pal, we got you covered.
What do those AnTuTu & GeekBench Scores Actually Mean?
In recent times, these benchmarks have been much influential and breeding ground for many widespread myths, remember the more megapixel-better camera thing? or the more cores better processor? These are some famous examples of how benchmarks can create myths in minds and often deviate you from the concepts from which you should be judging a mobile device. These benchmarks are simply run to see the fastest theoretical performance that the system could be tested at, without regard to battery life, operating systems, applications or real-world use cases. Many of them simply load up all of the cores to their maximum frequency, which phones never operate at other than benchmarks. As a result, these benchmarks have been given the label across the industry as "inaccurate or inappropriate benchmarks" that don’t accurately represent a user’s experience.
Does it matter?
Yeah, It does and while you are still puzzled with the question, “why should you care?”. Let us drop a few instances, which you’ll relate to quite nicely. Ever considered SoC pricing, they always are strongly in correlation with the perceived performance, Isn’t it?
Here’s an amazing one :
"£99 Tesco tablet beats £300 Apple rival in speed test: Consumer study shows price and brand is not guarantee to finding best performing device"
And while you are string meddling down with this, let us tell you that this came in as headline in ‘The Daily Mail’ and they are right, In GeekBench scores, Tesco one does beat the iPad, but really? This is set in stone that the Tesco one doesn’t even come close to the iPad in real benchmarks or experience. This is one of major examples of misunderstood benchmarks. And this we think should answer your question, why should you care?
Let’s discuss another facade, If you are in SoC business like the major ones, take Huawei, MediaTek, Qualcomm or Samsung. The motto for the day will be, “If you can’t beat them, join them” and you’ll be cramming your SoC’s with more cores, their gentlemen is the reason for the 8-core myth. Thankfully, Apple and Intel have stuck to their standards and decided not to go in wonderland by adding more and more cores and they deserve appreciation for denying to mislead customers.
The Poster Boys for Benchmarking: AnTuTu and GeekBench
Okay, so there’s no argument here, AnTuTu and GeekBench are the most used benchmarking tools, right now. AnTuTu and GeekBench are usually used by manufacturers and most reviewers to judge a mobile device based on a quantification provided by these benchmarks. The issue with these benchmarks is that they don’t actually judge a device based on its overall performance and experience, they just run theoretical tests on individual components of the SoCs or some other individual part of the whole device. Now, these are not always wrong, they do suggest the capabilities of a CPU or GPU up to a certain extent but are wrong interpretations for the combined system’s performance and experience.
What topples this is, that as these benchmarks test a device for certain individual parameters, Smartphone manufacturers and OEMs often trick these benchmarks successfully. This is flat out cheating people, you don’t believe me? check this out.
And while you would be consolidating your trust that this would have been done by minor companies or newcomers, shockingly you are wrong! Major players such as Samsung, LG, HTC and ASUS were all caught fooling the AnTuTu benchmarks by tweaking their devices to perform well in the parameters, AnTuTu works upon. How convenient?
Let’s find out more about these benchmarks, which will definitely help you out to get the best of these otherwise great tools.
AnTuTu Benchmark Score
Being the most famous and pervasive benchmark of the Android world, AnTuTu has a simplistic workflow, where on pressing a button “Run benchmark”, It would trigger a host of benchmarking procedures internally, quantifying the performance on a system based on several different system tests. AnTuTu initially started as a single app on the play store, but now offers a variety of tests under different apps. Good strategy huh! Including display tests, HTML5 tests and yeah, stability and battery too.
Sticking to the basics, It takes into account single thread floating point, single thread integer, multi-core floating point and multi-core integer performance of the CPU. Well, these are not very explanatory, but they do indicate the maximum capabilities of the CPU, given that one day apps will be demanding that much from the hardware.
AnTuTu is a more comprehensive analysis that also considers other aspects of the hardware, including memory. The benchmark also tests RAM performance in terms of read and write speeds. Multitasking performance and I/O are also put to test. Again, these tests are based on theoretical limits and hence how true to stand in real-world scenarios is still questionable. The last couple of tests involve 2D rendering performance by those bouncing and dropping shapes, followed by a very low 3D test that’s just lame in the name of 3D games. After gathering results from all these tests, the scores are compiled, combined to give you your device’s AnTuTu score, an aggregated quantification on the performance of the phone, what does this amount to practically? Nothing.
The issue with AnTuTu is that it is perceived as a full system benchmark and is expected to provide a full system benchmark, whereas all it lives up to is testing individual components of the hardware, on crude parameters that don’t amount to much in real world and based on the weightage of individual parameters, a score is calculated. That’s your AnTuTu score, well not your’s, your device’s. The weight given to these specific parameters might not agree to your usage or most of the others too. They are just as decided by AnTuTu as DxO mark does in case of cameras. Here’s a good resource to understand DxO mark ratings and understand what the number actually signifies.
AnTuTu’s main test application doesn’t incorporate any of the other tests they offer like HTML5, battery life, video, stability or screen test which could provide better insights into system performance. This is because if they did incorporate these tests, it would take too long to test and wouldn’t really be as popular of a test as it is today. Instead, they went a bit clever and divided these into individual apps, so if you want a detailed report, install all the apps, and run the tests individually, If you have the time and patience for that.
GeekBench by Primate Labs, is a cross-platform benchmarking tool that started out as a benchmarking tool specifically for iOS and Mac. It then got the capability to run and judge Android devices as well, and given its ability to provide a common platform for benchmarking both iOS and Android devices, the fanboys love it, the reviewers love it, leading to its widespread fame in just a few years.
GeekBench is a bit limited though in its approach and just measures certain aspects of CPU performance and memory. It is available on iOS, MacOS, Android, Linux and Windows too. Relatively, GeekBench isn’t that bad or misleading/misunderstood compared to AnTuTu. It has a limited approach and takes into account only CPU and memory, and no wonder that too under theoretical scenarios. That’s fine GeekBench, but you miss a really significant component, GPU. GPUs are becoming go-to commodity hardware and the computations in modern apps also take advantage of the power of GPU to provide a better experience. So missing out on GPU is not a cool thing.
The tests it runs are CPU integer and floating point calculations in both single core and multi-core modes as well as memory single core and multi-core. As a result, the benchmark may provide some insights as to the architectural comparison of the CPU in the system, but in the case of two different smartphones with the same SoC, this benchmark provides limited to no value.
So essentially, GeekBench only compares CPUs of the two or more devices in question, but what if both have the same SoC? Well, that’s just of no value here. This benchmark should only really be used to compare CPUs from different operating systems and platforms and how they stack up against each other, not a benchmark for comparing phones, especially not for a review.
But wait, If it isn’t all-inclusive and shouldn’t be used for reviews, why is it used so intensively all around. Yeah, they do use it intensively, but now you know that it isn’t doing anything of significance there. Geekbench should be commended for their transparency of how and what they’re testing exactly, not just an amalgamation of the whole sort of different score from a multitude of components in your device. This fact doesn’t wipe the fact that it is used inaptly by most, leading to misunderstandings and occasional fan-wars.
Benchmarking the Benchmarks: SuggestPhone’s Take
Benchmarks are a good tool to gauge the theoretical performance of your device but they shouldn’t be as influential as they are currently. Due to the widespread use of such benchmarks, and misunderstandings that have deeply rooted in the mids of people. We see big smartphone manufacturers tinker their devices to pass these benchmarks with flying colours, albeit in reality, they suck. And this is not a small issue, major tech giants such as Samsung spending millions and tons of engineering resources to cheat these exams which were meant to test them is not good for healthy benchmarking as well as customer trust.
After all, if so many people are using or mischaracterising AnTuTu and Geekbench, it lends them credibility even when it shouldn’t. It seems like those same resources could be working on further improvements to issues we all have, things like battery life.
What is unexpected is that many reputable tech-blogs and publications user these benchmark scores out in open without any explanations lending more credibility to these tests. Benchmarks are supposed to lend credibility to your experience and explain any kinds of performance differences between devices, be they good or bad, certainly they are not doing that, not which we can fully relate to.
What’s your view of the benchmarking practices? Which one in your opinion is a better benchmarking tool? Do tell us on Facebook and make sure to follow us on Twitter for quality technical articles. Also, we are all ears to answer your questions regarding smartphones and technology. Be sure to say Hi!
Latest Articles on our blog
In this article, we take a look at the top 10 best earphones under 500 and 700 available in India as of May 2019. These budget earphones offers decent sound quality and bass. The earphones are from companies like boAt, Leaf, Boult, MI and Realme which all are making a strong mark to offer you the best in this price range.
In this article, we list down the best car chargers in India as of 2019. We have listed the specs and features of the 9 best car mobile chargers. These car chargers are priced between Rs. 300 and 1000 and most of them offer dual USB ports with Qualcomm Quick Charge 2.0 or 3.0 technology for fast charging. These chargers are suitable for both Apple iPhones and Android phones.
In this article, we cover the detailed comparison of the processors in limelight for the phones in Rs. 10000 to 20000 price range: Qualcomm Snapdragon 660 vs Snapdragon 675 vs Helios P70 processors. A point by point comparison of cpu, gpu performance, camera, connectivity, audio quality and benchmarks is covered here.
In this article, we bring you the top 10 cases & covers for Redmi Note 7 and Redmi Note 7 Pro. We have different types of cases and covers like flip covers, soft transparent case, rugged armor case, shock proof case and printed covers. These covers are priced from around Rs. 200 to 600 in India. All these covers will protect your phone from drops and will give it a nice new look as well.
In this blogpost, we take a look at the 9 best wireless bluetooth earphones & headphones under 2000 in India 2019. These earphone and headphone use bluetooth technology to connect to your phone and stream music. If you are a regular wired earphone user, you would love the wireless experience.