iPhone X Beats Samsung Galaxy S9 in Benchmarking Tests

[MacRumors]

Early evaluations of Samsung's new Galaxy S9 and S9+ have ranked the S9 display and the S9+ camera above the iPhone X, but when it comes to performance, the iPhone X is still the clear winner.

In benchmark testing of the Samsung Galaxy S9 equipped with an Exynos 9810 chip, the iPhone X, and the iPhone 7 conducted by AnandTech, the iPhone X's A11 chip won in every comparison test, and in most cases, the Galaxy S9 also lost out to the A10 included in the iPhone 7.

Samsung is using two separate chips in its new Galaxy devices: the Exynos 9810 and the Snapdragon 845 from Qualcomm. The Exynos 9810 chip outperforms the Snapdragon 845, but doesn't quite match Apple's A11 Bionic chip.

On a single-core GeekBench 4 test, for example, the Exynos 9810 saw integer and floating point scores of 3,724 and 3,440, respectively, well below the 4,630 and 3,958 scores earned by the A11 and under the 4,007 integer score earned by the A10.

On a WebXPRT test that measures HTML5 and JavaScript-based tasks, the iPhone X's A11 chip scored 352, beating the 178 score earned by the Exynos 9810 and the 291 score earned by the Qualcomm Snapdragon 845.

Simliar results were seen in a Speedometer 2.0 test, with the iPhone X (A11), iPhone 8 (A11), and iPhone 7 (A10) winning out over both of the processors used in Samsung's new devices.

AnandTech was testing a demo version of the Exynos-based Galaxy S9 and came to the conclusion that there was something wrong with the device given its poor scores on the latter two tests compared to the Snapdragon 845, but even had the Exynos 9810 shown performance on par with the Qualcomm chip, Apple's iPhones still offer better performance.

When it came to graphics performance, the iPhone X also came out on top, beating out both Samsung chip variants.

Full benchmarks on the new Samsung devices and AnandTech's conclusions about the Exynos 8910 chip can be read over on the AnandTech site, but it's clear that Samsung is still unable to match the performance Apple can eke out of its chips by controlling both the software and the chip design.

It's also worth noting that there's one other area where Samsung can't compete with Apple as of yet - facial recognition. As CNET points out, the Galaxy S9 and the Galaxy S9+ continue to use a 2D facial and iris recognition system that can't compare to the security of Apple's 3D face scanning technology.

The Galaxy S9 and S9+ use a 2D facial recognition system​

The Galaxy S8, which used the same 2D feature, was able to be fooled by photos, and while the Galaxy S9 features iris technology that's "been enhanced" to recognize "unique iris patterns from greater distances" and to better withstand spoofing attempts, it's the same general system that was used in the previous-generation devices.

Because Samsung's facial recognition feature isn't as secure as Face ID, the South Korean company continues to pair it with fingerprint recognition, a biometric authentication method Apple is abandoning in favor of Face ID.

Apple's Face ID is a 3D system that uses a series of dots to map out a person's face. It can't be fooled by photographs.​

KGI Securities analyst Ming-Chi Kuo has said that Android smartphone makers are two and a half years behind Apple when it comes to facial recognition techniques, so it could be awhile before Samsung has a similar system that can replace fingerprint scanning.

Article Link: iPhone X Beats Samsung Galaxy S9 in Benchmarking Tests

 

[farewelwilliams]

interesting how comments on android fan sites are now saying "but performance doesn't matter anymore"

 

[kildraik]

And the original iPhone had a processor that is far more powerful than the computer in an F-22 Raptor.

Whoop-dee-doo.

 

[1050792]

Beats in benchmarks, loses in real world performance thanks to iOS 11 buggy mess.

 

[mtneer]

This is just day 0 when everything is brand new. It would be quite interesting to see if someone can do a running comparison at increasing battery wear levels - 100, 250, 500 cycles and so on. That will be quite representative of the entire ownership experience.

 

[kildraik]

Beats in benchmarks, loses in real world performance thanks to iOS 11 buggy mess.

iOS 11 is a bug. For some reason when I turn off Wi-Fi or Bluetooth in my control panel, it’s still ”on.”

Edit: I love how this triggered one person, and prompted another to *almost* inaccurately describe how the feature works. Bless their hearts.

 

[mi7chy]

Nobody buys a $1K phone to run useless synthetic benchmarks that claim it's faster than PC CPUs. They will buy it if it can substitute for a PC replacement with DeX.

 

[Apple_Robert]

The iPhone X has been fast for me. I enjoy using the new features. And even though I am a tech geek, I am not hung up on benchmark tests. I also have no need to try and justify my purchase to some Android user, nor do I think the Android user has anything he or she needs to defend. Use what makes you happy regardless of what benchmarks say.

 

[Gen0]

That's just it. It's pretty par for course that the iPhone *appears* faster in every benchmark, but if Apple specifically design the OS with benchmarking in mind that could literally be manufactured in. We know iOS app switching is 10x faster than Android, but we also know that's because they take a picture of the app when you close it and just show you that again. It's the same kind of 'sidestepping the problem' that Apple has done already is what I'm trying to say.

Samsung got stuck before (with the S2 or 3 I forget) as well as LG and others for overclocking their phones when a benchmark is opened.

Apple recently showed that there is no such thing as 'official processor speed' in their books when they downclocked older iOS phones that updated without catching a lawsuit. That actually means Apple could have and can be overclocking their phones under bench apps and it would actually be completely allowed... also, nobody could prove it, and if they could Apple never said anything to the effect of they don't or won't.

 

[Ds6778]

No, benchmarks don’t matter.



But geeze, it’s not even close.

 

[critter13]

the X is fast, im sure the real world difference is not noticeable but its interesting that the gap isn't closer or that the x even won out given the fact that its an older phone. I was blown away by how much faster the X is than my 6. At this point its all personal preference on OS and ecosystem between the top iOS and top Android phones.

 

[FelixDerKater]

A12 en route.

 

[deanthedev]

Why is this a surprise?

Exynos 9810 is terrible. The 9810 obviously throttles heavily. With a single core score that high (and Samsung really did make a big leap in single core performance - their highest in history), the multicore score should be much much higher. But it isn't higher because they obviously can't run all 4 cores wide open. Should have just put 2 fast cores in and let them run full speed instead of 4 throttled cores.

 

[dumastudetto]

Apple is so far ahead of the competition it's almost embarrassing.

 

[deanthedev]

Nobody buys a $1K phone to run useless synthetic benchmarks that claim it's faster than PC CPUs. They will buy it if it can substitute for a PC replacement with DeX.

DEX is complete garbage because Android Apps are garbage (there aren't any good Android tablet Apps so why would you even want to run them with a keyboard/mouse and large monitor).

 

[bruinsrme]

OMG based on one article I’m trading in my X. The next article I’m cancelling my preorder.
So conflicting

 

[deanthedev]

Beats in benchmarks, loses in real world performance thanks to iOS 11 buggy mess.

Guess you missed those "real world" tests (like encoding 4K video) where the Note 8 gets slaughtered by the iPhone X. Even the iPhone 7 was faster.

 

[haruhiko]

Yeah.. new phone is faster than a 4 month old phone...wait... is it the other way round?

 

[profets]

Weren’t they saying even the galaxy S8 beats the S9?

Lol whatever the case, doesn’t seem too impressive on Samsung’s part.

 

[ikir]

The correct titles should be:
iPhone SE and iPhone 7 beats S9 in many benchmarks

 

[44267547]

The average consumer likely doesn’t care about speed tests or even take full advantage of what their smart phones offer in terms of performance anyways. However, In terms of performance anecdotally, the “A” Series Chips is impressive with the iPhone.

 

[Eduardo75890]

iOS 11 is a bug. For some reason when I turn off Wi-Fi or Bluetooth in my control panel, it’s still ”on.”

It turns off temporally and turns on later like in a few minutes or so its normal

 

[BornAgainApple]

I opted for a ‘new’ year old iPhone 7 as the benchmarks with the 8/X were fairly close, and I saved some money in the process. Wasn’t too concerned with the wireless charging either. Very happy with my decision. Blows the doors off my work S7 Edge too.

 

[ErikGrim]

Android hardware perpetually two generations behind Apple in performance. This is really embarrassing.

 

[Gen0]

Why is this a surprise?

Exynos 9810 is terrible. The 9810 obviously throttles heavily. With a single core score that high (and Samsung really did make a big leap in single core performance - their highest in history), the multicore score should be much much higher. But it isn't higher because they obviously can't run all 4 cores wide open. Should have just put 2 fast cores in and let them run full speed instead of 4 throttled cores.

Firstly, the idea that Apple's CPU has 2 fast cores is technically a misnomer. It has 2 'dual issue' cores, which anyone else would call 4 cores in 2 modules because each core processes 2 separate streams at the same time in parallel, just shares some bits like power scaling.

Secondly, if the Exynos is running hot, it's doing so due to the design of the rest of the phone, not the Exynos. Samsungs thermal solution likely sucks, but for all we know that could mean that the 9810 is much faster in a different handset.

Guess you missed those "real world" tests (like encoding 4K video) where the Note 8 gets slaughtered by the iPhone X. Even the iPhone 7 was faster.

That's not a CPU task. The iPhone should win because it has a vector co-processor but that's not the CPU.