Galaxy S7 fails to impress in NAND read/write speeds; iPhone 6s still twice as fast or even faster

[vertsix]

By far one of the most significant improvements to the iPhone 6s was Apple's decision to use the NVMe controller from the Macbooks onto the iPhone to accelerate the flash performance, and I was expecting the Galaxy S7 to catch up (or surpass) the iPhone 6s in this aspect, especially because the Galaxy S6 had the fastest flash when it was released.

Unfortunately, Samsung disappointed me with very minor improvements to the flash performance of the Galaxy S7.

And results are even more shocking with write speeds:

It is important to mention that the faster flash and the controller of the iPhone 6s are what cause the more significant speed improvements, in collaboration with the ridiculously fast A9 CPU and faster RAM. So, theoretically, the iPhone 6s should still be faster when loading things from flash (such as applications).

What do you guys think? Is this important to you? Let me know.

SOURCE: http://www.anandtech.com/show/10120/the-samsung-galaxy-s7-review/3

 

[LIVEFRMNYC]

I'm lost. What's the difference between the benchmark you posted and this?

 

[mclld]

I'm sure you were really upset by it

 

[MRU]

And to the end user .. It makes no difference. Its a mobile phone, its not edting photoshop or final cut pro or heavy data processing ...

 

[Mr. Buzzcut]

Funny that. Touting Apple's specs. When they're better. Other times, UX trumps specs.

 

[grkm3]

Lol biased anandtech not using a memory app that uses the full potential of the memory controller

The gs7 has insane fast memory

Here is mine! And it's double what "cough" anandtech got

I mean cmon are you surprised a site that the founder and owner that still hangs around there that works for Apple posted bs memory charts to make the gs7 look bad?

Lol they found the one carrier in the whole world that don't use the optimsed Samsung web browser to also post its web browsing results to show inferior numbers.ever single gs7 on the planet comes with Samsung browser and are we surprised anandtech found the one and only phone that uses the unoptimized chrome to post numbers against the iPhone?

 

[Savor]

Like it even matters when I could still move files with a file explorer app and SELECT ALL option to mass delete media files on Android. You need iTunes which takes time just to boot up the computer to move or mass delete any files on iOS.

 

[Michael Goff]

The only importance of benchmarks on hardware these days is to try to get a grasp of how important different companies find different specs.

 

[vertsix]

And to the end user .. It makes no difference. Its a mobile phone, its not edting photoshop or final cut pro or heavy data processing ...

Have you even seen the performance difference between the 6 and the 6s?

Surely you haven't.

 

[McCool71]

Unfortunately, Samsung disappointed me with very minor improvements to the flash performance of the Galaxy S7.

Doesn't matter much - the S6 was already one of the fastest Android devices when it comes to memory, and they improved on that.

Memory speeds at this rate for a cell phone is purely a theoretical number with today's tech. It is measurable, but does not make any difference at all in the daily use of the device. The read speed alone is already far beyond what you would need to play a handful of 60fps 4K videos on your phone simultaneously...

 

[grkm3]

Have you even seen the performance difference between the 6 and the 6s?

Surely you haven't.
[doublepost=1457758993][/doublepost]

I don't know, man. That app looks pretty sketch and like it was made in 2010.

And I don't see how you say Anandtech is biased. Mind giving me some legitimate proof to prove your argument?

Anandtech is using androbench to get its numbers.if you dig deep you will see that they used a version that don't support multi io threads that the gs6 and gs7 are using.Samsung is using ufs 2.0 ssd flash in there phones and it's the fastest storage out.

Now they also used a newer version but purposely went into setting and changed the file size to a tiny cluster where the ufs flash is weak at.who the hell moves clusters at 4kb at a time? The standard setting is supposed to be 64mb and the gs6 and gs7 are well North of 420mb/sec and that my friend is why I never read any crap from that biased site.

Here this is taken from there gs6 review and as of right now the fastest phone tested in androbench (same app anandtech used)is 481mb per second and that score was just posted by a galaxy s7 edge!

 

[MRU]

Have you even seen the performance difference between the 6 and the 6s?

Surely you haven't.

No... I still use a Nokia 3310.


I think we can put most of the boost in speed & performance on the 6s over to the iphone 6, down to the much faster A9 and extra 1gb ram moreover than faster nand speed.

Have YOU seen slow app openeing speeds on the slower nand based Galaxy S7 ? Because no-body else has.

This is a nonsense thread, making mountains out of molehills and ultimately being a bit of a drama queen in one of the worst farces ever written ....

 

[John Mcgregor]

And to the end user .. It makes no difference. Its a mobile phone, its not edting photoshop or final cut pro or heavy data processing ...

Doesn't matter, but every time someone brings the specs up.

 

[vertsix]

I'm lost. What's the difference between the benchmark you posted and this?


View attachment 620749

Well, as some people are pointing out in this thread, Anandtech is supposedly being somewhat misleading.

So gg Samsung.
[doublepost=1457769905][/doublepost]

Doesn't matter, but every time someone brings the specs up.

True. Ultimately benchmarks and specs don't matter in some cases because other things weigh in, such as the software itself. In this case, Touchwiz.

Specs suddenly become less impressive because of the horrendous unoptimization of the software for real world use.

 

[grkm3]

I still see people mentioning single core performance but dissing multi core making up excuses for not needed extra cores etc.

My gs7 gets almost 2400 single core and 5500 muliti.Samsung has bridged the gap in single core and still dominates in multi

 

[mrex]

Anandtech is supposedly being somewhat misleading.

and no one follows them anymore... they lost their crediability since they started to lick apples ass and anand went to work for apple. since then they have told us only halve of the truth and telling nice stories about apple but articles from other manufacturers are always somewhat misdleading.

 

[I7guy]

and no one follows them anymore... they lost their crediability since they started to lick apples ass and anand went to work for apple. since then they have told us only halve of the truth and telling nice stories about apple but articles from other manufacturers are always somewhat misdleading.

Well no hyperbole in that statement. They may have lost their credibility with you, but I don't see them folding their tent and going home.

 

[gotluck]

Looks like gkrm3 is spot on about the testing method, props to you

I still think running the anandtech results by the android guys gives the best benchmark data. What other websites are you guys looking at for data driven info?

For example, it sounds like there is not a test equivalent to androidbench4.x that runs on ios. I would imagine that iPhone 6s storage uses more than a single io thread too eh?

 

[epicrayban]

Way to lose creditbility , anandtech.

 

[grkm3]

Funny how they found the one version of the gs7 that don't use Samsung's optimsed web browser to do there web performance tests too right?

 

[Michael Goff]

I still see people mentioning single core performance but dissing multi core making up excuses for not needed extra cores etc.

My gs7 gets almost 2400 single core and 5500 muliti.Samsung has bridged the gap in single core and still dominates in multi

View attachment 620821

Isn't your phone running an 820? I think Qualcomm did the matching. I think the Exynos has a slightly lower single and higher multi.

 

[Lloydbm41]

Sequential read/write speeds are nice to show theoretical maximum speeds, but random read/write speeds are what the average person is likely to see in everyday use. Ignore sequential and look at random.

P.S. I stopped believing Anandtech for anything.

 

[Fernandez21]

Funny how they found the one version of the gs7 that don't use Samsung's optimsed web browser to do there web performance tests too right?

If you look at all the reviews, most have the Verizon model, so this is probably just the phone Samsung sent them.

 

[vertsix]

I still see people mentioning single core performance but dissing multi core making up excuses for not needed extra cores etc.

My gs7 gets almost 2400 single core and 5500 muliti.Samsung has bridged the gap in single core and still dominates in multi

View attachment 620821

You have to consider it has twice as many cores as the A9, therefore making it less impressive.

Very solid single core score though.

But in the end it has Touchwiz so that result is pretty irrelevant in the real world because of the lack of optimization. If it had stock or Cyanogenmod, it'd be a different story.

 

[mi7chy]

iOS on iPhone is like MS-DOS in that it can't take advantage of multi-cores since it's limited to one foreground task and no background tasks. So, it's better to equip it with a faster single or fewer cores. Android, on the other hand, is like Windows 10 in comparison and benefits from multi-cores.

Here's Android running a game, Twitch.tv and a movie in the foreground with other background tasks.

As for storage, random performance is more important for real world performance and not like it matters on iOS anyway since you can't do things like connect it to a computer for use as direct storage.