Why are these scores so low?

[jagrave]

http://www.anandtech.com/show/7903/samsung-galaxy-s-5-review/7

I think some of these have been posted before, but I just saw them again. I'm just curious as to why a lot of the S5 scores are so low compared to the others that came out last year? Is the random read/write low performance actually noticeable? And could that be just because of their review unit? I remember someone saying that the m8 "cheats" some, but would that account for all the higher scores? And I realize there is probably some thermal throttling, but would that make that big of a difference?

 

[plasmastate]

I am really concerned about the random read write performance of the nand used in s5. It shouldn't have been this way. And no, throttling doesn't affect the nand performance.

Sent from my GT-I9500 using XDA Premium 4 mobile app

 

[jagrave]

I am really concerned about the random read write performance of the nand used in s5. It shouldn't have been this way. And no, throttling doesn't affect the nand performance.

Sent from my GT-I9500 using XDA Premium 4 mobile app

how noticeable do you think it'll be?

 

[Hellscythe]

how noticeable do you think it'll be?

it looks pretty noticeable as the range between the S5 and top performers are pretty significant.
I want to believe that this is a software fault, but i have a feeling it is not..

 

[simacca]

Surely if the phone runs smoothly in real world usage then these numbers should not really matter??


Sent from my iPad using Tapatalk

 

[CLARiiON]

That's a concern for me too.

Normally Samsung has very good NAND in each devices, GN3 is a recent example.

I am assuming, we might have different NAND from other vendors or from Samsung itself to meet the initial requirements. It's not the 1st time that this has happened. Need to wait for other reviews as well.

---------- Post added at 04:12 PM ---------- Previous post was at 04:09 PM ----------

BTW, never look at Anandtech's javascript benchmarks.

I never understood they always use Chrome for Android benchmarks which everyone knows sucks. The benchmark in native browser will blow away the competition.

 

[hung2900]

That's a concern for me too.

Normally Samsung has very good NAND in each devices, GN3 is a recent example.

I am assuming, we might have different NAND from other vendors or from Samsung itself to meet the initial requirements. It's not the 1st time that this has happened. Need to wait for other reviews as well.

---------- Post added at 04:12 PM ---------- Previous post was at 04:09 PM ----------

BTW, never look at Anandtech's javascript benchmarks.

I never understood they always use Chrome for Android benchmarks which everyone knows sucks. The benchmark in native browser will blow away the competition.

The funny thing here is that they considered the Sunspider as an CPU benchmark :rofl:

 

[CLARiiON]

@leoaudio13 can you run these benchmarks on your device and post the full results?

RL Benchmark: SQLite
Androbench

cheers mate!

---------- Post added at 05:36 PM ---------- Previous post was at 05:36 PM ----------

The funny thing here is that they considered the Sunspider as an CPU benchmark :rofl:

yep, just weird..

 

[leoaudio13]

@leoaudio13 can you run these benchmarks on your device and post the full results?

RL Benchmark: SQLite
Androbench

cheers mate!

---------- Post added at 05:36 PM ---------- Previous post was at 05:36 PM ----------



yep, just weird..

Will do now. Expect it in 15 mins

Sent from my SM-G900F using Tapatalk

 

[leoaudio13]

@leoaudio13 can you run these benchmarks on your device and post the full results?

RL Benchmark: SQLite
Androbench

cheers mate!

---------- Post added at 05:36 PM ---------- Previous post was at 05:36 PM ----------



yep, just weird..

Here u are :


Sent from my SM-G900F using Tapatalk

 

[jagrave]

Thank you Leo. So, do the low scores effect actual performance? I ran the test on my Rezound, and unless I did something wrong, the reads and writes were much faster on my phone. They were so much faster, that I do think it didn't work right. I even changed target path from /data to /sdcard and still fast. Hmm mm.....

Sent from my ADR6425LVW using Tapatalk

 

[leoaudio13]

Thank you Leo. So, do the low scores effect actual performance? I ran the test on my Rezound, and unless I did something wrong, the reads and writes were much faster on my phone. They were so much faster, that I do think it didn't work right. I even changed target path from /data to /sdcard and still fast. Hmm mm.....

Sent from my ADR6425LVW using Tapatalk

Not at all.... i found its quite fast doing the test. But someone please tell me the scores r ok or bad ? Cos i just quickly did the test then post the result

Sent from my SM-G900F using Tapatalk

---------- Post added at 11:21 PM ---------- Previous post was at 11:20 PM ----------

The funny thing here is that they considered the Sunspider as an CPU benchmark :rofl:

Wtf lol

Sent from my SM-G900F using Tapatalk

---------- Post added at 11:25 PM ---------- Previous post was at 11:21 PM ----------

Thank you Leo. So, do the low scores effect actual performance? I ran the test on my Rezound, and unless I did something wrong, the reads and writes were much faster on my phone. They were so much faster, that I do think it didn't work right. I even changed target path from /data to /sdcard and still fast. Hmm mm.....

Sent from my ADR6425LVW using Tapatalk

I think u shouldnt worry much about the actual performance. The fone is very snappie. Im very very picky about speed but this is the very 1st android fone from Sammy that I didnt have a thought 'bout selling it in the next 2 weeks for another fone to mess with. I want to keep it. And that suprised me,seriously. This and the LG G2. I also love my 5s but its screen is just too small now

Sent from my SM-G900F using Tapatalk

---------- Post added at 11:26 PM ---------- Previous post was at 11:25 PM ----------

That's a concern for me too.

Normally Samsung has very good NAND in each devices, GN3 is a recent example.

I am assuming, we might have different NAND from other vendors or from Samsung itself to meet the initial requirements. It's not the 1st time that this has happened. Need to wait for other reviews as well.

---------- Post added at 04:12 PM ---------- Previous post was at 04:09 PM ----------

BTW, never look at Anandtech's javascript benchmarks.

I never understood they always use Chrome for Android benchmarks which everyone knows sucks. The benchmark in native browser will blow away the competition.

I thought Toshiba Nand was better in term of durability ?

Sent from my SM-G900F using Tapatalk

 

[CLARiiON]

Here u are :
View attachment 2678716View attachment 2678717View attachment 2678718View attachment 2678719View attachment 2678720

Sent from my SM-G900F using Tapatalk

:good:

I am not sure what's wrong here.

The part numbers mentioned in both chipworks and iFixit teardown are quite good.

Chipworks mentions NAND solution KLMBG4GEAC and iFixit mentions KLMAG2GEAC-B0. Both these look fairly good performing NAND from marketing materials.

Source pdf

 

[leoaudio13]

:good:

I am not sure what's wrong here.

The part numbers mentioned in both chipworks and iFixit teardown are quite good.

Chipworks mentions NAND solution KLMBG4GEAC and iFixit mentions KLMAG2GEAC-B0. Both these look fairly good performing NAND from marketing materials.

Source pdf

Maybe different batch has different Nand ?

Sent from my SM-G900F using Tapatalk

 

[CLARiiON]

Maybe different batch has different Nand ?

Sent from my SM-G900F using Tapatalk

quite possible.

 

[Hellscythe]

the 32gb S5 might have better performance than the 16gb one because that was the case with the Note 3.