Antutu USA USER have 1527 with Spica

[sweetromain]

Hello,
I'm french and I would like to know "how this user did this score ?" , because me I've only 980, so I'm doing a comparative table :

HIM:
Memory: 170
CPU Integer: 381
CPU Float: 265
Graphic 2D: 226
Graphic 3D: 84
Database IO: 210
SD Card write: 42
SD Card read: 149
Android Version: 2.2.1

ME:
Memory: 111
CPU Integer: 243
CPU Float: 169
Graphic 2D: 146
Graphic 3D: 51
Database IO: 155
SD Card write: 20
SD Card read: 85
Android Version: 2.2.2

Please I would like some explanation

[KOala888]

maybe kernel without rtc sync?

Sent from my GT-I5700 using Tapatalk

[sweetromain]

the rtc sync takes so much resource it?

[uk_bm]

I don't think that's the point... If the clock isn't synchronised then scores can be artificially high as the benchmark action is measured against the time taken to complete.

[zenofilter]

Hello,
I'm french and I would like to know "how this user did this score ?" , because me I've only 980, so I'm doing a comparative table :

HIM:
Memory: 170
CPU Integer: 381
CPU Float: 265
Graphic 2D: 226
Graphic 3D: 84
Database IO: 210
SD Card write: 42
SD Card read: 149
Android Version: 2.2.1

ME:
Memory: 111
CPU Integer: 243
CPU Float: 169
Graphic 2D: 146
Graphic 3D: 51
Database IO: 155
SD Card write: 20
SD Card read: 85
Android Version: 2.2.2

Please I would like some explanation

+1,
Me too...

[podtalje]

It was already answered that the reason is in the rtc sync.

To explain it in more plain words, there were some versions of kernel where the clock was running too slow.
When kernel measured 8 seconds, in fact it was 12 seconds that passed.

The result of this was that the bencmarks scores were up to 50% higher. But as you can see the results are just
the result of wrong clock timing and in real world situations doesn't mean a thing.

[sweetromain]

Thanks podtalje for your explanation, I think to understand for CPU score but I've don't understand the score for graphic because the sync rtc for just the CPU side, no ?

[podtalje]

It is not just for CPU, it is general for the clock.

For example benchmarking program renders a 3D sequence as fast as it can and for that it requires let say 10 seconds.
From that it calculates the score. If it would render the sequence in 8 seconds then the score would be 20% higher.

But when rtc sync is off the benchmark tool thinks that the sequence finished in 8 seconds when in fact if you would measure it with you clock it would be 12 seconds.
So benchmark will show 50% better score when in fact this is not true.

My suggestion is to stop worrying about this benchmarks.

[sweetromain]

Okay thanks podtalje

[KOala888]

he already explained everything very clearly
that's how some time i got 1260 total in quadrant and permanent 30 fps in 2d quadrant