I wonder with geek bench...

TechGod

macrumors 68040
Original poster
Feb 25, 2014
3,177
914
New Zealand
So we all know geek bench tests single core and multicore but what happens with big.LITTLE chipsets like the exynos? Does it only bench the more powerful cluster and disable the other one?

If so, its kinda stupid for Apple fanboys to claim that the A9 dual core keeps up with an octacore when it's just a quad core.

I mean I have problems with companies calling their big.LITTLE configs octacores and stuff as well but that's for another thread...
 

Truefan31

macrumors 68040
Aug 25, 2012
3,435
722
So we all know geek bench tests single core and multicore but what happens with big.LITTLE chipsets like the exynos? Does it only bench the more powerful cluster and disable the other one?

If so, its kinda stupid for Apple fanboys to claim that the A9 dual core keeps up with an octacore when it's just a quad core.

I mean I have problems with companies calling their big.LITTLE configs octacores and stuff as well but that's for another thread...
I don't think it works like that. Hence why the a9 is impressive vs the competition. Have you looked at all the benchmarks? The iPhone 6s/6s plus is basically being compared to a surface pro 3.

The question is how a dual core is on par if not beating most if not all quad/hex/octal cores.
 

TechGod

macrumors 68040
Original poster
Feb 25, 2014
3,177
914
New Zealand
I don't think it works like that. Hence why the a9 is impressive vs the competition. Have you looked at all the benchmarks? The iPhone 6s/6s plus is basically being compared to a surface pro 3.

The question is how a dual core is on par if not beating most if not all quad/hex/octal cores.
Yes I have...that's not my point. My point is that its not A dual core A9 vs an octacore, its a dual core A9 vs the more powerful of the two clusters that make up the big.LITTLE octacore processors.
 

Truefan31

macrumors 68040
Aug 25, 2012
3,435
722
Yes I have...that's not my point. My point is that its not A dual core A9 vs an octacore, its a dual core A9 vs the more powerful of the two clusters that make up the big.LITTLE octacore processors.
So for an octacore even if that were the case it'd be the dual a9 vs quad core? Again I don't think geek bench works like that.
 

TechGod

macrumors 68040
Original poster
Feb 25, 2014
3,177
914
New Zealand
So for an octacore even if that were the case it'd be the dual a9 vs quad core? Again I don't think geek bench works like that.
So you're saying that Geekbench and ONLY geekbench enables both CPU clusters whereas Samsung or Qualcomm don't otherwise allow both clusters to be online at the same time?
 

Truefan31

macrumors 68040
Aug 25, 2012
3,435
722
So you're saying that Geekbench and ONLY geekbench enables both CPU clusters whereas Samsung or Qualcomm don't otherwise allow both clusters to be online at the same time?
No I'm saying that all cores are being tested whether it's 2 or 10. The score is a mean of the arithmetic tests performed.

I'm not trying to be a fanboy but what Apple has done with the a9 and the new faster storage is impressive and most didn't expect this kind of power. It's a benefit no doubt but again Apple didn't focus on this because they care about the user experience first and foremost. I'm sure we'll see great tech from Samsung and others. I'm sure the next exynos chip will be impressive too. And they'll be plenty of people here touting its power. But at least recognize that Apple basically surprised people again who thought they wouldn't/couldn't add the power they did. A dual core a9 with 2 gb ram is at the least on par/beating everything out right now. It shows how optimized Apple makes things.
 
Last edited:

MRU

Suspended
Aug 23, 2005
25,318
8,810
Other
Benchmarking apps on Android are taking into account all cores and YES all 8 cores or 6 cores in case of the 808 chipset are running together and active when these tests are taking place.
 

Truefan31

macrumors 68040
Aug 25, 2012
3,435
722
Benchmarking apps on Android are taking into account all cores and YES all 8 cores or 6 cores in case of the 808 chipset are running together and active when these tests are taking place.
Thank you. I thought that too.
 
  • Like
Reactions: MRU

Tsepz

macrumors 68040
Jan 24, 2013
3,622
3,179
Johannesburg, South Africa
No I'm saying that all cores are being tested whether it's 2 or 10. The score is a mean of the arithmetic tests performed.

I'm not trying to be a fanboy but what Apple has done with the a9 and the new faster storage is impressive and most didn't expect this kind of power. It's a benefit no doubt but again Apple didn't focus on this because they care about the user experience first and foremost. I'm sure we'll see great tech from Samsung and others. I'm sure the next exynos chip will be impressive too. And they'll be plenty of people here touting its power. But at least recognize that Apple basically surprised people again who thought they wouldn't/couldn't add the power they did. A dual core a9 with 2 gb ram is at the least on par/beating everything out right now. It shows how optimized Apple makes things.
Don't know who you are trying to convince...
I had no doubt that the apple A9 would do well as it only has to run a HD1080P display in the iPhone6S Plus at best and an OS that is much lighter than the competition with less overhead.

I expect it to get better results. I will be very concerned the day an iPhone gets worse results than the rest. In actual fact, I expect iPhones to do better than most Androids a generation newer, as they tend to have a lot more to run.
 

MRU

Suspended
Aug 23, 2005
25,318
8,810
Other
Surely the phone would pick up that CPU power is needed and only power up the powerful cores?
The new octacore chips run all 8 cores together when needed.

Only first gen big.little architecture ran with either 4 cores only. But for a long time now all octa core chipsets and hexacore have been able to run with all cores when needed.

This feels like a thread from late 2013 than 2015. Things have moved on massively and the original synopsis outlined in the opening post is now inaccurate and wrong.
 
  • Like
Reactions: TechGod and Oohara

Truefan31

macrumors 68040
Aug 25, 2012
3,435
722
Don't know who you are trying to convince...
I had no doubt that the apple A9 would do well as it only has to run a HD1080P display in the iPhone6S Plus at best and an OS that is much lighter than the competition with less overhead.

I expect it to get better results. I will be very concerned the day an iPhone gets worse results than the rest. In actual fact, I expect iPhones to do better than most Androids a generation newer, as they tend to have a lot more to run.
By all means then it's not intended for you. But many didn't anticipate 50% jump in single core and up to 80-90% in gpu. Also the big jump in storage speeds due to basically putting a new controller that mimics the MacBook ssd.
 

Tsepz

macrumors 68040
Jan 24, 2013
3,622
3,179
Johannesburg, South Africa
By all means then it's not intended for you. But many didn't anticipate 50% jump in single core and up to 80-90% in gpu. Also the big jump in storage speeds due to basically putting a new controller that mimics the MacBook ssd.
All on the same resolution as last year, which is much lower than 2014 Androids.

A similar thing happened to the Xperia Z5 Compact, running the same ol' HD720P res in its case as the old Z1 Compact, with newer CPU, GPU and RAM.
 

TommyA6

macrumors 65816
May 15, 2013
1,056
516
All on the same resolution as last year, which is much lower than 2014 Androids.

A similar thing happened to the Xperia Z5 Compact, running the same ol' HD720P res in its case as the old Z1 Compact, with newer CPU, GPU and RAM.
Resolution has no effect on Geekbench scores.
If you care about resolution that much you can compare Android phones to 6s Plus, which renders everything at 2208x1242 and then downscales to 1080p, which is more taxing.
 

Tsepz

macrumors 68040
Jan 24, 2013
3,622
3,179
Johannesburg, South Africa
Resolution has no effect on Geekbench scores.
If you care about resolution that much you can compare Android phones to 6s Plus, which renders everything at 2208x1242 and then downscales to 1080p, which is more taxing.
Interesting, thanks for clarifying.
Does OS also not have an effect?

Is it really more taxing to downsample a low res to a lower res vs. running 2K??? I have major doubts about that.
 
Last edited:

Tsepz

macrumors 68040
Jan 24, 2013
3,622
3,179
Johannesburg, South Africa

TommyA6

macrumors 65816
May 15, 2013
1,056
516
How is it that OS doesn't have an effect? Does Geekbench close down every single app/process and have both devices running clean? I just ran it and it seems to be allowing my phone to go on as if any other app is running.
I don't own an Android device, so I don't know for sure, but you can try running it with some background process running and see whether it makes a noticeable difference in score. All the scores I've seen online point to it not having a huge effect.
 

Tsepz

macrumors 68040
Jan 24, 2013
3,622
3,179
Johannesburg, South Africa
I don't own an Android device, so I don't know for sure, but you can try running it with some background process running and see whether it makes a noticeable difference in score. All the scores I've seen online point to it not having a huge effect.
Will try do some digging this weekend.

I always have trouble comparing iOS and Android benchmarks, as these two OSs are so different.

When I got my iPad 4 in 2012 I also had a Galaxy Tab10.1, and comparing their benchmark performance and the way each shows Battery stats was a lost cause, so I just went with the one that performed best with what I did, and the iPad whiped the floor with that 10.1, early days of Android tablets were shocking.
 

Oohara

macrumors 68030
Jun 28, 2012
2,754
1,549
So we all know geek bench tests single core and multicore but what happens with big.LITTLE chipsets like the exynos? Does it only bench the more powerful cluster and disable the other one?

If so, its kinda stupid for Apple fanboys to claim that the A9 dual core keeps up with an octacore when it's just a quad core.

I mean I have problems with companies calling their big.LITTLE configs octacores and stuff as well but that's for another thread...
Well Christmas time is right around the corner and that's when we here in Germany like to eat a lot of

 

TechGod

macrumors 68040
Original poster
Feb 25, 2014
3,177
914
New Zealand
The new octacore chips run all 8 cores together when needed.

Only first gen big.little architecture ran with either 4 cores only. But for a long time now all octa core chipsets and hexacore have been able to run with all cores when needed.

This feels like a thread from late 2013 than 2015. Things have moved on massively and the original synopsis outlined in the opening post is now inaccurate and wrong.
Alright thanks for clearing this us. How is temperature manger? Surely these things get hot?
 

TechGod

macrumors 68040
Original poster
Feb 25, 2014
3,177
914
New Zealand
The new octacore chips run all 8 cores together when needed.

Only first gen big.little architecture ran with either 4 cores only. But for a long time now all octa core chipsets and hexacore have been able to run with all cores when needed.

This feels like a thread from late 2013 than 2015. Things have moved on massively and the original synopsis outlined in the opening post is now inaccurate and wrong.
In GeekBench, are the single core scores taken from the higher performing cluster? I'd assume so.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.