Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Wha-? "Larger performance advantage than indicated by the benchmark"? Keep dreaming...

No, he's actually right (for the majority of apps).

For example, let's say you have two processors, both of which score 1000 on Geekbench, one with 4 cores, the other with two.

That puts their per-processor Geekbench score at approximately 250/core and 500/core. So, an app, which runs on a thread will run twice as fast on the dual-core chip than it will on the quad-core chip. (That's rough, because not all of the sub-benchmarks scale the same way between processors, but the base reasoning is sound.)

For an app which uses all available cores equally, the performance would be roughly equal on the two processors. If any given thread bogs down a core, it will slow down the overall performance of the application. Having faster individual cores helps mitigate this. Having more cores helps spread the workload of threads more evenly to minimize the odds that a particular set of threads will do so. It's a tradeoff, but for single threaded (or even lightly threaded) apps, faster cores are better than more cores.
 
If this is what Cortex A15 brings, 2013 is going to be a very interesting year for Smartphones. We're reaching the point where the Motorola Atrix concept might be more than just a gimmick and might just become a true reality.

iPhone that docks into a laptop shell and becomes a full OS X system ? Docks into your TV to serve as an Apple TV ?

I can't wait until I have 1 and only 1 computing device that fits into my pocket, yet can be used on my desk to perform all my tasks while connected to a keyboard/mouse/monitor or then transform into a HTPC type device for my Home Theater, etc..

I haven't read every single post of yours as I usually just skip over them. They [your posts] usually come across as Anti-Apple banter, but this has to be one of the smartest posts I've seen here in a while. Good call, I could totally see this happening. This could be where Apple is going with iOS/OSX being more closely integrated, iCloud integration, using AirPlay possibly, the new connector... this "one device" thing sounds incredibly intriguing.
 
If you look at pure CPU scores the A6 isnt that fast. It gets lots of points for vastly improved memory bandwidth.

I wouldnt put to much faith in the CPU power its self until more tests have been run. Its funny how one leaked bench has the iSheep proclaiming the second coming of christ lands on friday.

dual core 1ghz vs quad core 1.4 ghz in a synthetic benchmark?

am i dreaming or is this discussion still happening?
 
Funny.

No benchmark on Galaxy SIII US and JP versions, which also uses a custom designed dual-core Cortex A15 SoC.

And no benchmark on LG Optimus G, which uses a custom designed quad-core Cortex A15 SoC.
 
Last edited:
No, he's actually right (for the majority of apps).

For example, let's say you have two processors, both of which score 1000 on Geekbench, one with 4 cores, the other with two.

That puts their per-processor Geekbench score at approximately 250/core and 500/core. So, an app, which runs on a thread will run twice as fast on the dual-core chip than it will on the quad-core chip. (That's rough, because not all of the sub-benchmarks scale the same way between processors, but the base reasoning is sound.)

For an app which uses all available cores equally, the performance would be roughly equal on the two processors. If any given thread bogs down a core, it will slow down the overall performance of the application. Having faster individual cores helps mitigate this. Having more cores helps spread the workload of threads more evenly to minimize the odds that a particular set of threads will do so. It's a tradeoff, but for single threaded (or even lightly threaded) apps, faster cores are better than more cores.
I've already explained that this analogy is invalid and is based on a flawed simplification of how processors work. If a two-core processor achieves a score 1000, you simply cannot assume that one core would achieve half of that score - it does not work that way. You don't double the processing power by adding another core. This score is given to the whole architecture, and it depends on my factors. If you want to learn more, read this white paper.
 
You dont have to do that on Android either. But at least you can background applications you want to continue running and no frozen into RAM. iOS doesnt have the option

Entirely true, but in practical use the difference between the two approaches is negligible. (And don't get me started on the various android phone commercials which show off multi-tasking by showing someone switching between two different movies and 3 different games *WHILE THEY'RE ALL STILL RUNNING*. Nobody leaves a movie playing on their phone while they start watching another one, while they've got three un-paused games they're busy playing despite the fact that they can't see what's on the screen. Unless the movies & games auto-pause, in which case there's no difference at all at the user level between the two methods.)

iOS does *real* multi-tasking, too. It has since the beginning (otherwise you couldn't have received a call without being in the phone app). They just don't expose public APIs to allow 3rd party developers to create arbitrary background processes that keep running when the app isn't active.
 
Wow.

Two cores running @1.02 GHz each outperforming S3. Engineering people, engineering.

This is all less about specs and more about engineering.

What "engineering" and where is it "outperforming S3"??
Galaxy S3
http://browser.primatelabs.com/geekbench2/1037491 2433
http://browser.primatelabs.com/geekbench2/1039203 2399
http://browser.primatelabs.com/geekbench2/1037296 2293
http://browser.primatelabs.com/geekbench2/1038481 2087
http://browser.primatelabs.com/geekbench2/1036620 1957 (2 cores)

iPhone 5
http://browser.primatelabs.com/geekbench2/1030202 1601

Dreamers...
 
Last edited:
I've already explained that this analogy is invalid and is based on a flawed simplification of how processors work. If a two-core processor achieves a score 1000, you simply cannot assume that one core would achieve half of that score - it does not work that way. You don't double the processing power by adding another core. This score is given to the whole architecture, and it depends on my factors. If you want to learn more, read this white paper.

I understand that. In fact, I pointed out that fact. I said:
(That's rough, because not all of the sub-benchmarks scale the same way between processors, but the base reasoning is sound.)
It was actually included in the text you quoted in your response to my post.

However, barring some *really* weird processor design decisions, a quad-core processor which scores the same as a dual-core processor on a balanced, multi-core benchmark (like Geekbench) will have less powerful individual cores than the dual-core processor with the same score.

In fact, the overhead related to core-count will mean that the theoretical quad core CPU I discussed would have *slightly* better than 1/2 the per-core performance of the dual-core.

Having a ratio of 1.9:1 rather than 2:1 doesn't make the underlying logic and reasoning invalid. Especially since I *acknowledged* that I was using rough numbers in my post.
 
Truth be told, I really don't care what GB says.

I will never buy another android phone. The GS4 could score a million points, but you'd still be stuck with android, so what's the point.
 

To be fair we have to compare without resorting to user overclocking and optimizing the OS. You have to take what is given by the factory and use that. True it is much easier to install a custom ROM on an Android device than jail breaking an iOS device.
 
What "engineering" and were is it "outperforming S3"??

In the beginning of the thread? Your examples was uploaded today, with the top performers using an over clocked cpu. Why are they magically higher then the previous results uploaded to geekbench.
 
I understand that. In fact, I pointed out that fact. I said:

It was actually included in the text you quoted in your response to my post.

However, barring some *really* weird processor design decisions, a quad-core processor which scores the same as a dual-core processor on a balanced, multi-core benchmark (like Geekbench) will have less powerful individual cores than the dual-core processor with the same score.

In fact, the overhead related to core-count will mean that the theoretical quad core CPU I discussed would have *slightly* better than 1/2 the per-core performance of the dual-core.

Having a ratio of 1.9:1 rather than 2:1 doesn't make the underlying logic and reasoning invalid. Especially since I *acknowledged* that I was using rough numbers in my post.

*Sigh* Have you even read that paper? In the example given in the paper, two cores of the same frequency have roughly a 25% performance increase, which is way lower than your 1.9:1 ratio.
 


Funny how all these outperforming scores showed up today. Couldn't possibly be android fans doing something to cheat the system.
 
Guys,

The banchmarks should be checked out precisely at the web site before making judements:

Iphone 5,2 Geekbench Score: 1601
(There is only 1 sample)
http://browser.primatelabs.com/geekbench2/1030202

Samsung Galaxy S III Geekbench Score: 2052
http://browser.primatelabs.com/geekbench2/1036389

There are hundreds of samples for SGS3, those are between 1800-2000, when all the 4 cores are on:
http://browser.primatelabs.com/geekbench2/search?utf8=✓&q=samsung+galaxy+S+III

The SGS3 has a Power Saving mode as well when the performance is limited.

Let's wait for some independent and objective test results. I am really interested how is the new iOS 6 on iP5 is against the newest Android (4.1.1 JB) on SGS3 in terms of speed.
Still, it is pretty powerful for a dual-core 1 GHz processor.
 
There are many, many tests that were conducted months ago with higher scores than 1601 on stock S3s (without overclocking).

Maybe, the interesting question is how much of that is utilized in everyday use, and how much of it is sitting on idle. ie real world use.
 
really?

Multiple cores and threads matter more on Andriod cause they have true multitasking. Apple doesn't and can get away with a much smaller battery and less cores.

In any case,
How much power are you guys using on your phones? What apps push your phone? Games? A graphics benchmark would be better for that. Show me real like streaming and 1080P decoding tests and we can compare properly.
 
Multiple cores and threads matter more on Andriod cause they have true multitasking. Apple doesn't and can get away with a much smaller battery and less cores.

I doubt that matters much, there are a lot of active processes at any given moment and it doesn't address multithreading within one process. Multitasking only concerns user applications. On my mac I have 51 processes and 259 threads active at the moment for example.
 
Let me see if I get this straight. The iPhone 5 is pretty much as fast as the SGSIII, but people are still whining?

I think people are forgetting that everything that made the previous iPhone's great is also valid for the new Iphone5.

Could someone enlighten me as to what feature the SGSIII has that makes it an iPhone-killer? Any useful feature it has that doesn't exist on the iPhone?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.