Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Yes it does. Many of the API calls you perform are handled in different threads and even a mediocre programmer does not perform heavy lifting in the main thread to avoid UI stutter.

Wrong. Show a single API call (besides a call to a threading API such as pthreads or parallel programming / concurrency API such as GCD) where results are computed on a separate thread. The programmer manages this, not the API. I'll be waiting for your example.
 
For Comparison

My iPad 1 & iPhone 3Gs scored as follows (iOS 5.1.1 on both):

iPad 1,1 ~469
iPhone 3Gs ~285
 
Hows does this compare to the almighty S3?:rolleyes:

Sadly, not impressive:

Iphone 5 (1601 points):
http://browser.primatelabs.com/geekbench2/1030202

Galaxy S3 4-core, 1GB (Non-US version) (1847 points):
http://browser.primatelabs.com/geekbench2/998950

Galaxy S3 2-core, 2GB (US version) (1683 points):
http://browser.primatelabs.com/geekbench2/884649

The S3 scores (just like the iPhone 4S scores and no doubt the upcoming iPhone 5 scores when lots of people start running the test) is all over the place, a lot of people obviously run this tests with lots of stuff going on in the background that impacts the result.
 
Google is already preparing an update for all Android OSes that will increase the processor speed on any Android smartphone by at least 30%. Google knows that those Droidbois can't tolerate sucking the iPhone's dust. By next week, the iPhone 5's processor specs will be somewhere in the middle of the Android pack. Most of those Android processors have lots of headroom to be over-clocked just for this purpose. If they lose a couple of hours of battery time, so what. It's having the fastest processor that really matters.
 
The negligible speed difference does not magically turn the iPhone 5 into an innovative, visionary device. It still is a catch-up device that only brings minor improvements to iPhone customers.

Yeah it only doubles the performance of any other iOS device ever. I definitely consider that a minor improvement.
 
I think this pretty much settles it, then. Samsung is DEAD.

Android users, you may now STEP FORWARD and leave the MR premises forthwith.

This is an Apple forum, after all.

I bet you still live with your mother . Certain of it
 
I think this pretty much settles it, then. Samsung is DEAD.

Android users, you may now STEP FORWARD and leave the MR premises forthwith.

This is an Apple forum, after all.

If you didnt have any other phones, these scores would be useless. What would compare it to?

Of course, we know Apple wants *zero* competition.... maybe you guys actually bought into that mentality?
 
that is as fast or faster than my 2005 amd pc I was running(running 32bit mode)
holy ****
2-4k geekbench scores are going to be coming within the next year or two
crazy
 
Blown away? We're talking about a questionable 2,5% performance advantage, and there is already proof that some tests carried out on the S3 had a higher scores than the one in the table.

The main point is the quad-core vs dual core, 2 GB vs 1 GB here. It's like Samsung revving up a Lamborghini Aventador against an M3 in front of the traffic lights, but still losing the drag race.

Samsung chose to bloat about specs.. They'd better made sure their flagship can hold up a candle, which it can't atm :)
 
The APIs on iOS do not do not take advantage of multi-core CPUs automatically. Apple provides technologies to facilitate parallel programming such as GCD, sure, but the actual work is still up to the programmer. Parallel programming is inherently difficult. It is the job of the programmer to express the parallelism and concurrency using an appropriate technology -- a job which is time consuming, error prone, often provides diminishing returns, and requires skills that many programmers still don't have.

I think you haven't used GCD and blocks much. Blocks are dead easy. GCD makes running things on background threads _really_ easy. Graphics and animation, the two things that make your app look and feel smooth, are automatically multithreaded.
 
Wrong. Show a single API call (besides a call to a threading API such as pthreads or parallel programming / concurrency API such as GCD) where results are computed on a separate thread. The programmer manages this, not the API. I'll be waiting for your example.

Core Location? After you request to have access to that information you have delegate functions that respond to change in location, errors, etc.
 
Can't wait to get the iPhone 5 but the score on the S3 is a little low, just ran a benchmark on my quad core S3 and scored 1822. The iPhone 5 still has much better CPU though as it's only running in dual and not quad. If the S3 only had a dual core then it's score would be far lower.

 
uh oh..could it be? s3 geekbench 2 score is 1789. quick google search lists other scores that are higher than the 1560 score posted.

http://kaniskc.co.cc/post/25956766027/samsung-galaxy-s3-geekbench-review
Scores on the non-LTE, quad-core GS3 vary from 1400 to 1850 depending on the OS and settings. To get more than 1600 on the GS3, you have to disable power management and other background services. And as others have noted, Geekbench takes advantage of all four cores, whereas most applications can only take advantage of 1-2, so this score overstates actual performance.

The U.S. LTE GS3 (dual-core) scores up to 1450 or so with latest OS, but real world performance is fairly close to the non-LTE quad-core version, because, again, most apps use only 1-2 cores effectively.
 
This is faster than my 24GB 3.33GB 6-core MacPro? Holy *****.

Oops, off by a factor of 10.
 
Last edited:
As a developer in training I assure you that nothing iOS runs is on any virtual machine and there is no garbage collection. The os iOS is a lot smarter than android.

Objective-C runs on top of a virtual machine???

Whoever makes/made that claim clearly has no freaking clue what he/she is talking about.
 
Last edited:
You can't compare these results between platforms. Also geekbench is far from an accurate benchmark anyway. The dumbest thing I saw on that list is the nexus s scoring significantly lower than the droid incredible. Its a fact that the samsung hummingbird soc in the nexus s is much faster than the snapdragon s1 in the incredible. Also my nexus 7 scored 1800 and its not even overclocked. I noticed that the snapdragon s4 s3 isn't in the list either. I bet it performs better in this test than exynos version.
 
Can't wait to get the iPhone 5 but the score on the S3 is a little low, just ran a benchmark on my quad core S3 and scored 1822. The iPhone 5 still has much better CPU though as it's only running in dual and not quad. If the S3 only had a dual core then it's score would be far lower.

[url=http://img.photobucket.com/albums/v604/Radeon/Forum%20Photos/th_Screenshot_2012-09-16-22-40-44.png]Image[/URL]

I was going to run it on my US Sprint GS3, but they want me to pay for a benchmark. LOL, no thank you.
 
The main point is the quad-core vs dual core, 2 GB vs 1 GB here. It's like Samsung revving up a Lamborghini Aventador against an M3 in front of the traffic lights, but still losing the drag race.

Samsung chose to bloat about specs.. They'd better made sure their flagship can hold up a candle, which it can't atm :)

They have the right to brag about specs, because it is a powerful phone, and according to some tests even more powerful than 1601 points. We're comparing a phone released a couple months ago with a phone that will be released soon, and they have comparable performance.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.