The Iphone 5 outperforms Samsung's Galaxy S3 in early benchmark tests

[Steve Williams]

Samsung might want to rethink its ads
By Carly Page

THE IPHONE 5 has managed to outperform its main rival, the Samsung Galaxy S3, in an early benchmark test, despite having a lesser dual-core processor.
Compared to the Samsung Galaxy S3's 1.4GHz quad-core Exynos processor, the Iphone 5's 1GHz dual-core chip certainly sounds measly on paper. However, early benchmarking tests have shown that the Iphone 5 outperforms its Samsung rival when it comes to speed, scoring 1,601 using the Geekbench tool. In comparison, the Galaxy S3 came in 2.5 percent slower with a score of 1,560, 9to5Mac reports.
The test also shows that the latest Iphone is more than twice as fast as the latest Apple Ipad, which scored just 791 using the same benchmarking tool. That's not the only tablet the Iphone 5 surpassed, as it beat the Asus Nexus 7 too, which scored 1,591.
However, it's worth noting that results using tools such as Geekbench can easily be misleading, so we'll be sure to benchmark the two handsets ourselves when we can, perhaps later this week.
This latest development might make Samsung feel a little sheepish. Earlier today, the firm unwrapped its latest Iphone bashing advert, which mocks the Iphone 5 for its meagre processor and just 1GB of RAM. If these results are accurate, Samsung might want to rethink its ad campaign.

The Inquirer

 

[amirm]

Hmmm. I went to the source of the data and something is lost in translation. Here are the scores for Samsung (two versions of the OS):
http://browser.primatelabs.com/android-benchmarks

Samsung Galaxy Note 10.1 3G
Samsung Exynos 4412 1400 MHz (4 cores)
1831
Samsung Galaxy S III
Samsung Exynos 4412 1400 MHz (4 cores)
1751

Samsung Galaxy S III (Korea)
Samsung Exynos 4412 1400 MHz (4 cores)
1674

And:
iPhone 5 (CDMA)
Apple A6 1000 MHz (2 cores)
1586

http://browser.primatelabs.com/ios-benchmarks
iPhone 5
Apple A6 1000 MHz (2 cores)
1578

Not sure why the article numbers don't match the site it references with that data.

BTW, this is a rather synthetic benchmark and not the preferred workload one. It measures simple, pure computational operations. It does not simulate what we really like to know: how fast the browser, games, image gallery, etc. runs. In computing world, that is where we have moved to as that represents the realistic performance we can expect from the whole system. Simple tweaks in the processor can create high gains in these operations yet not result in a faster system.

Also, never trust a single number. Each app is different in what parts of the CPU it exercises and little comes out of a composite score of dozen plus tests as this one is.