The web is flooded today with the latest Apple benchmarks and review sites are crowning the new A7 chip fastest ever, until Intel responded by posting a video with even faster times…
Benchmarketing in full swing. The original data are from AnandTech, but the graphic is mine to show the new line-up after Intel posted even faster results. http://www.youtube.com/watch?v=UXCobb134Vo
P.S in case you missed the latest developments on mobile: Intel has a new and radical improved SoC named the Bay Trail. The version in the graph is intended for tablets so don't expect a phone to beat the iPhone 5s real soon now.
PS 2: Benchmarketing is more than responding to the competition. It's mostly picking favorable data points. Sunspider is a popular benchmark meaning it's optimized to death.
PS 3: Laymen's version: these data won't mean anything for your actual results in everyday usage but you will see them all over the place as trophy charts… #Tech
And they tested with one hand tied behind their back… IE 11?
Does anyone know how much RAM these phones have yet?
Everyone does +Fraser Cain this test is with native browsers so Apple can (and does) optimize the hell out of Safari.
Anand Shimpi states that the iPhone 5s comes with 1GB LPDDR3 RAM +Mike Elgan
1GB LPDDR2 +Mike Elgan Same as on the 5 hence the 'adaptive' multitasking meaning you can't control which apps can switch.
5 S has LPDDR3, 5 C has slower LPDDR2 (same as the iPhone 5). LPDDR3 DRAM chips can shuffle information in and out at 1,600 Mbps (compared to 1,066 Mbps for LPDDR2), with up to 12.8 GB/s of bandwidth.
And Sunspider is basically a terrible benchmark. It ceased being relevant even to Javascript VM writers ages ago — it tests a random grab bag of non-representative features (e.g., pushing heavily on the native regex engine), runs a fixed number of times (leading to bad sampling error), and the like.
If I wanted to show off the performance of a CPU, I wouldn't use a Javascript benchmark in the first place. But if I did use a JS benchmark, I'd use V8's suite, not Sunspider.
if I can't buy a phone with this Intel processor today, the comparison is kind of meaningless…
+Marc Jeuken have a link for the ddr3 mem?
sure sunspider is an optimizing target, not a predictor of real world performance +Joel Webber and if +Marc Jeuken if right that Apple now uses ddr3 I'm surprised they didn't gain more compared to the iPhone 5.
Sure +Markos Giannopoulos although based on Apple's history this processor will stay their flagship for a year from now.
New phones with the Intel will be launch before Christmas so not totally irrelevant.
+Max Huijgen Sure — and I'm not suggesting that any artificial benchmark is a terribly useful indicator of real-world performance on any computing device. I'm just amazed that they're still focusing on a demonstrably bad benchmark 😛
+Ted Ewen
+Joel Webber the continuous problem of benchmarking and the reason I call it bench marketing
Just when the test itself is completely outdated it becomes an optimizing target.
+Max Huijgen until there is an actual product that exists in the market and can be bought, Apple's claim is valid.
Beyond that, there is benchmark speed and there is often the perception of speed in real usage which sometimes can be much different.
Meh. Performance is important, but power usage is much more important. Is there a comparison of the power usage of these devices?
+Markos Giannopoulos isn't that exactly what I'm saying in the post itself?
Not yet +Alok Meshram but I expect that the tested Intel SoC will use more power.
+Alok Meshram True enough, but don't forget that better performance at a given power translates to more idle time for the CPU. Of course, if the Intel SoC uses much more power to get this level of performance, it could be a wash or worse.
Yip +Joel Webber 'Race to idle' is one of the rare cases where progress is good for the battery and the user.
I'm no Apple fanboy, but I think they deserve respect for this result after only a very short time in the cpu business.
Fully agre +Steven Roose Playing with the big boys after just a few years is an amazing accomplishment.
I do not say AP performance does not matter but given the universal requirement for connectivity to do mobile computing, the performance of AP alone no longer tells the whole (real world) story of application experience on mobile devices IMHO.
+Markos Giannopoulos iPhone 5s doesn't go on sale until next week, so does that water down that in the comparison as well?
+Esa Edvik i think reviewers already have it the exact same product that you can buy on Monday and can run the tests. So it's not the same products that have not been announced to will exist sometime in the future 🙂
Appale