Frax app sees 90 percent performance boost on Apple’s 64-bit A7-powered iPhone 5s

“Is Apple’s A7 chip twice as fast at processing and graphics, as Apple promised when announcing the new iPhone 5S?” Stephen Shankland reports for CNET. “At least for Ben Weiss, the answer is yes — but his app is unusually well suited to take advantage of the new processor’s features.”

“The app in question is Iter9’s new Frax app for generating lavishly detailed fractal imagery by running mathematics calculations on both the CPU and GPU,” Shankland reports. “It gets a major speed boost on the iPhone 5S, Weiss said, evident in faster rendering times for the swirling psychedelic images laboriously calculated one pixel at a time.”

Shankland reports, “Frax is currently a 32-bit app, Weiss said, and it runs 50 percent faster on the A7-powered iPhone 5S compared with the A6-powered iPhone 5. And then came the second speed boost: When he compiled the first 64-bit version, Frax ran 25 percent faster than the 32-bit version on the iPhone 5S. Together, that means the 64-bit app runs nearly 90 percent faster on an iPhone 5S than the 32-bit version on the iPhone 5.”

Read more in the full article here.

MacDailyNews Take: A 90 percent performance boost. That’s quite the “marketing gimmick.”

Related articles:
Apple’s 64-bit A7 already powering advanced new audio, video features in apps and games – October 4, 2013
Apple takes the high-end with their 64-bit A7; leaves rivals to scrap for survival in low-end – October 2, 2013
Apple’s 64-bit A7 is no gimmick: New iPhone 5s offers major performance leap – September 25, 2013
Ars Technica: Apple’s Touch ID and 64-bit A7 are deceptively large advances in the iPhone’s evolution – September 24, 2013
Apple iPhone 5s reviews are universally positive, many crown iPhone 5s the best smartphone – September 19, 2013
Engadget reviews Apple iPhone 5c: A breath of fresh air that will be wildly popular this holiday season – September 18, 2013
Apple’s 64-bit iPhone 5s is by far the fastest smartphone in the world – September 18, 2013
John Gruber reviews Apple iPhone 5s: ‘This is what innovation, real innovation, looks like’ – September 18, 2013
AnandTech reviews iPhone 5s: Apple’s 64-bit A7 is seriously impressive – September 18, 2013
TechCrunch reviews Apple iPhone 5s: The best smartphone available – September 18, 2013
USA Today’s Baig reviews Apple iPhone 5s: ‘Makes the best smartphone even better’ – September 18, 2013
Mossberg reviews Apple iPhone 5s: ‘The best smartphone on the market’ – September 18, 2013

23 Comments

  1. The day will come where we can do “Laptop Class” work with an iOS device.

    I don’t envision creation so much as display or minor editing of big and complex documents. Just as I type this, though, I realize that miniaturization is upon us and an iPad that can work like a laptop is almost here.

  2. wait, what does “runs 90% faster” mean? if you had been running the mile in 4 minutes (240 sec) then does 90% faster mean you can run it in 24 seconds? i think what they are trying to say is that it took a little more than half the time to run, which is not 90% faster.

    1. Maybe a review of grade 4 math would help you. Although you have a computer, for many it is difficult to use it for anything but surfing the net or posting on discussions like this one. We can hardly wait until you get to sixth grade and improve your understanding of basic math.

      1. i am not so good at 4th grade math, but i can do 3rd grade math pretty well. if case c is 25% “faster” than case b which is 50% “faster” than case a, then case c would be 62.5% “faster” than case a. 25% “faster” does not mean one-fourth the time, that would be 75% “faster”. you are all confusing yourselves because for 50% it is the same thing (symmetry). according to your logic 1% faster would be really fast, and 0% “faster” would be no time at all. thus 25% faster than 50% would be .5 x (1 – .25) or 37.5%, not 88.5% or 87.5% as were mentioned below, and not close to 90% “faster”. what i was annoyed about to begin with, though, was the use of the word faster. it is a misleading word to use when you are talking about decreasing something that has a lower limit of zero. it is an inverse relationship, not linear as in the case when the time is increasing. if case a takes 100 sec and case b takes 1 second and case c takes .1 sec, it seems like case c isn’t that much better than case b, when there is effectively an order of magnitude difference between case b and case c. now, i have to admit that in my hurry to type my thoughts above i said “little more” rather than “little less”, but the point is the same: it isn’t any where near 90%.

  3. Let’s say the A5 32 bit app takes 100 time units to run. Then 50% would be 50 time units for the A5s 32 bit app. And the 64 bit app requires only 25% of the A5s run time: 25% of 50 time units equals 12.5 units. This indicates the A5s 64 bit app runs 88.5% faster than the 100 time units for the A5 32 bit app: 12.5/100 = 12.5% of the base time for the A5 32 bit app, or 88.5% faster. This is “nearly 90%”. Any questions?

    1. the article did not say only 25% of the time, it said 25% “faster”; there is a big difference. as hannahjs implied below, word problems suck because you need to be able to comprehend english before you can do the math. and problems involving percentages are almost always guaranteed to cause confusion.

  4. Since the dual core A7 64 bit chip easily beats the quad core Qualcomm chips running at twice the clock speed and the fact that very very few mobile apps are written to take advantage of 4 cores, wouldn’t it be a fair assessment that the quad core chips are a marketing gimmick?

    1. That’s true, the app hasn’t been specifically tuned yet for 64-bit and GL ES 3.0. But that will happen, and I expect good results.

      To clarify the math: “25% faster” means that the app can process 25% more pixels per second than it could before. (And on the GPU side, at least 50% faster, though it was difficult to measure accurately because it kept pegging at 60fps.) With some dedicated time spent tuning for the 64-bit architecture, I think I can get the CPU performance to 2x over the 32-bit implementation.

      In a nutshell: the A7 chip is very impressive due to its overall architectural changes; mostly the doubling of registers and new instruction set. Recompiling for the new architecture was a tremendous win for very little effort.

  5. Why does the author keep reinforcing the fact that 64 bit is really doing nothing here. It’s all in the updated CPU and not really the 64 bit goodness. According to the author several times on the article. So what is he trying to imply then?

    1. The “64-bit” architecture of the A7 introduces several new features: a new instruction set, twice as many registers, and 64-bit memory addressing. The 64-bit addressing is the least relevant, mostly because today’s iDevices still have less than 32 bits (4GB) of physical memory to address. But the new instructions and extra registers give a very significant performance advantage.

  6. This is another nail in Apple’s coffin. The majority of the world’s population buys Android-based devices. According to recent samplings of real-world analytics, they use their phones as phones, and do little in the way of Internet web browsing or commerce, let alone rendering the Mandelbrot set hour after hour.

    1. If, as you allege, people buy Android phones just to make phone calls, then why on earth incorporate the ability to run apps or browse the web. By your argument, that functionality on Android is pointless, which is a pretty good summation of Android.

      Thanks for so eloquently making the case for the superiority of iOS over Android.

      =:~)

  7. Let’s simplify it:

    If iPhone 5 runs at a pace of 100 actions per minute (just for example), then a 32 bit 5S app runs at 100 + 50% = 150 actions per minute, (or 100 actions in 40 secs) and a 64 bit 5S app runs at 150 + 25% = 187.5 actions per minute (or 100 actions in 32 secs).

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.