Benchmarking the iOS 5 update

17 Comments

Sure, there are plenty of new features with iOS 5, but at what cost? Sometimes, the cost for software updates comes in the form of degraded performance of day-to-day core features that you may depend upon. Let’s see if that’s the case with iOS 5.

iOS 5 update tested

Once again, I used GeekBench 2 ($0.99), Gauge Mathematical Tool (free), BenchTest ($0.99), SunSpider (free) and V8 (free) to test the performance of each OS version on each device. I tested the original iPad 1 Wi-Fi, the iPhone 4 (GSM), the iPod touch 4th Gen, and the iPad 2 Wi-Fi+3G (GSM).

Conclusions

If you use the web more than anything else on your iOS device, then you should be eager to apply each and every update that Apple puts out, since they all improve Safari’s performance every time. iOS 5 is no exception in that regard.

Considering all of the new features you get with iOS 5.0, the hit on performance in other areas is in most cases negligible.  In fact, with the exception of the original iPad (the older device that was tested), all devices actually gained in performance in most areas. That is simply amazing for the first release of a major OS update. This is nothing like the 4.0 release, where we saw incremental performance gains over the lifetime of 4.0 to restore what was lost to older device; iOS 5.0 is a solid release for all where performance is concerned.

17 Comments

Boudewijn

Could you also check how much memory is free on the different iOS versions?

I’m having quite often problems with low memory on the iPad 1 so especially games keep crashing, forcing me to close all other apps int he background and retry.
I hate to see even more less memory free by upgrading to iOS 5 which doesn’t have any features that I want to use anyway…

Geoffrey Goetz

Unfortunately I cannot go back and see how each device managed memory under typical daily usage, as I cannot go back to prior versions of iOS.

It is important to keep in mind that developers may not have the same mix of Apps installed that is creating a memory challenge on your iPad 1. Some developers may not have the opportunity to test their Apps on each of the possible devices du to limited time and resources. I would suggest informing the developer directly of the situation you are facing, and see if they have a fix already in the works.

Dan Mitchell

I upgraded last night east coast at about 7:30. I had no problems and all my apps except 1 (Alarm Clock) transferred. I have been noticing one problem today, my E-Mail loads very slow and then to return from reading an E-Mail to the E-Mail list almost seems to freeze for about 10 seconds. Everything else is fine.

Geoffrey Goetz

Regarding the 3GS, I have not had access to a 3GS over the last year I have been collecting such data. I have data points for every iOS release on the devices I do have.

If you are looking for 3GS before and after data, GeekBench makes posted results available for online viewing. Just search for “iphone2,1 5.0” to show the results of the iPhone 3GS running iOS 5.0, and search for “iphone2,1 4.3.5” to show the results of the iPhone 3GS running iOS 4.3.5. Pulling data this way can reduce the opportunity for error in the results as noted in an earlier comment. http://browse.geekbench.ca/

Daniel C Bittencourt

The First table (iPad 1 WiFi) is full of errors!
How come in the BenchTest test the best score was in iOS 4.2.1 if in the iOS 5.0 the score was 4.055?
And in the Drawing into View iOS 4.3, 4.3.5 and iOS 5 are marked as best ones?
And how many times the tests were conduced? There should be an error margin in tests like these!

Geoffrey Goetz

The results shared are snapshot point in time measurements. Apologies on the highlighting. The iPad 1 was actually the only device I tested that deviated from the others.

Daniel C Bittencourt

OK. I understood now. Glad I helped. But it’s really interesting that iPad 1 didn’t follow iPhone 4 results. I think memory really makes difference here.

Alex

Yes, please can you test the 3GS, as it would be very useful to see the change in speed.

As i can remember the change from iOS 3 to 4 dramatically decreased the productivity in 3G, so was wondering if it will do the same to 3GS.
Thanks

echomrg

i absolutely agree with the other commenters.
considered the way iOS 4 practically bricked the old 3G rendering it almost useless some information on benchmarks on a 3GS would be very useful.

M

Josiv

My 3gs went from 28k to 40k testing with Browsermark. Really sped things up using iOS5. : )

doomtints

Why didn’t you test it on the 3GS? That’s the one device where I would expect to see a noticeable performance hit.

Comments are closed.