iOS / iPhone and Android benchmark charts

[David (PassMark)]

We have now released PerformanceTest mobile for both the iPhone and Android. Like the full Windows release you can benchmark your CPU, video, memory and disk.

All the details & the downloads for the benchmark apps are on this page.
http://www.passmark.com/products/pt_mobile.htm

We have also be compiling results into charts on 2 different sites,

Performance benchmarks chart for Android devices can be found here,
http://www.androidbenchmark.net/

Performance benchmarks chart for iOS devices (iPhone, iPad, iPod) can be found here,
http://www.iphonebenchmark.net/

Screen shot from iPhone.

[godfather]

Very nice. Will it at all be possible to use this to compare iphones to PC hardware?

Also, will it support tablets? and which ones?

Thanks!

[David (PassMark)]

Any tablet running Windows (x86), iOS, or Android should be OK.

You can somewhat compare the numbers for the individual tests with a PC, but it isn't always a fair comparison. For example on Android, all the code is written in Java, which tends to be slower than the C++ and assembler code used in Windows. There are also compiler differences and some differences in the actual tests. For example on Windows we have access to much more RAM than on iOS and Android. There are also SIMD instructions on the Intel CPUs that aren't available on iOS and Android. There is also normally a CD/DVD drive in Windows PC's.

[keithwins]

Android & PC comparisons

This seems strange to me. The best Android Passmark ratings are around 2300, which is on par with a Core i5 at 2.4 GHz. That seems unbelievably fast for the Android. Is that right? Am I missing something, other than the caveats above (Android code in Java, etc).

Keith

[David (PassMark)]

The overall tests results are a scaled weighted average. Which uses different weightings between Windows and iOS and Android as the tests are different.
We are trying to keep it the same between iOS and Android.

What you can compare is the individual tests. For example Integer maths, which is measured in millions of operations per second (MOps).

So for example,
The new Asus Transformer Prime TF201 running Android with the NVIDIA Tegra 3 Quad-core CPU gets around 160 MOps for Integer maths.

A dual core AMD Athlon 64 X2 Dual Core 5400+ also gets around 160MOps. This was a CPU from 2006.

But you also need to look at more than just the Integer maths score as different CPUs will be strong in different areas.

[keithwins]

It's almost like you're trying to say

the subtleties matter. Bah!

Seriously though, thanks for a great set of tools & reports.

[Hun73rdk]

Is there enyway to get the file around google play
i cant download it i have desire with andriod miui 2.3 and i get this message.

This app is incompatible with your 3 HTC HTC Desire.

[Richard (PassMark)]

We have not put any phone on the block list, but it seems some users are having trouble downloading from Google Play.

An alternative to Google Play is to download it from Amazon AppStore.

[Hun73rdk]

We have not put any phone on the block list, but it seems some users are having trouble downloading from Google Play.

An alternative to Google Play is to download it from Amazon AppStore.

i cant get it from amazon.com its US citicens only

• We could not process your order because of geographical restrictions on the product which you were attempting to purchase. Please refer to the terms of use for this product to determine the geographical restrictions.
  
  We apologize for any inconvenience this may have caused you.

[Richard (PassMark)]

We contacted Google Play support to see why our App may be blocked for some devices.

None of the manifest settings should exclude any particular device. The app does require that the device should be running Android 1.6 and above, which shouldn't be a problem for a most devices as 99% of users are at least running 2.1 (as reported via Android website.)

Will update once we get more information.

[emc_2u2]

Hello Richard.

First ofall thank you guys for your nice softwares, and in advance to the attention you'll pay to this request of mine.

Second off, today I dowloaded from the App store your Passmark for iOS to test my almost brand new (2 month old) new iPad (3, recognized as Verizon LTE model even if I bought it from Vodafone Italy), with 64 GB static memory(the exact model number is: MD368TY); well, while all the other tests are aligned to the updated values (and reltive graphs) you published on your web sites, the 2D *and* 3D results I've got from the model I own are HALF the ones I can see from your charts... Is this, to your opinion, a matter for which I should ask Apple to investigate and exchange my iPad, or is there any logical answer to this behaviour with your software other than a defective unit I might have?

TIA and keep up th great work.

Regards,

ME.

[Richard (PassMark)]

Thanks for bringing this issue up. I took a look into the raw numbers being submitted and determined that there were some bad results being submitted and in combination with an issue with our chart generating script for the iOS benchmarks. I apologize, as this should have been noticed sooner.

In V1.1.0 in the iOS PT App, we fixed some issues we were seeing in the 2D and 3D tests for newer Apple devices. Some devices would score exceptionally high on graphics test. The graphics tests would be performed but nothing was displayed, or selective frames (ghosting) would be displayed on the screen resulting in high results. The issue was corrected in V1.1.0. We (thought?) we updated the chart generating script to filter the bad results out, but it seems it never got pushed live.

[emc_2u2]

On updated charts

[QUOTE=Richard (PassMark);1352 Thanks for bringing this issue up. I took a look into the raw numbers being submitted and determined that there were some bad results being submitted and in combination with an issue with our chart generating script for the iOS benchmarks. I apologize, as this should have been noticed sooner.

In V1.1.0 in the iOS PT App, we fixed some issues we were seeing in the 2D and 3D tests for newer Apple devices. Some devices would score exceptionally high on graphics test. The graphics tests would be performed but nothing was displayed, or selective frames (ghosting) would be displayed on the screen resulting in high results. The issue was corrected in V1.1.0. We (thought?) we updated the chart generating script to filter the bad results out, but it seems it never got pushed live.[/QUOTE]


Dear Richard,

thank you instead for your kind and prompt reply.

Let me better understand this: you mean the "reference" charts images on the appropriate web site of yours regarding the iOS version of Passmark currently show incorrect values for the graphics subsystem of new iPad, and this will be fixed with the new reports of v1.1.0 (which is the version I downloaded and used)?

Also, I forgot to mention that when I tap "submit", the app prompts me with a 5 digits (reference) number and immediately takes me to your web site but in the "general" page and not specific to the test I just asked to do (I did it 3 times and it always behaved like that).

Last but not least, I'd also would like to truly thank you people at Passmark from the deep of my heart because your app seems to be the only one to use as a reference tool to test the correct functionality of the iOS systems (it would be great to have some more objective info about the actual graphics of the device, like the current or "real" (max) resolution, to name the most important), and Apple to my opinion should integrate it in its basic system tools (the mere model number just tells me my device SHOULD come from a certain production batch, but it's not a guarantee of its "integrity" and correct functionality).


TIA again and have a great day.

Regards,

ME.

[Richard (PassMark)]

Thank you for the positive feedback. There are plans to further develop the graphics test in the future releases. However, there isn't a timeline on when the changes will occur.

The graphics charts on the site have been corrected. The script was updated to filter out the prior bad results and display results from v1.1.0.

[emc_2u2]

Positive thinking

Thank you for the positive feedback. There are plans to further develop the graphics test in the future releases. However, there isn't a timeline on when the changes will occur.

The graphics charts on the site have been corrected. The script was updated to filter out the prior bad results and display results from v1.1.0.

You guys obviously deserve all the credits for my (and not only) most sincere thanking.

Sorry for the late reply, I've checked the new graphs and now my device seems aligned; odd results, though, but I see these handheld devices have several differences compared to more traditional hardware thus it's difficult to correctly "calibrate" benchmarking software for them.

If you might need any further input or if you think I might help you in any way, just let me know and I'll be more than happy to contribute.

Greetings,

ME.