iPad 2 CPU performance

[firewood]

I just benchmarked an iPad 2 as around 4.5X slower than an MBA 11 1.6 GHz at general integer computing, but only around 2.5X slower at some floating-point number crunching. The MBA 11 is about 1.75X heavier, so it wins in CPU performance-to-weight ratio. But not by that much over something not even being advertised as a fairly powerful number-crunching computer.

The iPad 2 also benchmarked between 1.4X and 2X faster than a 1st gen iPad. The benchmarks I ran were in fairly generic single-threaded portable C code compiled with full optimizations using Xcode.

 

[iPadThai]

I just benchmarked an iPad 2 as around 4.5X slower than an MBA 11 1.6 GHz at general integer computing, but only around 2.5X slower at some floating-point number crunching. The MBA 11 is about 1.75X heavier, so it wins in CPU performance-to-weight ratio. But not by that much over something not even being advertised as a fairly powerful number-crunching computer.

The iPad 2 also benchmarked between 1.4X and 2X faster than a 1st gen iPad. The benchmarks I ran were in fairly generic single-threaded portable C code compiled with full optimizations using Xcode.

iPad 2: $499+tax
MBA: $999+tax

iPad2: portability > MBA. In fact, I use more of my iPad 2 than my MBP 17" 2.2i7quad.

Those benchmarks mean jack nothing. When I pick up the iPad 2, it's so butter smooth and everything runs so well (games and web), I know the iPad 2 > ipad 1. Compared to the MBA? Well, if you need an MBA, then you need a laptop.

 

[firewood]

Those benchmarks mean jack nothing.

These benchmarks mean a lot, since I plan to do a lot of scientific number crunching and data visualization on the iPad 2, and develop apps for other professionals who also plan to do such. The iPad is not just for fun and games.

YMMV.

 

[iPadThai]

These benchmarks mean a lot, since I plan to do a lot of scientific number crunching and data visualization on the iPad 2, and develop apps for other professionals who also plan to do such. The iPad is not just for fun and games.

YMMV.

I don't think the iPad 2 is meant for anything but fun and games. Sure you can use it for other computational purposes but I don't think you really think it can replace a laptop for serious work right? 99% of the buyers of iPads don't care about the benchmarks.

 

[firewood]

I don't think the iPad 2 is meant for anything but fun and games.

The FDA has approved iPad apps for professional medical use, and the FAA has approved iPad apps for commercial aviation use. The banking and financial industries are developing lots of enterprise iPad apps for internal use.

That's a lot more than fun and games.

And it's also just the beginning.

 

[iPadThai]

The FDA has approved iPad apps for professional medical use, and the FAA has approved iPad apps for commercial aviation use. The banking and financial industries are developing lots of enterprise iPad apps for internal use.

That's a lot more than fun and games.

And it's also just the beginning.

Well, then, I'm the minority who just web surf and play games and garageband on the ipad 2. Even if your benchmarks prove that it's slower than an MBA, the portability of these things just trumps the MBA in performance. But like I said, perhaps you really need more power of a laptop. The ipad 2 is more of a media consumption device rather than creation. That's just my 2 cents.

Anyway, I love my iPad!

 

[WMuntean]

I also plan on running some computational intensive code on the iPad 2, however, within a comfortable range. For my purposes, I intend to provide the user with options to run either an unaltered version of computations or a stripped down -- time saving -- version. Its my hope that most will find the stripped down version sufficient.

 

[firewood]

Even if your benchmarks prove that it's slower than an MBA

Not by that much!

the portability of these things just trumps the MBA in performance.

That was almost exactly the point.

It doesn't "trump" an MBA or MBP in performance, but it's clearly close enough, especially for some scientific number crunching.

I also plan on running some computational intensive code on the iPad 2, however, within a comfortable range. For my purposes, I intend to provide the user with options to run either an unaltered version of computations or a stripped down -- time saving -- version.

I wrote some stripped-down lower-res compute code for the iPhone 3G. But the iPad 2 seems fast enough that I'm strongly considering throwing out all that stripped-down code, and just using the just the unaltered full-blown compute kernels. Decent performance and less code to maintain.