iPad uses iOS the others use variations of Android with who knows whats loaded in the background.

But the simple reason is different OS's provide different performance characteristics as they handle processes and memory loads differently.

Thank-you for the reply... now my knowledge and processor IQ just went up a notch... : )

Yep. All that hardware specs means nothing if your software can't utilize it. That much is clear even way back in the days.

It's not just due to the OS, it is also due to the other hardware coupled with the A9. For instance, more RAM means application data can be loaded quicker, rather than from the HD. The GPU and screen size/resolution also effect benchmarks - the amount of effect depends on the type of test.

Also the different hardware vendors may have modified some of the firmware instruction sets to make it more efficient.

But that's a big reason why these benchmarks are used, to have some sort of common ground that more accurately compares the different hardware/software combinations.

actually , it's more a case that most current Linux ARM app writers have apparently not learn the basic lessons from the old Linux PPC days, where apple dev's clearly and logicically used to optimise all their app's C routines and wrote macros with SIMD code use etc.

as can be clearly seen here for instance SIMD makes a massive speed difference ,as can be seen in just one sample routine from the x86
http://pastebin.com/T0jt8VUB
"x264: All tests passed Yeah :)
nop: 196
optimize_chroma_dc_c: 415
optimize_chroma_dc_sse2: 203
optimize_chroma_dc_ssse3: 200
optimize_chroma_dc_sse4: 190
optimize_chroma_dc_avx: 177
"
its not uncommon to see up to *18 times faster* speeds in SIMD than C routines for instance.

until today's ARM NEON dev's stop relying on their crappy auto vectorising GCC to get a boost and start looking at the actual output of the compiler to see it generate brain dead SIMD code in many cases, then write better SIMD code by hand for their most used C routines , most apple dev's will continue to beat Linux ARM dev's on speed and data throughput it seems

ARM NEON SIMD was a 2009 GSOC project for x264

checkasm --bench
from the x264 git should give you interesting result's on the ARM NEON capable SOC too, even if its missing lots of SIMD code right now. good for seeing the generic C routines speed's and compared on different ARM SOC for instance, try it and report back you're speeds .

this real life Benchmark test came from the old x264-dev logs if anyone's interested in the real life Number's.

and StippenG's number's came from an older Quad A9/NEON developer board at his Uni apparently, so a current Marvell ARM v7 A9 quad /SIMD at 1.6 GHz for instance plus any/all the higher clocked 1GHz dual core ARM cortex A9 would produce a better result today OC.

640x360 at Ultrafast: 38.59 seems like a very good real life start for encoding on ARM cortex even without those extra SIMD patches being written yet

"2010-08-24 15:39:19 StippenG Some X264 Benchmarks (Rush Hour 640x360,preset=medium, crf=24): 4-core Cortex-A9 @ 400 MHz gives 5.55 fps,Beagleboard (A8 @ 720MHz) gives 1,65. Really nice speedup, considering the much higher frequency of the A8

2010-08-24 15:39:56 Dark_Shikari It'd go a lot faster if you used a faster preset.
2010-08-24 15:40:01 Dark_Shikari Or if you wrote some of the asm we're missing
2010-08-24 15:40:26 Dark_Shikari But yeah, that scales surprisingly well. about 3.5x faster

2010-08-24 15:40:27 StippenG Yes. Superfast gives 22.07. Ultrafast: 38.59
2010-08-24 15:41:23 StippenG Guess the out-of-order execution and shorter pipeline is really quite a bit better for performance

2010-08-24 15:41:35 < Dark_Shikari> Well, the A9 is known to be a lot faster
"

finally , making the effort to actually port "yasm" to ARM NEON would be a very good thing if you care to try and improve speed there

We also don't know the clocks of the A5. Maybe it's not safe to assume it's running at 1.0ghz?

Er you're very right about that. Geekbench reports 900MHz :)

Take care,
Anand

http://www.apple.com/ca/ipad/specs/

Its 1GHz. Geekbench reports the instantaneous speed, so you'll hear different numbers from that depending on what it ramps its speed down to to save power.

It's all about the OS at that point, just like how iOS 4.3 gives 2.5x increase in Javascript performance compared to iOS 4.0 even using the same original iPad.

Subject Comment

Please login or register to post a comment.
User Name Password